Dissertation > Excellent graduate degree dissertation topics show

Spam messages based on ensemble learning multi-level classification technology research

Author: LeiYang
Tutor: FuYan
School: University of Electronic Science and Technology
Course: Computer Software and Theory
Keywords: SMS Spam Integrated Learning Text Classification
CLC: TN929.53
Type: Master's thesis
Year: 2011
Downloads: 83
Quote: 0
Read: Download Dissertation

Abstract


In recent years, mobile phone text messages for its cheapness and convenience of the characteristics increasingly become a favorite contact, SMS application from the initial communication tool between people gradually extended to a service mode of delivery, such as flight information, weather forecasts subscriptions. While the mobile phone short message service brought explosive growth great convenience to people's lives, but at the same time, followed by a lot of junk SMS also gradually on people's lives have a great negative impact. Spam messages only interfere with people's normal life, while giving a negative impact on social stability. In this paper, spam filtering techniques for the study, first introduced the traditional spam filtering technology, the basic principles, such as black and white list method, keyword filtering method and content-based identification of machine learning method, as well as their advantages and disadvantages, and the focus compare a variety of machine learning algorithms for spam classification ability. Secondly, this paper introduces SMS classification aimed at improving the stability and accuracy of the various multi-level classification ensemble learning methods, and with Stacking as our SMS filtering system integration learning algorithm, experiments two kinds of integration strategies, we finally get a more effective message classification systems. The main work is reflected in the following aspects: (1) experimentally analyzed and compared various machine learning algorithms for spam messages as well as their advantages and disadvantages classification ability. (2) presents a classifier training strategy by merging those confusing categories of data as a training set for the first level classifier, and then these confusing categories of data extracted from a training set as another the second stage classifier, using two classification strategies help to improve classification performance. (3) proposed a mutual information based feature selection algorithm, after using the improved algorithm to achieve good results. (4) using Stacking integrated learning technology integrates multiple base classifiers and proposes two integration strategies, in a real data set with respect to the individual classifiers to improve the classification accuracy. (5) the actual implementation of a multi-level classification of spam messages prototype system. (6) presents the future direction of the system's many improvements.

Related Dissertations

  1. Research on Text Classification Based on Biomimetic Pattern Recongnition,TP391.1
  2. Tourism Comments on the Internet’s Semantic Analysis and Usefulness Research,TP391.1
  3. Based on Data Distribution Characteristics of Text Classification,TP391.1
  4. Research on Improved K Neighbor Support Vector Machine Algorithm Faced Text Classification,TP391.1
  5. Research for Event Extraction Method in Specific Domain Based on Tree Conditional Random Field,TP391.1
  6. Online Education News Text Categorization System Design and Implementation,TP391.1
  7. One kind of empirical data on the workload of a software bug fixes Prediction Model,TP311.53
  8. Study on Presentation and Implementation Strategy about Chinese Comprehensive Learning in Junior Middle School,G633.3
  9. Research on cross-language text categorization,TP391.1
  10. Classification model based monitoring of e-commerce Prohibited Research and Implementation,TP393.09
  11. Based on semantic analysis of text mining research,TP391.1
  12. Research of Web Text Classification Based on Decision Tree Classification Algorithm,TP391.1
  13. A Feature Extraction Method Using Base Phrase and Keyword,TP391.1
  14. Design and Implement of the SMS Spam Monitoring System of ZheJiang Telecom,TP277
  15. Android-based spam messages processing system research and design,TP391.1
  16. Fault Recognition Method Based on Neighborhood Structure Analysis of Feature Space,TH165.3
  17. Gender -based classification of text mining,TP391.1
  18. The Research and Implementation of an Semantic-based Automatic Chinese Text Categorization System,TP391.1
  19. On the cell phone spam messages a civil governance,D923
  20. Research on the Design Features and Teaching Strategy in Chinese Integrative Learning,G633.3

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Wireless communications > Mobile Communications > Cellular mobile communications systems (mobile phones, mobile phone handsets )
© 2012 www.DissertationTopic.Net  Mobile