An innovative approach to classify and retrieve text documents using feature extraction and Hierarchical clustering based on ontology

摘要

Data retrieval is a key process of acquiring information as per requirement. The necessity of proper information has increased. The most basic tools which provide this service are browser. It traverses the data as per user's query and gives the search results of all related information. Hence, it becomes a time consuming process to find required information. In this paper, the focus is done on content based data mining using ontology and text feature extraction. Content based data mining process focuses on domain of the data. Ontology, itself is a domain based data set information system that will help to achieve required data retrieval in a more appropriate way. The proposed system uses k means clustering algorithm for creation of flat clusters. Flat clusters are the primary classification or clusters of data that are used for Hierarchical clustering. For the proposed system Hierarchical Fuzzy Relational Eigenvector Centrality-based Clustering Algorithm (HFRECCA) is used. This technique of clustering is very fast and gives more accurate results. For more appropriate data retrieval, this system uses text feature extraction algorithm. This algorithm will help to reduce the noisy data from data sets. A noise free data will help to perform better data retrieval process. Implemented system works over various types of text file such as PDF, .txt, DOC, DOCX. This system is also compatible with other types of files like WebPages, images etc.