Feature Selection Based on Multiviewpoint And Link Similarity Measure. Document Clustering

      Аннотация: To explore and utilize huge amount of text documents is a major question in the area of information retrieval and text mining. All the methods aiming to find groups of entities utilizes similarity or dissimilarity measure. It is necessary to analyse how similarity measure behave on text documents before developing or modifying a good similarity measure for document clustering to understand the effectiveness of the technique. A similarity function embedded in a criterion function is to a large extent is responsible to analyze the intrinsic structure of the data. If appropriate similarity measures are used with specific clustering technique the efficiency and accuracy of the information discovery task can be enhanced. Use of appropriate measures not only improves the provenance and credit-ability of the retrieved information but also helps to overcome the time and cost complexity of the process. This book focuses on identifying the various similarity measure for Clustering. An imperative method for measuring similarity between text documents is illustrated to cluster the documents using hierarchical clustering and feature selection method using Matlab.

Ключевые слова: Document Clustering, Similarity Measure, Cosine Similarity, Hierarchical Clustering