Self Appreciating Concept Based Model For Cross Domain Document Classification
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2014, Vol 16, Issue 3
Abstract
Abstract : In text mining, text categorization is an important technique for classifying the documents. Most of the times statistical approaches that are based on analysis of the term in the form of frequency of the term, that is the number of occurrences of one or more words in the document are used for classification. Even statistical analysis indicates the importance of the term, but it is hard to analyze when multiple terms have the same frequency value, but one term is more important in terms of meaning than the other. Also, there are a wide variety of documents being generated that belongs to different domains which differ in formats, writing styles, etc. These domains can be news articles, e-mails, online chats, blogs, wiki articles, twitter posts, message forums, speech transcripts, etc. Often a classification method that works well in one domain does not work as well in another. The proposed system tries to implement a concept based text classification model that classifies the cross-domain text data based on the semantics or theme of the text data. Also the proposed approach makes the training system stronger and stronger at all possible positive tests of the categorizer. This system is called as a Self Appreciating Concept Based Classifier (SACBC).
Authors and Affiliations
Dipak A. Sutar
Keywords
Related Articles
- EP ID EP152519
- DOI 10.9790/0661-16329095
- Views 86
- Downloads 0
Wireless Controlling Of Remote Electrical Device Using Android Smartphone
Abstract: Controlling the home appliances and electronics gadgets wirelessly through SmartPhone technology can be more easy. That is, through this paper, we target to control a light emitting device (LED) from any...
"Randomized Visual Cryptography scheme for color images”
Abstract: In this paper, we propose a new color visual cryptography scheme which is based on modified visualcryptography. Visual Cryptography is based on cryptography where n images are encoded in a way that onlyth...
A New Modified Fatha Method For Arabic Text Steganography Hybrid With Aes Encryption
Abstract: Growing the size of attacks that listed during the passing of data between the sending and the receiving has really demand to get high robust method for transfer information in secure manner. Both of cryptogra...
The effect of Encryption algorithms Delay on TCP Traffic overdata networks
Abstract: Security is a big concern for data networks users. Data encryption considered to be one of the bestsolutions for security issues. There are some standard encryption algorithms that used to encrypt transfe...
A Multi-Agent System to Support ICT based Distance Learningthrough Modeling of Learner Needs: The Caseof Bachelor of Education at the University of Nairobi
Abstract:In distant learning environments, most learners need feedback from their lecturers and administration. The feedback supports them to achieve their learning objectives. However, the feedback that the most distant...