Journal of Intelligent & Fuzzy Systems: Applications in Engineering and Technology

Intrusion detection is an important requirement in wireless mesh network and the intrusion detection system (IDS) provides security by monitoring data traffic in real time. This work proposes support vector machine (SVM) classifier to identify the intrusion in the network. The traffic data collected from the wireless mesh network (WMN) is given as input to the SVM. The irrelevant and redundant input variables increase the complexity of designing IDS and may degrade its performance. Hence, feature selection techniques, which select the relevant features from the original input is essential to improve the performance of IDS in WMN. In this work, a hybrid genetic algorithm (GA) and mutual information (MI) based feature selection technique is proposed for IDS. The performance of IDS with the proposed feature selection technique is analyzed with IDS having mutual information, genetic algorithm and GA+MI based feature selection techniques using ADFA-LD dataset. Experimental results have demonstrated the effectiveness of proposed intrusion detection system with hybrid feature selection technique in wireless mesh network. The superiority of SVM classifier with hybrid feature selection technique is also verified by comparing with artificial neural network classifier.

Cloud computing offers an economical, convenient and elastic pool of computing resources over the internet. It enables computationally weak client to execute large computations by outsourcing their computation load to the cloud servers. However, outsourcing of data and computation to the third-party cloud servers bring multifarious security and privacy challenges that needed to be understood and address before the development of outsourcing algorithm. In this paper, the authors propose solutions for matrix-chain multiplication (MCM) problem. Our goal is to minimize the execution burden on the client without sacrificing the confidentiality and integrity of the input/output. Conventionally, the complexity of matrix-chain multiplication is O (n³). After leveraging the facility of outsourcing, the client-side complexity reduces to O (n²). In the proposed algorithm, the client employs some efficient linear transformation schemes, which preserve the data confidentiality. It also developed a novel result verification scheme, which verifies the result with modest burden and high probability and maintain the integrity of computed result. The analytical analysis of algorithm depicted that the algorithm is simultaneously meeting the design goals of correctness, security, efficiency and verifiability. We conduct many experiments to validate the algorithm and demonstrate its practical usability. The algorithm is implemented on public cloud “Amazon EC2”, and found that the proposed outsource algorithm performs 11.655 times faster computation of matrix-chain multiplication than the direct implementation.

In recent years, domain generation algorithms (DGAs) are the foundational mechanisms for many malware families. Mainly, due to the fact that DGA can generate immense number of pseudo random domain names to associate to a command and control (C2) infrastructures. This paper focuses on to detect and classify the pseudo random domain names without relying on the feature engineering or any other linguistic, contextual or semantics and statistical information by adopting deep learning approaches. A deep learning approach is a complex model of traditional machine learning mechanism that has received renewed interest by solving the long-standing tasks in artificial intelligence (AI) related to the field of natural language processing, image recognition, speech processing and many others. They have immense capability to extract optimal feature representations by taking input as in the form of raw input texts. To leverage this and to transfer the performance enhancement in aforementioned areas towards characterize, detect and classify the DGA generated domain names to a specific malware family, this paper adopts deep learning mechanisms with a known one million benign domain names from Alexa, OpenDNS and a corpus of malicious domain names generated from 17 DGA malware families in real time for training in character and bigram level and a trained model has been evaluated on the OSNIT data set in real-time. Specifically, to understand the effectiveness of various deep learning mechanisms, we used recurrent neural network (RNN), identity-recurrent neural network (I-RNN), long short-term memory (LSTM), convolution neural network (CNN), and convolutional neural network-long short-term memory (CNN-LSTM) architectures. Additionally, to find out an optimal architecture, experiments are done with various configurations of network parameters and network structures. All experiments run up to 1000 epochs with a learning rate set in the range [0.01-0.5]. Overall, deep learning approaches, particularly family of recurrent neural network and a hybrid network (where the first layer is CNN and a subsequent layer is LSTM) have showed significant performance with a highest detection rate 0.9945 and 0.9879 respectively. The main reason is deep learning approaches have inherent mechanisms to capture hierarchical feature extraction and long range-dependencies in sequence inputs.

Long Short-term Memory (LSTM) is a sub set of recurrent neural network (RNN) which is specifically used to train to learn long-term temporal dynamics with sequences of arbitrary length. In this paper, long short-term memory (LSTM) architecture is followed for Android malware detection. The data set for evaluation contains real known benign and malware applications from static and dynamic analysis. To achieve acceptable malware detection rates with low computational cost, various LSTM network topologies with several network parameters are used on all extracted features. A stacked LSTM with 32 memory blocks containing one cell each has performed well on detection of all individual behaviors of malicious applications in comparison to other traditional static machine learning classifier. The architecture quantifies experimental results up to 1000 epochs with learning rate 0.1. This is primarily due to the reason that LSTM has the potential to store long-range dependencies across time-steps and to correlate with successive connection sequences information. The experiment achieved the Android malware detection of 0.939 on dynamic analysis and 0.975 on static analysis on well-known datasets.

Without an iota of doubt, security, safety, and privacy are the most critical aspects of any Industrial Internet of Things (IIoT) environment. Among the existing intrusion detection methods, knowledge-based methods discover only the recognized attacks, the behavior-based methods suffer from high false positives, and specification-based methods demand the complete knowledge about the elements present in the IIoT environment. Examining the heterogeneous data from different and distributed sensors and sending the correct commands to actuators are vital to the increasingly industrialized economy. This work proposes an Intrusion Detection System (IDS) for the IIoT environment that combines both the anomaly and specification-based approaches. The resulting system overcomes the limitations of the contemporary techniques by detecting unidentified attacks. All kinds of data emanating from any IIoT setup comprising sensors and actuators are logged, and specification rules are constructed from it. Any violations of the created rules are treated as attacks. The validation is carried out through simulation using the Mininet tool with the dataset obtained from the real-world water treatment facility at the Singapore University of Technology and Design (SUTD). The results show only 3.2% of false positives with the detection rate of 96.4%.

Smart grid networks to deliver power, are futuristic well planned, designed and managed systems in real time. The potential threat to the dynamic security of such systems is chaotic nature of real power flow on transmission lines. Smart grid needs to be secure in static and dynamic sense under contingent conditions. At present static security assessment is based on monitoring of conventional mega-watt performance index along with other voltage and current related performance indices. This paper proposes an intelligent technique based on novel interleaved performance index. Proposed technique uses interleaving of two features. One feature is related to computation of Lyapunov Exponent for chaotic behavior related to the divergent trend of power flow increase and other related to conventional megawatt performance index for overload monitoring. Novel proposed index is named as interleaved Mega Watt-Lyapunov Exponent performance index. Chaos quantification and overload of real power flow on transmission lines, both can be evaluated using proposed index. This paper presents formulation and computation of novel proposition for smart grids. Benchmark model with computed values of novel performance index under different contingencies and load profiles along with a comparison to conventional index, is demonstrated and discussed at length in the paper.

