Towards Design of an Integrated System for Bengali Text Document Retrieval, Clustering, Cluster Label Generation and Document Summarization
摘要
In this paper, we present an integrated Bengali information retrieval system with three important modules—text document retrieval in response to a user query, Clustering documents, and summarization. Through our experiments, we have demonstrated that incorporating all three types of approaches into the architecture of the information retrieval system enhances retrievals by giving users a better focus on information arranged according to topics. Given a query in Bengali, our developed system retrieves relevant documents, separates the retrieved documents into topic clusters, and assigns labels to the clusters by extracting keyphrases from each cluster. A summary for each document within a cluster is also generated. Our developed information retrieval system is empowered with document clustering, cluster labeling, and summarization modules facilitating the users to quickly perform an information search on the list of relevant documents returned by the retrieval engine. Our system is made to give you access to comprehensive data in a convenient and concise form. Our method shows that it is feasible to assembling clustering, cluster labeling, and summarization modules to design a more effective IR system.