Information retrieval approach based on indexing text documents: Application to biomedical domain
Kabil Boukhari, Mohamed Nazih Omri · 2017
The use of the controlled vocabulary for documents indexing makes it possible to index documents with other terms not present in the document. The proposed approach, based on Vector Space Model (VSM), allows matching partially document and MESH (Medical Subject Headings) terms. The stemming method applied for the preparation of terms allows favoring the partial matching. In this paper we propose an Information Retrieval Approach based on Indexing Text Documents which combine two methods statistical and semantic one. A measure of similarity is then considered to measure the correspondence between a document and term in MESH thesaurus. To filter the extracted concepts and keep the relevant ones, we exploited the MESH architecture for better indexing. The proposed approach is implemented to evaluate its effectiveness and the experimentation and the results analysis shows that the proposed method gives good results compared to standard measures.