The mobile Internet Protocol (IP) is a mobility based communication protocol that provides guidelines for the routing of mobile nodes in a network. The mobile IP manipulates IP addresses as a natural identifier for each mobile communicant. Here, each mobile device recognizes itself via two IP addresses: a home-of address and a care-of address. Owing to the mobility nature of the these devices, the location update of their current care-of address plays a vital role in receiving continuous services without interference. Many investigations have been explored on the location update of mobile devices along with the security and computation issues. However, the efforts on the security services have not received much attention in these investigations. Consequently, there is an increasing need for optimized binding update that balances security and efficiency. In this paper, a new Binding Update using Twofold Encryption (BUTE) is proposed, for balancing both security and efficiency of binding update for IPv6 mobility. It exhibits the alleviation of the attacks such as rerun, man-in-the-middle, false binding update and denial-of-service. The proposed BUTE is simulated using network simulator-2 and the experimental results are analyzed. Also, it is validated for security attributes using Automated Validation of Internet Security Protocols and Applications (AVISPA) – a security tool. Finally, the numerical results reveal that the proposed BUTE provides a significant reduction in communication cost and binding update delays.

Recently, a color image encryption algorithm is suggested by Lakshmanan et al. in [IEEE Transactions on Neural Networks and Learning Systems 2016, doi: 10.1109/TNNLS.2016.2619345]. In encryption algorithm, a piece-wise linear chaotic map is adopted to create a permutation matrix to perform shuffling of pixels of plain-image. The encryption of shuffled image is proceeded by extracting keystreams from chaotic inertial delayed neural network. This paper evaluates the security of encryption algorithm to unveil its inherent defects and proposes to present a complete cryptanalysis. To demonstrate the break of algorithm, we apply proposed chosen-plaintext attack that successfully recovers the exact plain-image from encrypted image without secret key. The proposed cryptanalysis shows that the encryption algorithm is inapt to realize a secure communication, the security claims by authors are not valid as the algorithm has serious security flaws and susceptible to proposed attack.

Malicious uniform resource locator (URL), termed as malicious website is a foundation mechanisms for many of internet criminal activities such as phishing, spamming, identity theft, financial fraud and malware. It has been considered as a common and serious threat to the Cybersecurity. Blacklisting mechanism and many machine learning based solutions found by researchers with the aim to effectively signalize and classify the malicious URL’s in internet. Blacklisting is completely ineffective at finding both variations of malicious URL or newly generated URL. Additionally, it requires human input and ends up as a time consuming approach in real-time scenarios. Machine learning based solutions implicitly rely on feature engineering phase to extract hand crafted features including linguistic, lexical, contextual or semantics, statistical information of URL string, n-gram, bag-of-words, link structures, content composition, DNS information, network traffic, etc. As a result feature engineering in machine learning based solutions has to evolve with the new malicious URL’s. In recent times, deep learning is the most talked due to the significant results in various artificial intelligence (AI) tasks in the field of image processing, speech processing, natural language processing and many others. They have an ability to extract features automatically by taking the raw input texts. To leverage this and to transform the efficacy of deep learning algorithms to the task of malicious URL’s detection, we evaluate various deep learning architectures specifically recurrent neural network (RNN), identity-recurrent neural network (I-RNN), long short-term memory (LSTM), convolution neural network (CNN), and convolutional neural network-long short-term memory (CNN-LSTM) architectures by modeling the real known benign and malicious URL’s in character level language. The optimal parameter for deep learning architecture is found by conducting various experiments with various configurations of network parameters and network structures. All the experiments run till 1000 epochs with a learning rate in the range [0.01-0.5]. In our experiments, deep learning mechanisms outperformed the hand crafted feature mechanism. Specifically, LSTM and hybrid network of CNN and LSTM have achieved highest accuracy as 0.9996 and 0.9995 respectively. This might be due to the fact that the deep learning mechanisms have ability to learn hierarchical feature representation and long range-dependencies in sequences of arbitrary length.

With the advancement of cloud computing, data-owners prefer to outsource their confidential data to cloud as it provides storage facility at reduced administration and maintenance cost. When these confidential documents leave the user’s premises, security of the data becomes a prime concern, as cloud service provider (CSP) and end users are in different trust domains. To provide data privacy, it is preferred that the end user should encrypt the data before outsourcing to the cloud. Encryption provides security but makes data utilization a challenging task, i.e., searching encrypted documents is difficult. Various schemes exist in the literature but they are either restricted to single keyword search or are inefficient in terms of search time required. In this paper, we propose an efficient approach for secure information retrieval using the concept of bucketization. This reduces the average number of comparisons per query to sub-linear. Experimental analysis on Reuters-21578 data-set demonstrates that the proposed scheme provides a recall of 100% and precision of 98.45% while decreasing the search time by 98.85% using bucketization.

Threats related to computer security constantly evolving and attacking the networks and internet all the time. New security threats and the sophisticated methods that hackers use can bypass the detection and prevention mechanisms. A new approach which can handle and analyze massive amount of logs from diverse sources such as network packets, Domain name system (DNS) logs, proxy logs, system/service logs etc. required. This approach can be typically termed as big data. This approach can protect and provide solution to various security issues such as fraud detection, malicious activities and other advanced persistent threats. Apache spark is a distributed big data based cluster computing platform which can store and process the security data to give real time protection. In this paper, we collect only DNS logs from client machines in local area network (LAN) and store it in a server. To find the domain name as either benign or malicious, we propose deep learning based approach. For comparison, we have evaluated the effectiveness of various deep learning approaches such as recurrent neural network (RNN), long short-term memory (LSTM) and other traditional machine learning classifiers. Deep learning based approaches have performed well in comparison to the other classical machine learning classifiers. The primary reason is that deep learning algorithms have the capability to obtain the right features implicitly. Moreover, LSTM has obtained highest malicious detection rate in all experiments in comparison to the other deep learning approaches.

With the increasingly growing internal and external attacks on computer systems and online services, cybersecurity has become a vibrant research area. Countering intrusive attacks is a daunting task with no universal magic solution that can successfully handle all scenarios. A variety of machine-learning and computational intelligence techniques have been applied extensively to detect and classify these attacks. However, the effectiveness of these techniques greatly depends on the adopted data preprocessing methods for feature extraction and engineering. This paper presents an extended taxonomy of the work related to intrusion detection and reviews the state-of-the-art techniques for data preprocessing. It offers a critical up-to-date survey which can be an instrumental pedagogy to help junior researchers conceive the vast amount of research work and gain a holistic view and awareness of various contemporary research directions in this domain.

The present work describes different research techniques for collecting and organizing speech database in different scenario at the institute and successfully structuring the text independent speaker identification database in Indian context. In order to get the Multi-Scenario dataset, each speaker performed multiple sessions recording in reading style with English and Hindi language with same passages but under different conditions. This work analyzed different scenario affecting the performance of speaker recognition system when tested under dissimilar training conditions. Here four different scenarios are considered; sensor and environment, language, aging and health. To study the effect of sensor, language and environment on the performance of ASR system a database of 200 speaker was created. Under different environmental conditions, four different types of sensors in parallel configuration were used to study the sensor mismatch conditions over testing and training phase. The database contains speech samples of the individual in English and Hindi in read speech styles under two environment i.e. a controlled recording chamber and library. To study the aging effect, an aging NSIT speaker database (AG-NSIT-SD) of 53 famous personalities was collected from online source varying over a period of 10–20 years. Further to study the effect of health, a cough and cold NSIT speaker database (CC-NSIT-SD) of 38 speakers was also collected to study the performance of system. Apart from this, the effect of different noise types on the speaker identification was also studied on different sensors.

