Abstract
In this paper, we propose a modified version of the Context-aware Stemming algorithm, itself based on the well-known Porter stemmer in an effort to maximize the proportion of the meaningful stems and thus, the search effectiveness without compromising the other performance measures. Several stemmers are presented and a synergetic hybrid solution is proposed. Indeed, the Semantically Enriched Context-Aware Stemming algorithm combines features from algorithmic stemmers and dictionary stemmers with respect to conceptual indexing techniques in order to improve retrieval performance; proposing root words much comparable to lemma. Moreover, a new query-document mapping technique is proposed based on a previous work and the experimental results conducted with the WT2G dataset show that our algorithm is noticeably more efficient as compared to Porter and CAS algorithms.