Award Date

2009

Degree Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science

Advisor 1

Kazem Taghva, Committee Chair

First Committee Member

Ajoy K. Datta

Second Committee Member

Laxmi P. Gewali

Graduate Faculty Representative

Muthukumar Venkatesan

Number of Pages

77

Abstract

Document clustering or unsupervised document classification is an automated process of grouping documents with similar content. A typical technique uses a similarity function to compare documents. In the literature, many similarity functions such as dot product or cosine measures are proposed for the comparison operator.

For the thesis, we evaluate the effects a similarity function may have on clustering. We start by representing a document and a query, both as a vector of high-dimensional space corresponding to the keywords followed by using an appropriate distance measure in k-means to compute similarity between the document vector and the query vector to form clusters. Based on these clusters we decide the best distance metric for the document set used. Next, we compute time complexities for different similarity functions for the same model and document set based on the number of iterations and number of clusters.

Keywords

Canberra distances; Chi-Square; Data mining; Distances; Document clustering; Euclidean distances; Execution time; Information retrieval; K-means clustering algorithm; Similarity functions

Disciplines

Computer Sciences | Databases and Information Systems

Language

English


Share

COinS