Retrieving Semantically Relevant Documents using Latent Semantic Indexing

Abstract

Nowadays, with the development of the internet, it is available to collect very large amounts of data and searching effective information from these data develop into an essential work. The major purpose of an Information Retrieval system is to retrieve all the relevant documents, which are relevant to the user query. Popular search engines such as Google, Yahoo, Alta Vista and Bing give the services of the form of modern information retrieval. Term matching techniques may retrieve irrelevant or inaccurate results because of synonyms and polysemys words, so effective concept-based techniques are needed. This system examines the utility of conceptual indexing to improve retrieval performance of a domain specific information retrieval system using Latent Semantic Indexing (LSI). LSI is an indexing and retrieval method that uses a mathematical technique called Singular Value Decomposition (SVD) to figure out patterns in the relationship between the term used and the meaning they convey. LSI makes use of the words that occur together in documents to capture the hidden related meanings among documents and thus can improve the ability to rank relevant documents. This system is able to accept a user query such as a phrase or sentences, search the most semantically related documents and rank and retrieve such documents according to their similarity values. In this system, Cosine Similarity Method is used to find the relevancy and also let the user to view the results by descending order of the similarity values. This system ensures to support the searching time and provide the rate of latent semantic relevancy. The accuracy result of the system is calculated by precision, recall and f-measure. This system introduces to search the symptoms and signs of disease which are collected from https://www.medicinenet.com/ symptoms and signs/symptomchecker.htm#introView. It basically works as a web page search system. The proposed system expected that it helps the people who want to find the information about biomedical diseases.