Hierarchical Temporal Memory is a new machine learning algorithm intended to mimic the working principle of the neocortex, part of the human brain, responsible for learning, classification, and making predictions. Although many works illustrate its effectiveness as a software algorithm, hardware design for HTM remains an open research problem. Hence, this work proposes an architecture for HTM Spatial Pooler and Temporal Memory with learning mechanism, which creates a single image for each class based on important and unimportant features of all images in the training set. In turn, the reduction in the number of templates within database reduces the memory requirements and increases the processing speed. Moreover, face recognition analysis indicates that for a large number of training images, the proposed design provides higher accuracy results (83.5%) compared to only Spatial Pooler design presented in the previous works.

A well connected of network of smart devices that can be accessed through internet is broadly known as Internet of Things (IoT). IoT has diversified application exploiting various technologies that requires integration of computing devices with end users. The use of IoT provides efficient management large and heterogeneous assets, establishes a real time network for strong connectivity and set up useful coordination among various resources. For example, in defense system, IoT can be applicable in monitoring war fighter’s health, border area surveillance, proactive equipment maintenance etc. However, maintaining privacy and security through secure remote authentication is an essential prerequisite for hazard-free use of IoT based services. In this paper, we provide an efficient and secured authentication protocol for remote IoT based services. Our proposed scheme does not exploit computation costly operation and avoids heavy storage in medical application server, thus making it suitable for battery limited devices. Finally, we perform the formal security verification using the widely accepted verification tool, called the ProVerif 1.93, to show that presented scheme is secure.

Face Recognition is widely used applications such as of mobile phone unlocking, credit card authentication and person authentication in airports. The face biometric authentication system can be easily spoofed by printed photograph, replay video of the legitimate user and 3D face mask. This paper proposes hybrid feature descriptors to detect the face spoofing attack (printed photograph and replay video attacks). The proposed method extracts three different feature descriptors such as Color moment, Haralick texture and Color Local Binary Pattern (CLBP) feature descriptors. The extracted features are concatenated and classified by Logistic Regression. The performance of the proposed method is evaluated on the Michigan State University Mobile Face Spoofing Database (MSU-MFSD) dataset and found to achieve better results than state-of-the-art methods.

Biometric refers to the automatic identification of a person based on physiological or behavioural characteristics. Current modes of biometric systems are fingerprint, voice, face, signature, palm print, iris scan etc. The conventional biometric systems are unable to meet these authentication requirements as it can be forged. Hence, a novel biometric system which can overcome these limitations is proposed. Tongue is a unique vital organ which is well protected within the mouth and not affected by external factors. Dorsum of the tongue exhibits a great amount of information along with its visual differences in shape, texture and pattern which can be called the tongue print. As tongue exhibits rich textural patterns, Local Binary Pattern (LBP) algorithm is used for extracting features. Extracted features are then trained by a linear Support Vector Machine (SVM) for personal identification. From the database consisting of 136 tongue print images of 34 individuals, we achieved an accuracy of 97.05% for identification. Our study is the first of its kind where texture patterns are extracted from tongue images using Local Binary Pattern for biometric authentication. We achieved a level of accuracy compared to the technique used in other studies.

Spell checking plays an important role in conveying correct information and hence helps in clear communication. Spell checkers for English language are well established. But in case of Indian languages, especially Malayalam lacks a well developed spell checker. The spell checkers that currently exist for Indian languages are based on traditional approaches such as rule based or dictionary based. The rich morphological nature of Malayalam makes spell checking a difficult task. The proposed work is a novel attempt and first of its kind that focuses on implementing a spell checker for Malayalam using deep learning. The spell checker comprises of two processes: error detection and error correction. The error detection section employs a LSTM based neural network which is trained to identify the misspelled words and the position where the error has occurred. The error detection accuracy is measured using the F1 score. Error correction is achieved by the selecting the most probable word from the candidate word suggestions.

Social media is considered to be a vibrant area where millions of individuals interact and share their views. Processing social media text in Indian languages is a challenging task, as it is a well-known fact that Indian languages are morphologically rich in structure. On transferring such an unstructured text into a consistent format, the data is exposed to feature extraction method. In the huge corpora, information units i.e. entities holds the basic idea of the content. The main aim of the system is to recognise and extract the named entities in the social media twitter text. The proposed system relies on the proficient co-occurrence based word embedding models to extract the features for the words in the dataset. The proposed work makes use of text data from the Twitter resource in the Tamil language. In order to enhance the performance of the system, tri-gram features are extracted from the word embedding vectors. Hence, systems are trained using N-gram embedding features and named entity tags. Implementation of the system is using machine learning classifier, Support Vector Machine (SVM). On comparing the performance of the proposed systems, it can be seen that glove embedding shows better results with the accuracy of 96.93%, whereas the accuracy of word2vec embedding is 84.53%. The improvement in the performance of the system based on glove embedding with regard to the accuracy may be due to the imperative role of the co-occurrence information of glove embedding in recognising the entities.

Parallel processing is crucial for accelerating computation in many high-performance applications and modern technologies including computational modeling, optimization and simulation, Web and DNS servers, peer-to-peer systems, grid computing and cloud computing. Due to the heterogeneity nature of various processing nodes and the differences of workloads of various tasks, some processors can be idle while others are overloaded. In this paper, we present a simple, yet efficient, solution inspired by the intelligence of ant colonies to adequately mitigate the load imbalance and communication overhead problems in multiprocessor environments. The proposed approach is based on defining and maintaining data structures to dynamically track the load of each processor. We implemented the proposed algorithm and evaluated its performance under different scenarios against the baseline round-robin algorithm. The results showed that the proposed algorithm has more effective properties than the round-robin algorithm.

Illegal cyber activities can be curbed by means of authorship analysis which intends to identify the authors of a document by scrutinizing the writing style involved in it. One of the major threats associated with online media is the propagation of false statements on behalf of celebrities with the aim of tarnishing their public image especially as a part of online political campaigns. The scenario calls for the need of analyzing the authorship of documents with less contents and capturing the author style from among a large number of candidate authors belonging to the same domain. This is a less explored area of authorship analysis as the task is challenging because traditional methods fail to acquire accuracy when the contents of different authors are pertaining to same topic. Here we propose a method that accomplishes the task of analysis in such an environment, by employing psycholinguistic, lexical, and syntactic aspects of an author combined with word co-occurrences obtained by modeling the style word pattern of the text. The method identifies an author’s individualistic form of expression of emotional aspects, sociolinguistic aspects and word co-occurrences, to obtain an author-style pattern for each candidate author. An author-specific model is generated. The questioned document is fed into the different models so formed, and the final decision regarding the authorship is made based on the ensembled learning method. The experimental results of the proposed method has secured a precision of 0.98 in best case and 0.45 in worst case, thereby illustrating an improvement in the accuracy of authorship attribution of short texts, in comparison with the existing methods.

A novel Non-Linear Fractional order PID controller (NLF-PID) is designed for control of coupled and non-linear 2-link rigid robot. The structure of proposed controller comprises of non-linear hyperbolic function of instantaneous error and current state cascaded with Fractional Order PID (FO-PID). Non-linear function provides adaptive control ability while incorporation of fractional operator enhances flexibility of designed controller. To examine the comparative merits of NLF-PID controller, Non-linear PID (NL-PID), FO-PID and traditional PID schemes are also implemented. Design variables of controllers are optimally tuned using multi objective Non-dominated Sorting Genetic Algorithm II (NSGA-II) for small variation in control and error signal. Results prove that NLF-PID provides robust and efficient control of robotic arm as compared to other designed controllers for reference tracking, model uncertainty, disturbance and noise due to inherent shortcoming of sensor.

