Improved Method for Pattern Discovery in Text Mining
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2014, Vol 16, Issue 3
Abstract
Abstract: The proposed technique uses two processes, pattern deploying and pattern evolving, to refine the discovered patterns in text documents. The experimental results show that the proposed model outperforms not only other pure data mining-based methods and the concept based model, but also term-based state-of-the-art models, such as BM25 and SVM-based models. PBTM firstly generates pattern based topic representations to model user’s information interests with multiple topics; then PBTM selects quality patterns for estimating the relevance of documents. The proposed approach incorporates the semantic topics from topic modeling and the specificity of the representative patterns. The proposed model has been evaluated by using RCV1 and TREC topics for the task of information filtering. Comparing with the state-of-the-art models, PBTM demonstrates excellent strength on document modeling and relevance ranking.
Authors and Affiliations
Rakhi A. Dixit, , Prof. V. A. Chakrawar
Keywords
Related Articles
- EP ID EP157833
- DOI 10.9790/0661-16361722
- Views 93
- Downloads 0
Study of Cloud Computing
Cloud computing is rising as the next generation platform for computation. May be in future Cloud computing will be the main platform to save the world; this makes people can have everything they need on it. Main...
Opinion Mining Method for Sentiment Analysis
Abstract: We are living in a world full of data. Every passing second, large data is generated by Social Media, ECommerce, Stock Exchange and many other platforms. Now-a-days, microblogging sites are used for manypurpose...
Thorny Issues of Stakeholder Identification and Prioritization in Requirement Engineering Process
Identifying the stakeholder in requirement engineering process is one of the critical issues. It performs a remarkable part for successful project completion. The software project largely depends on several  ...
A Comparison Based Study on Biometrics for Human Recognition
A biometric system provides automatic recognition of an individual based on a unique feature or characteristic possessed by the individual. These biometric characteristic may physiological or behavioral. Un...
Artificial Neural Network For Recognition Of HandwrittenDevanagari Character
Abstract: Information gathering is one of the most important thing in the digital world, manyorganizationsstill use handwritten documents for storing their information. So it is important task to convert this infor...