XPages multi-dbs search engine and semantic search

Abstract

Due to the difficulties encountered by users of SmartShare - management software marketed by EASI - to find their documents within the various databases, a multi-DBs search engine must be incorporated into the software. Today, in the scientific world and in practice, more and more search engines no longer work simply on the superficial representation of words but infer the meaning of these words in order to better respond to queries. These search engines are known as semantic search engines. Therefore, the aim of this project-dissertation is twofold: the development of a multi-DBs search engine for SmartShare and the analysis of semantic extension possibilities. First, a literature review is conducted to provide an overview of existing intelligent search engines, their characteristics and architectures. Then, the characteristics to be provided to the search engine are listed on the basis of a functional analysis of the need carried out beforehand. Finally, a minimum viable product is developed and the business plan is drawn up. Then come the research question: Could the search engine developed for EASI be extended with semantic features? In order to answer this question, another research sub-question arises: Among features of existing semantic search engines, which ones are adapted to the search engine developed for EASI? To answer this sub-question, the different features stated in the literature review are cross-referenced with the characteristics to be provided to the search engine in order to meet the needs of users. Four semantic extensions are then studied in detail: TF-IDF as a sorting method, two mapping algorithms to integrate the meaning of search queries, topic detection, and the extension to multimedia retrieval.