Markov game based controllers are robust but lack guarantee on the stability of the designed controller. In this work, we attempt to address this shortcoming by proposing a lyapunov fuzzy Markov game controller for safe and stable tracking control of two link robotic manipulators. Lyapunov theory has been used to generate fuzzy linguistic rules for implementing a reinforcement learning (RL) based Markov game controller. We employ fuzzy inference system as a generic function approximator to deal with the “curse of dimensionality” issue. Proposed RL based Markov game controller is self-learning, adaptive and optimal. We implement the proposed control paradigm on: a) Two link robot manipulator and b) SCARA manipulator for the cases: i) controller handles disturbances and parameter variations, and ii) disturbances and no parameter variations. We give comparative evaluation of our approach against: a) fuzzy Q learning controller, and b) fuzzy Markov game controller. Simulation results illustrate stable and superior tracking performance and advantage in terms of lower control torque requirements.

Engineering systems are governed by set of differential equations and work under various uncertainties of parameters. Type–1 fuzzy logic controller are widely used in engineering systems for expert control of system using logic and mathematics together. This paper proposes the first-ever use of Interval Type–2 fuzzy logic controller design to control chaos and associated instability in a nonlinear dynamical power system. Interval Type–2 fuzzy designs have an edge over type-1 fuzzy sets and interval type–2 fuzzy logic controller is well suited for the uncertainty present in weak and chaos sensitive systems. Uncertainty in parameters affects differential equation very badly for such systems. Uncertainty in engineering applications needs an extra layer of handling system control mathematically and logically. Comparison of Type–1 and Type–2 fuzzy logic controller based on time domain waveform, phase plane trajectory and integral square error, clearly proves the efficacy of Type–2 fuzzy logic controller in controlling dynamic behavior of a nonlinear dynamical system as demonstrated through results discussed in the paper. This paper discusses control of chaos driven voltage instability issue in the case of nonlinear dynamical power system, as an application.

This article presents a scalable and optimized recommender system for e-commerce web sites to maintain a better customer relationship management and survive among its competitors. The proposed system analyses the clickstream data obtained from an ecommerce site and predicts the preference level of the customer for the products clicked but not purchased using efficient classifiers such as decision trees, artificial neural networks and extended trees. Collaborative filtering technique is used to recommend products in which similarity measures are used along with efficient rough set leader clustering algorithm which helps in making accurate and fast recommendations. To determine the effectiveness of the proposed approach, an experimental evaluation has been done which clearly depicts the better performance of the system as compared to conventional approaches.

The microblogging service Twitter has witnessed a rapid increase in its adopters ever since it’s discovery in October 2006. Today it has become a medium of communication as well as spread of information. Hashtags are created in twitter by users whenever an event of significant importance occurs and hence they become trending on twitter network. Once hashtags are created on twitter platform, the tweeters may communicate within a particular community of interest following and tweeting to any particular hashtag conversations. In this paper, we propose the design of Community based Hashtag Recommender System (CHRS) for twitter users. This will help the users by expanding their hashtag base and hence strengthening the hashtag conversation for a particular event. The tweets collected over a period of time for some particular hashtags have been categorised to communities based on sentiment analysis of the tweets. Once the process of community detection completes, the existing users are found in the tweets. Further the idea of Hashtag frequency- Inverse Community Frequency (HF-ICF) has been suggested and deployed to find hashtags which uniquely distinguish the users found earlier. Finally relevance score is computed based on the idea of collaborative filtering approach to recommendation, for various hashtags used by the users. A prototype of the system is developed using the statistical tool R and experimental analysis has been carried out. Tweets of national concern in India pertaining to ‘demonetisation’ have been collected and used for experimental purposes.

Recommender systems (RS) suffer from cold start and data sparsity problem. Researchers have proposed various solutions to this problem in which cross domain recommendation is an effective approach. Cross domain recommender system (CDRS) utilizes user data from multiple domains to generate prediction for the target user. This paper proposes a proactive cross domain recommender system. This paper also introduces a parallel approach in cross domain recommendation using general purpose graphic processing unit (GPGPU). This will help to accelerate the computation in the multi-agent environment as data processing in multiple domains takes significant amount of time. A prototype of the system is developed in tourism domain using Cuda, JCuda, Java, Android studio and Jade. The system uses four domains which is restaurant, tourist places, shopping places and hotels. The performance of the parallel CDRS system is compared with non-parallel CDRS in terms of their processing speed. Also the system is compared to the normal Collaborative Filtering approach to measure accuracy of the proposed system using MAE as well as precision, recall and F-measure. The results show a significant speedup for the presented system over non-parallel system.

Aktas and Cagman propounded soft group in 2007 and Abdulkadir Aygunoglu and Halis Aygun defined fuzzy soft group in 2009. In 2016, J. Vimala and J. Arockia Reeta proposed lattice ordered fuzzy soft group and derived its pertinent properties. In this work, fuzzy soft cardinality and fuzzy soft relative cardinality are promoted in lattice ordered fuzzy soft group. Then we give decision making method that can be applied successfully for solving many problems with uncertainties.

Credit scoring is a procedure to estimate the risk related with credit products which is calculated using applicants’ credentials and applicants’ historical data. However, the data may have some redundant and irrelevant information and features, which lead to lower accuracy on the credit scoring model. So, by eliminating the redundant features can resolve the problem of credit scoring dataset. In this work, we have proposed a hybrid credit scoring model based on dimensionality reduction by Neighborhood Rough Set (NRS) algorithm and layered ensemble classification with weighted voting approach to improve the classification performance. For classifiers’ raking, we have proposed a novel classifier ranking algorithm as an underlying model for representing ranks of the classifiers based on classifier accuracy. It is used on seven heterogeneous classifiers for finding the ranks of those classifiers. Further five best ranking classifiers are used as base classifier in layered ensemble framework. Results of the ensemble frameworks (Majority Voting (MV), Weighted Voting (WV), Layered Majority Voting (LMV), Layered Weighted Voting (LWV)) with all features and after feature reduction by various existing feature selection algorithms are compared in terms of accuracy, sensitivity, specificity and G-measure. Further, results of ensemble frameworks with NRS are also compared in terms of ROC curve analysis. The experimental outcomes reveal the success of proposed methods in two benchmarked credit scoring (Australian credit scoring and German loan approval) datasets obtained from UCI repository.

Neural machine translation is an approach to learn automatic translation using a large, single neural network. It models the whole translation process in an end-to-end manner without requiring any additional components as in statistical machine translation systems. Neural machine translation has achieved promising translation performances. It has become the conventional approach in machine translation research nowadays. In this work, we applied neural machine translation for English-Punjabi language pair. In particular, attention based mechanism was used for developing the machine translation system. We also developed the parallel corpus for English-Punjabi language pair. As of now, we are releasing version-1 of the corpus and it is freely available for any non-commercial research. To the best of author’s knowledge, there is no relevant literature on neural/statistical machine translation implementation for English-Punjabi language pair as of this writing. To evaluate the system, BLEU evaluation metric was used. To quantify system’s performance, the results obtained were further compared with existing systems such as AnglaMT and Google Translate. The BLEU score of the developed system exceeds both of these systems marginally.

