Home           Contact us           FAQs           
 
   Journal Page   |   Aims & Scope   |   Author Guideline   |   Editorial Board   |   Search
    Abstract
2013 (Vol. 5, Issue: 03)
Article Information:

3-Layered Bayesian Model Using in Text Classification

Chang Jiayu and Hao Yujie
Corresponding Author:  Chang Jiayu 

Key words:  3-layered Bayesian model, coefficient of correlation, degree, matrix of correlation, naive Bayesian, , ,
Vol. 5 , (03): 986-989
Submitted Accepted Published
June 23, 2012 July 28, 2012 January 21, 2013
Abstract:

Naive Bayesian is one of quite effective classification methods in all of the text disaggregated models. Usually, the computed result will be large deviation from normal, with the reason of attribute relevance and so on. This study embarked from the degree of correlation, defined the node’s degree as well as the relations between nodes, proposed a 3-layered Bayesian Model. According to the conditional probability recurrence formula, the theory support of the 3-layered Bayesian Model is obtained. According to the theory analysis and the empirical datum contrast to the Naive Bayesian, the model has better attribute collection and classify. It can be also promoted to the Multi-layer Bayesian Model using in text classification.
Abstract PDF HTML
  Cite this Reference:
Chang Jiayu and Hao Yujie, 2013. 3-Layered Bayesian Model Using in Text Classification.  Research Journal of Applied Sciences, Engineering and Technology, 5(03): 986-989.
    Advertise with us
 
ISSN (Online):  2040-7467
ISSN (Print):   2040-7459
Submit Manuscript
   Current Information
   Sales & Services
   Contact Information
  Executive Managing Editor
  Email: admin@maxwellsci.com
  Publishing Editor
  Email: support@maxwellsci.com
  Account Manager
  Email: faisalm@maxwellsci.com
  Journal Editor
  Email: admin@maxwellsci.com
  Press Department
  Email: press@maxwellsci.com
Home  |  Contact us  |  About us  |  Privacy Policy
Copyright © 2009. MAXWELL Science Publication, a division of MAXWELLl Scientific Organization. All rights reserved