A Novel Approach for Semi Supervised Document Clustering with Constraint Score based Feature Supervision

Apply

 A Novel Approach for Semi Supervised Document Clustering with Constraint Score based Feature Supervision

Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2014, Vol 16, Issue 2

Abstract

Abstract: Text document clustering provides an effective technique to manage a huge amount of retrieval outcome by grouping documents in a small number of meaningful classes. In unsupervised clustering method the unlabeled input data is used to estimate the parameter values. In a semi supervised document clustering both labeled and unlabeled input data is used for document clustering. A semi supervised clustering with feature supervision and constraint score is proposed in this paper. This proposed system which handles document clustering and feature Supervision simultaneously and this system finds the number of clusters automatically. Feature supervision uses pairwise constraints that performs supervision between the each documents. The semi-supervised constraint score that uses both pairwise constraints and the constraint score is to compute relevant features and irrelevant feature on document data set. A variational inference algorithm uses the Dirichlet Process Mixture model for the document clustering.

Authors and Affiliations

S. Princiya, , M. Prabakaran

Keywords

variational inference algorithm semi supervised clustering feature supervision Feature Selection filter method Constraint score Dirichlet Polynomial Allocation (DPA) model Dirichlet Process Mixture Model (DPM

EP ID EP152486
DOI 10.9790/0661-16292834
Views 114
Downloads 0