Cloud computing offers utility-based IT services on-demand to the users on a pay-per-use-basis. The cloud centers consist of physical machines (PMs) with virtual machines (VMs). These data centers consume a large amount of energy due to the improper resource utilization and lack of efficient scheduling algorithms to perform the task-resource mapping. These issues lead to huge energy consumption along with high maintenance costs and carbon emissions. In this paper, a Power Efficient Scheduling and VM Consolidation (PESVMC) algorithm is proposed to address these issues and the associated challenges. The numerous existing research works concentrated on the application of energy management techniques to hardware level support for the reduction of energy consumption. The proposed algorithm emphasizes on the software level by taking the flexibility of the virtualization technology and it consists of two phases, VM Scheduling phase, and VM Consolidation phase. In the Scheduling phase, the tasks with maximum runtime are allocated to VM, which is expected to consume minimal energy. In the VM Consolidation phase, overloaded and underloaded hosts are determined based on the double-threshold scheme. Further, Live Migration technique is applied for migrating the VMs from over-utilized or underutilized hosts to other hosts with the normal utilization. A power efficient utilization factor is introduced to determine the underloaded hosts. This utilization factor is proven to reduce the number of migrations, which can cause additional energy consumption. Energy efficient Scheduling combined with VM Consolidation is successful in maximizing the resource utilization and minimizing the energy consumption. The experimental evaluation is performed using WorkflowSim and the proposed algorithm achieves significant energy conservation and resource utilization.

This paper presents, a novel nature-inspired optimization paradigm, named as owl search algorithm (OSA) for solving global optimization problems. The OSA is a population based technique based on the hunting mechanism of the owls in dark. The proposed method is validated on commonly used benchmark problems in the field of optimization. The results obtained by OSA are compared with the results of six state-of-the-art optimization algorithms. Simulation results reveal that OSA provides promising results as compared to the existing optimization algorithms. Moreover, to show the efficacy of the proposed OSA, it is used to design two degree of freedom PI (OSA-2PI) controller for temperature control of a real-time heat flow experiment (HFE). Experimental results demonstrate that OSA-2PI controller is more precise for temperature control of HFE in comparison to the conventional PI controller.

In the recent era, evolutionary meta-heuristic algorithms is popular research area in engineering and scientific field. One of the intelligent evolutionary meta-heuristic algorithms is Teaching Learning Based Optimization (TLBO). The basic TLBO algorithm follows the isolated learning strategy for the whole population. This invariable learning strategy may cause the misconception of knowledge for a specific learner, which makes it unable to deal with different complex situations. For solving the complex non-linear optimization problems, local optimum frequently happens in the generating process. To resolve these kinds of problem, this paper introduces Neighbour based TLBO (NTLBO) and differential mutation. The concept of neighbour learning and differential mutation is introduced to improve the convergence solution after each run of experiment. Neighbour learning method maintains the explorative and exploitation search of the population and discourages the premature convergence. The efficiency of the proposed algorithm is evaluated on eight benchmark functions of Congress on Evolutionary Computation (CEC) 2006. The proposed NTLBO present extensive comparative study with the state-of-the-art forms of the meta-heuristic algorithms for standard benchmark functions. The result shows that the proposed NTLBO gives the superior performance over recent meta-heuristic algorithms.

Because of its increasing usage, internet has become an integral component of our daily lives. In this paradigm, users can share their perceptions and collaborate with others easily through social communities. The e-healthcare community service is particularly recommended by individual patients who are remotely located, have embarrassing medical conditions, or have caretaker responsibilities that may prohibit them from obtaining satisfactory face-to-face medical and emotional support. However, participation in such online social collaborations may be constrained due to cultural and language barriers. This paper proposes a privacy-preserving collaborative e-healthcare system that connects and integrates patients or caretakers into different groups. This system helps them to chat with other patients with similar problems, understand their feelings, and much more. However, patients’ private and sensitive information cannot be disclosed to anyone at any point of time. The recommended model uses a special technique, k-centroid multi-view point similarity algorithm, to cluster e-profiles based on their similarities. Finally, a distributed hashing technique is used to encrypt the clustered profiles to persevere patients’ personal information. The suggested framework is compared with well-known privacy-preserving clustering algorithms to compute accuracy and latency by using popular similarity measures.

The diagnostic prediction models in medical sciences are more relevant today than ever before. The nature and type of the data do have a profound impact on the prediction output. As the nature of data changes, the choice of intelligent methods also has to be altered adaptively to attain promising results. A highly customised data oriented model which encompasses multi-dimensional information can aid and improve the prediction process. This paper proposes an adaptive soft set based intelligent system which is designed to receive a set of input parameters related to any disease and generates the risk percentage of the patient. The system produces soft sets with the given inputs by fuzzification; followed by rule generation. The rules are analysed to obtain the risk percentage and based on its intensity, the system proceeds with the disease diagnosis. Four different approaches are introduced in this study to enhance the risk prediction accuracy, namely subset of parameters method, adaptive selection of analysis metrics, weighted rules method and the unique set method. The best model is acquired from these approaches in an adaptive fashion by the algorithm. Our method of risk prediction is applied for prostate cancer detection as a case study and we provide exhaustive comparison of the different approaches employed within the algorithm. The results prove that this synergistic approach gives better prediction results than the existing methods. The combination of unique set and weighted approach gave the best predictive solution for the proposed system.

High-dimensional data analysis is quite inevitable due to emerging technologies in various domains such as finance, healthcare, genomics and signal processing. Though data sets generated in these domains are high-dimensional, intrinsic dimensions that provide meaningful information are often much smaller. Conventionally, unsupervised clustering methods known as subspace clustering are utilized for finding clusters in different subspaces of high dimensional data, by identifying relevant features, irrespective of labels associated with each instance. Available label information, if incorporated in clustering algorithm, can bias the algorithm towards solutions more consistent with our knowledge, leading to improved cluster quality. Therefore, an Information Gain based Semi-supervised- subspace Clustering (IGSC) is proposed that identifies a subset of important attributes based on the known label for each data instance. The information about the labels associated with data sets is integrated with the search strategy for subspaces to leverage them into a model based clustering approach. Our experimentation on 13 real world labeled data sets proves the feasibility of IGSC and we validate the clusters obtained, using an improvised Davies Bouldin Index (DBI) for semi-supervised clusters.

Deep Neural Networks (DNNs) have powerful recognition abilities to classify different objects. Although the models of DNNs can reach very high accuracy even beyond human level, they are regarded as black boxes that are absent of interpretability. In the training process of DNNs, abstract features can be automatically extracted from high-dimensional data, such as images. However, the extracted features are usually mapped into a representation space that is not aligned with human knowledge. In some cases, the interpretability is necessary, e.g. medical diagnoses. For the purpose of aligning the representation space with human knowledge, this paper proposes a kind of DNNs, termed as Conceptual Alignment Deep Neural Networks (CADNNs), which can produce interpretable representations in the hidden layers. In CADNNs, some hidden neurons are selected as conceptual neurons to extract the human-formed concepts, while other hidden neurons, called free neurons, can be trained freely. All hidden neurons will contribute to the final classification results. Experiments demonstrate that the CADNNs can keep up with the accuracy of DNNs, even though CADNNs have extra constraints of conceptual neurons. Experiments also reveal that the free neurons could learn some concepts aligned with human knowledge in some cases.

