Particle Swarm Optimization for Web Document Retrieval Based on Term-Document Matrix

Retrieving relevant information from the enormous volume of data is the top most priority of information retrieval systems. As technology arises, various search engines help us to track the relevant data and information on the user’s needs, but the amount of information available is huge and complex. Sometimes, user finds it difficult to find the data according to their need. When people provide the same queries, the same set of relevant documents will be returned by classical search engines. It gets hard for the users to retrieve the relevant documents. To solve this problem, a developed document search technique has been proposed, which is based on Particle Swarm Optimization. It helps to optimize web document information retrieval. PSO algorithm is easy to use and can have better results as it is a faster method compared to other methods. PSO algorithm is being applied on Term-Document Matrix which consists of weights calculated by TF-IDF weighting formula. Term-Document Matrix is the search space on which the PSO is applied by considering the query as a particle and the position of the particles is initialized randomly. This method is tested experimentally on a big TREC 2019 dataset. By applying PSO, the results are compared with the results obtained through traditional approaches.