Document Clustering: A Review

Abstract

ABSTRACT:As the internet is exploding with huge volume of text documents, the need of grouping similar documents together for versatile applications have hold the attention of researchers in this area. Document clustering can facilitate the tasks of document organization and web browsing, search engine results, corpus summarization, documents classification, information retrieval and filtering. However several attempts have been made to develop efficient document clustering algorithms but most of the clustering methods suffer from challenges in dealing with problems of high dimensionality, scalability, accuracy and meaningful cluster labels. This paper intends to provide a brief summary over methods studied and current state of documents clustering research, including basic traditional methods as well as advanced fuzzy based, GA, PSO, HS oriented techniques etc. Also document representation model and its challenges, dimensionality reduction mechanisms, issues in document clustering, and cluster quality evaluation criteria are discussed.

Document clustering

“…Some detailed document clustering reviews are addressed in Shah and Mahajan (2012) and Premalatha and Natarajan (2010) which consist in describing the general document clustering process and its challenges, focusing mainly on extensions of ‐means applied in the context of document clustering and the conventional hierarchical clustering algorithms. Furthermore, in addition to the aforementioned works, Bisht and Paul (2013) analyze also the frequent itemset based clustering approach which consists of a set of techniques that do not require the vector space model representation of the corpus.…”

Document Clustering based on the Similarity of Data with Efficient Time Consumption

“… Document Representation: Most of the clustering approaches use the vector space model for document representation. The m*n matrix is represented by the collection of n documents with the m unique words where each document is a vector of m dimension [5].  Document Clustering: At this stage, the target documents are grouped into different clusters by selected features [8].…”