Computing the all pair shortest paths in a graph is a widely used solution, but a time-consuming process too. The popularly used conventional algorithms rely solely on the computing capability of the CPU, but fail to meet the demand of real-time processing and mostly do not scale well for larger data. In this paper, we propose the ex-FTCD (extending Full Transitive Closure with Dijkstra’s) algorithm for finding the all pair shortest path by merging the features of the greedy technique in Dijkstra’s single source shortest path method and the transitive closure property. Experiments show that the process improves computing speed and is more scalable. We re-designed the algorithm for the parallel execution and implemented it in mapreduce on Hadoop that supports the conventional map/reduce jobs. This work also includes the implementation on Spark that supports the in-memory computational capability which uses Random Access Memory for computations. The experiments show that the numbers of iterations are relatively small for even large networks.

Image blurring artifact is the main challenge to any spatial, denoising filters. This artifact is contributed by the heterogeneous intensities within the given neighborhood or window of fixed size. Selection of most similar intensities (G-Neighbors) helps to adapt the window shape which is of edge-aware nature and subsequently reduce this blurring artifact. The paper presents a memristive circuit design to implement this variable pixel G-Neighbor filter. The memristive circuits exhibits parallel processing capabilities (near real-time) and neuromorphic architectures. The proposed design is demonstrated as simulations of both algorithm (MATLAB) and circuit (SPICE). Circuit design is evaluated for various parameters such as processing time, fabrication area used, and power consumption. Denoising performance is demonstrated using image quality metrics such as peak signal-to-noise ratio (PSNR), mean square error (MSE), and structural similarity index measure (SSIM). Combining adaptive filtering method with mean filter resulted in average improvement of MSE to about 65% reduction, increase of PSNR and SSIM to nearly 18% and 12% correspondingly.

EEG is the most effective diagnostic technique to determine epilepsy in a patient. The objective of this research work is to apply classification techniques on EEG signals to determine whether the patient has suffered from epileptic seizure. This is carried out through the extraction of various time and frequency domain features. The two classifiers, i.e. Artificial Neural Network (ANN) and Support Vector Machine (SVM) are used and compared using various evaluation parameters. The simulation results and corresponding quantitative analysis shows that ANN classifier is superior to SVM.

Today’s Image processing tools have matured to a level where its users can effortlessly modify or enhance the images according to their requirement. A misuse of such tools has created a necessity for authenticating images to ensure its correctness. Image Forensics deals with the study of different kinds of manipulation on images and their detection. Image forgery detection algorithms detect forgery related artifacts which can be distinguished using specific image properties. Texture-based features have been widely used to detect forgery induced texture variations in the images. In this paper, we propose Region and Texture combined features for Image Forgery Detection. The Region-based approaches like – Edge-based Region Detection, Saliency-based Region Detection, and Wavelet-based Region Detection are captured, and on these regions, the texture feature- Rotation invariant Co-occurrences among adjacent LBP (RiCoLBP) is applied. The features thus obtained are optimized using Non-Negative Matrix Factorization and fed to a Support Vector Machine (SVM) for classification. The method is extensively evaluated on three benchmark datasets for image forgery detection namely CASIA v1.0, CASIA v2.0 and CUISDE. The performance reveals improved detection accuracies when compared to the state-of-the-art methods in detecting forged and authentic images.

In this paper, a robust image watermarking technique has been proposed in lifting wavelet transform (LWT) domain. Neural network is incorporated in the watermark extraction process to achieve improved robustness against different attacks. The integration of neural network with LWT makes the system robust to various attacks maintaining an adequate level of imperceptibility. The 3-level LWT coefficients are randomized and arranged in 2×2 non-overlapping blocks. Each block is modified according to a binary watermark bit. Randomization of coefficients and blocks has been done to enhance the security of the system. The binary watermark bit is also encrypted using another key. The scheme provides an average imperceptibility of 43.88 dB for a watermark capacity of 512 bits. The robustness has been observed against all the intentional and non-intentional attacks. The technique provides satisfactory robustness against different attacks such as noising attacks, de-noising attacks, lossy compression attacks, image processing attacks and some geometric attacks. The algorithm has been tested on a large image database containing different class of images.

In this paper, a two step approach using Variational Mode Decomposition (VMD) and ℓ₁ trend filter is proposed for enhancing speech signals degraded by white Gaussian noise. In the first step, VMD decomposes the noisy speech signal into Intrinsic Mode Functions (IMFs) corresponding to different frequency components of the signal. In the second step, ℓ₁ trend filter retrieves the speech information by filtering out the noisy sub frames. The noisy sub frames are identified using a threshold based on noise variance in the corresponding IMFs. The proposed work experiments on speech signals degraded by white Gaussian noise in the range 10 dB– 30 dB. The performance of proposed method is compared with some of the well known speech enhancement techniques: Spectral Subtraction (SS) and Minimized Mean Square Error (MMSE) using the subjective and objective quality measures. The proposed, two-step approach of VMD-ℓ₁ trend filter achieves better performance compared to the considered methods.

Optic disc (OD) detection is an important step in a number of algorithms developed for automatic extraction of anatomical structures and retinal lesions. In this article, a novel system, eye gaze– based OD detection, is presented for detecting OD in fundus retinal image using the knowledge developed from the expert’s eye gaze data. The eye gaze data are collected from expert optometrists and non-experts group while viewing the fundus retinal images. The task given to them is to spot the OD in fundus retinal images. Eye gaze fixations were used to identify the target and distractor regions. The image-based features were extracted from the identified regions. The top-down (TD) knowledge is developed using feature ranking and fuzzy system. This TD knowledge is further used for building the TD map. The success rates for various standard datasets are: DRIVE dataset, 100%; DRIONS-DB, 98.2%; INSPIRE, 97.5%; High Resolution Fundus Images, 100%; DIRECTDB0, 96.9%; ONHSD, 91.9% and STARE, 81.4%.

Satellites that orbit the Earth at lower altitudes are the predominant type that are deployed in remote sensing missions. Although thee are many well documented advantages in having the Earth-observing satellites in such lower orbits, the altitude parameter often introduces significant variabilities to the orbital elements, the predominant among them being the perturbative forces due to atmospheric drag. The drag parameter causes deviation of the satellite from its actual orbital trajectory. While updated near-Earth atmospheric drag models have helped resolve this issue, there are the other forces that are secondary in importance, such as the non-spherical Earth’s shape factor effect and luni-solar perturbations that also affect on remote sensing satellites, particularly of the low-Earth Orbit (LEO) and geosynchronous transfer orbiting (GTO) types, that are considered for this particular study. After verifying the equivalent perturbed acceleration terms using the Cowell’s method, variations caused to the classical orbital elements in general, and the altitude element in particular, are characterised in terms of the corresponding distortions caused in the imaging data obtained by the Landsat platform, that are then shown as resulting in a image quality degradation.

Automotive embedded applications are gaining widespread importance worldwide. Almost every car is equipped with hundreds of software-enabled technologies which are expected to have a drastic spike in the coming years ahead. With the advent of more and more automatic and autonomous techniques, there arises an extensive requirement of applying latest development and testing strategies. Apart from using single traditional V-Model, the design and test cases are generated based on more advanced double and triple V-models. The front panel of one of the autonomous car systems is designed and more refinement is obtained for optimum test path generation using prefix graph algorithm and edge-pair coverage criteria.

A quality software development is inclined to the software developer skills. The research focus on recommending the skills of individuals based on the eye movement data. The paper sketches a study conducted on students who are future developers. A fuzzy based recommendation system was implemented to recommend two skills, code coverage and debugging skills that are primary in source code review. The code coverage inference system recommends individual code coverage as maximum, average and minimum and the debugging fuzzy inference system recommends debugging skills as skilled, unskilled and expert.

Determination of change prone classes is crucial in providing guidance to software practitioners for efficient allocation of limited resources and to develop favorable quality software products with optimum costs. Previous literature studies have proposed successful use of design metrics to predict classes which are more prone to change in an Object-Oriented (OO) software. However, the use of evolution-based metrics suite, which quantifies history of changes in a software, release by release should also be evaluated for effective prediction of change prone classes. Evolution-based metrics are representative of evolution characteristics of a class over all its previous releases and are important in order to understand progression and change-prone nature of a class. This study evaluates the use of evolution-based metrics when used in conjunction with OO metrics for prediction of classes which are change prone in nature. In order to empirically validate the results, the study uses two application packages of the Android software namely Contacts and Gallery2. The results indicate that evolution based metrics when used in conjunction with OO metrics are the best predictors of change prone classes. Furthermore, the study statistically evaluates the superiority of this combined metric suite for change proneness prediction.

In this paper, we study the convergence of fuzzy tensors. The notions of directed paths and fuzzy directed path systems are introduced. Their convergent theorems are proposed by using cut matrices principles. Also, two algorithms for identifying the convergence of fuzzy tensors are presented. Finally, numerical examples are proposed to show the feasibility of the presented methods.

On the basis of α-minimal resolution principle, an α-n(t)-ary resolution dynamic automated reasoning method—multi-ary α-ordered linear minimal resolution method is studied in lattice-valued propositional logic system LP(X) and lattice-valued first-order logic system LF(X) based on lattice implication algebra (LIA). Firstly, multi-ary α-ordered linear minimal resolution method is established in LP(X), while its theorems of both soundness and completeness are proved. Then, multi-ary α-ordered linear minimal resolution method is further established in the corresponding lattice-valued first-order logic LF(X), along with its soundness theorem, lifting lemma, and completeness theorem. Then, the validity of multi-ary α-ordered linear minimal resolution based on lattice-valued logic is analyzed. At last, an multi-ary α-ordered linear minimal resolution algorithm in LP(X) is designed, and it is proved to be sound and complete, then it is further extended in the corresponding LF(X). This lays the foundation for the further study on α-n(t)-ary resolution dynamic automated reasoning program.

Unmanned Aerial Vehicle (UAV) is an aircraft with no onboard pilot. UAV can be remotely controlled using pre-programmed flight plans. Automatic trajectory planners are an integral and critical component of advanced mission planning systems. The goal of the trajectory planning is to compute an optimal or near-optimal path satisfying UAV dynamic constraints and safety considerations. Information such as terrain data, threat information and UAV dynamic constraints should be provided beforehand and based on this information, the best trajectory should be proposed. In this paper, the 3D trajectory planning problem for a Terrain Following, Terrain Avoidance and Threat Avoidance (TF/TA/TA) mission is considered and a fuzzy logic system is proposed to find a near optimal TF/TA/TA flight path. Then the planned trajectory is compared with the trajectories obtained from Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) algorithm.

In this paper, a new version of distance measure for hesitant fuzzy linguistic term sets is developed. The new version of distance measure provides an objective way to handle the diverse dimensions of two HFLTSs, i.e., without shortening the HFLTS to be a linguistic interval or extending the shorter one by adding linguistic terms. The Hamming distance, the Euclidean distance and the corresponding forms are put forward. By the relationship between the distance measure and the similarity measure, some parallel similarity measures are also investigated. As applications of the proposed measures, the issues of pattern recognition, clustering analysis and multi-criteria decision making are considered.

Elliott Wave Theory has the advantages of universality and accuracy. It accurately depicts the way the stock market works and has become an important tool in securities market modeling. Elliott Wave Theory includes five rising waves and three descending waves, which have important quantitative features related to the Fibonacci series and the golden ratio. At the same time, China’s Shanghai Composite Index reflects the systemic risk of the stock market to a certain extent. If China’s Shanghai Composite Index could be accurately predicted, we could take the necessary precautions to prevent risk in the system. Therefore, this paper uses gray model features, which are highly adaptable and can handle parameter changes. The cross-sectional data of the inflection points of the Elliott wave line are selected as the original data to fit the small sample number required for gray modeling. According to the special mapping relationship between Elliott Wave Theory and the Fibonacci sequence, by combining the important properties of the Fibonacci sequence and the golden ratio, the background value of the gray GM (1, 1) model is optimized, the new models F₁ - GM and F₂ - GM are presented, and the important properties of the new models are studied. Finally, the Elliott wave line is drawn using the Chinese Shanghai Composite Index, and the inflection point data for the whole wave, rising waves and falling waves of the wave line are used as empirical data. The results show that the new model can choose an optimal model according to the data characteristics and is more effective. The new model can also provide new information for the forecasting of stock price indexes and provide help and reference for stock price index forecasts.

This paper develops the mathematical framework and the solution of a system of type-1 and type-2 fuzzy fractional and arbitrary order dynamical systems. The theory of arbitrary order differential equation is developed with fuzzy initial values, fuzzy boundary values and fuzzy parameters. There are numerous natural phenomena which can be modelled as arbitrary order dynamical systems whose initial conditions and/or parameters may be imprecise in nature. The imprecision of initial values and/or parameters are generally modelled by fuzzy sets. Here, the concept arbitrary order dynamical systems is developed with the introduction of fuzzy fractional derivatives, fuzzy fractional partial derivatives, fuzzy stochastic process, fuzzy stochastic random variable and fuzzy Brownian motion. The generalized L ^p -integrability, based on the extension of the class of differentiable and integrable fuzzy functions is applied and fuzzy arbitrary order differential equations are represented as fuzzy integral equations. A novel numerical scheme for simulations of numerical solution for fuzzy arbitrary order differential equations have also been developed. Some illustrative examples have been provided for different fuzzy type-1 and type-2 arbitrary order general dynamical systems models related to mathematical finance and problems in fluid flow.

In this paper, bipolar fuzzy planar graph is defined and studied several properties. The bipolar fuzzy planar graph is defined in a very interesting way. The parameter “degree of planarity” measures the planarity of a bipolar fuzzy graph. The other relevant terms such as strong edges, bipolar fuzzy faces, strong bipolar fuzzy faces are defined here. A very close association of bipolar fuzzy planar graph is bipolar fuzzy dual graph. This is also defined and several properties of it are studied. Bipolar fuzzy planar graph and bipolar fuzzy dual graph have many applications in different fields including design of social network, design of subway tunnel or routes, gas or oil pipelines, image segmentation, etc.

In this paper, an efficient and polynomial edge detection algorithm based on a hierarchical graph-partition approach is presented. After transforming a digital image into a graph network, the proposed algorithm proceeds by iteratively dividing the image into regions, and then transforming this hierarchical region map into a sequence of boundary maps. This allows the proposed algorithm to operate naturally with colour or hyperspectral images, as well as to detect edges at different levels of detail in a simultaneous and consistent manner. Such a sequence of edge maps can be seen as jointly approximating the different levels of detail that humans may use when recognizing objects in an image. This idea is taken to base the evaluation methodology of the proposed algorithm, that extends the usual boundary-based evaluation methodology based on the matching of the automatic maps with a set of human ground truth, reference maps. The computational experiences carried out to benchmark the performance of the proposed algorithm over the BSDS500 dataset suggest that the proposed method attains a statistically significant better performance than some well-known detectors as Canny or Sobel.

Obtaining interesting and topic-relevant information is a very important task in Web mining. Text classification using a small proportion of labeled data and a large proportion of unlabeled data, also called semi-supervised learning, is a well-known problem. Despite plenty of research on text classification, however, how to effectively and efficiently apply valuable frequent patterns and deal with high-dimensional data in text classification is still an open issue. Due to the increasing data volumes and plenty of high-dimensional data, both distance measures and time complexity could be influenced by the noisy data. This paper targets on this problem and presents a novel method for text classification called CTFP (Classification based on TFP-tree), which uses TFP-tree (Text-Frequent-Pattern-tree) to generate frequent patterns in tremendous amount of texts and conduct text classification in a relatively low dimensional data space. It effectively reduces the data dimensionality during constructing the classifier. Substantial experiments on three datasets (RCV1, SRAA and Reuters-21578) show that our proposed method can achieve better performance than many existing state-of-the-art methods on precision, efficiency and many other evaluation metrics.

On a hypergroup, one can define a topology such that the hyperoperation is pseudocontinuous. The purpose of this paper is to study examples of topological hypergroupoids. We show that there is no relation (in general) between pseudotopological and strongly pseudotopological hypergroupoids. In particular, we present a topological hypergroupoid that does not depend on the pseudocontinuity nor on strongly pseudocontinuity of the hyperoperation.

Uncertain set is a set-valued function on an uncertainty space, and used to model unsharp concepts. In uncertain set theory, the distance between uncertain sets has been proposed. However, the existing distance measure formula is complicated and calculated by supremum and integral. Based on the membership function and inverse membership function, this paper proposes an easier way to calculate the distance between uncertain sets. At the same time, the distance measure formulas of two types of uncertain sets are also derived. Especially, the formula of triangular uncertain sets is expressed by some parameters. At last, the paper discusses the application in classification via the proposed distance measure formulas.

In this paper, the embedding of m-polar fuzzy graphs which constructed on the surface of spheres is introduced. The m-polar fuzzy genus graphs with its genus value, strong and weak m-polar fuzzy genus graph are defined. Also isomorphism properties on m-polar fuzzy genus graph are discussed. The relation between planarity value and genus value of m-polar fuzzy graph is established. Euler polyhedral equation is established in terms of genus value of the m-polar fuzzy genus graph. Finally, an application of m-polar fuzzy genus graph is given in topological surface.

In this study, we aim to introduce the notion of a parameterized connectedness degree of an L-fuzzy soft set in L-fuzzy (E, K)-soft topological spaces. We first define the notion of a parameterized separatedness degree of L-fuzzy soft sets in L-fuzzy (E, K)-soft topological spaces in terms of closure operators. Then we introduce the notion of a parameterized connectedness degree of L-fuzzy soft sets and generalize some connectedness properties, well known in general topology, to L-fuzzy (E, K)-soft topological spaces.

Large amounts of data are generated by the intelligent transportation system (ITS) everyday. It exceeds the storage and processing capacity of conventional systems, and also doesn’t fit the structures of current database. Therefore, it is necessary to use efficient methodology addressing the challenges. Vehicle logo recognition (VLR) is a significant application in ITS. VLR is difficult due to the geometric distortions as well as various imaging situations simultaneously. However, traditional methods and hand-crafted features have many limitations. Convolutional neural network (CNN) enjoys the success in many machine vision tasks. Inspired by the excellent performance of CNN, we design and develop a novel VLR distributed system framework based on Hadoop ecosystem and deeplearning. We propose a Mapreduce based CNN called MRCNN to train the networks, which significantly increases the training speed and reduces the computation cost simultaneously. Furthermore, unlike previous classical CNN starting from a random initialization, we propose a novel genetic algorithm (GA) global optimization and Bayesian regularization approach called GABR in order to initialize the weights of classifier, which help prevent the overfitting and avoid the local optima. Compared with other algorithms, the proposed method performs best and increases the recognition accuracy with good initial weights optimized by GABR. The results show that the distributed system framework and proposed algorithms are suitable for real-world applications of VLR.

In this paper, a robust linear programming is considered, where all of its coefficients in the objective function and constraints are rough intervals or IT2 rough interval coefficients. First, we allow the IT2 rough intervals to transform into rough intervals using [α₁,  α₂] level. Then, a robust two-step solution method (RTSM) is developed to solve the robust linear programming problem with IT2 rough interval coefficients (LPIT2RIC). Finally, an example is presented to demonstrate the results.

This article discusses a vendor-buyer inventory model with permissible delay in payments and controllable lead time in which the order quantity, lead time and the number of shipments delivered from vendor to the buyer in a production cycle are decision variables. Here the production process is imperfect. The lead time is crashed and the crashing cost is an exponential function of lead time. Based on the lead time demand, two models are developed, that is, lead time demand follows a normal distribution in the first model and then distribution free approach is considered in the second model to minimize the joint total expected cost per unit time. Efficient computational algorithm is designed to find the optimal solution. Numerical examples are provided to illustrate the results obtained. Sensitivity analysis is carried out to study the changes in the effect on optimal solution and some managerial phenomena are obtained through sensitivity analysis.

Shannon entropy is an important tool in the study of information theory. A new fuzzy entropy was introduced by Dumitrescu [5, 6] in the context of fuzzy sets or messages. From the perspective of the fuzzy version, this paper defines the conditional fuzzy entropy of a sequence of fuzzy process, and studies the commutativity behavior of sequential fuzzy entropy in the fuzzy process. Moreover, the commutativity behavior of functions and invariant measures is discussed in those different fuzzy processes. However, our demonstration needs more techniques of topological dynamics and fuzzy analysis.

The main objective of this paper is to present a novel decision making algorithm using matrix representation of the inverse soft set defined in [10]. Therefore, we first introduce cardinality inverse soft matrix theory and its operations, products and algebraic structures in detail. Afterwards, an algorithmic solution employing the cardinality inverse soft matrix to find the optimum object and the ranking order of objects is proposed. The performance of algorithm named soft sum-row decision making algorithm is demonstrated by solving various decision problems. Also, we compare it with existing algorithms based on the soft set theory, soft matrix theory and inverse soft set theory. Moreover, we give Scilab codes of the algorithm and argue that this codes make the process of decision making composed of many objects, criteria and decision makers faster and easier.

Three-way decisions model proposed by Yao gives a semantic interpretation of positive region, negative region and boundary region. This model was developed in the framework of classical rough set, the approached target concept is a crisp set, the employed knowledge is a equivalence relation. In this paper, we extend the three-way decisions model to rough fuzzy set. Specifically, the target concept is extended to a fuzzy set, while the used knowledge is also a equivalence relation. An example is given to illustrate the computation processes of the proposed three-way decisions model. The extended model can deal with the problems described by fuzzy decision tables with symbolic-valued conditional attributes and fuzzy decision attributes.