Dissertation > Excellent graduate degree dissertation topics show

Research on Content-based Spam Filtering

Author: ZhouZuo
Tutor: ZhuXiaoDong
School: Jilin University
Course: Software Engineering
Keywords: Spam filtering Content identification Bayesian algorithm
CLC: TP393.098
Type: Master's thesis
Year: 2011
Downloads: 92
Quote: 0
Read: Download Dissertation

Abstract


The rapid development of Internet, e-mail has become a major tool for people to pass information. But in the e-mail the convenience of its rapidity and simplicity to people at the same time, it also brings to highlight the growing \Spammers use spam to spread reactionary, pornographic, violent content and commercials contain a lot of useless information. It will not only delay the user's valuable time, more potential threats to be interested in the stability of the whole society. At present, domestic and spam prevention and control of existing research, including content-based filtering technology. This paper lists the mainstream anti-spam technology, and to study and compare the advantages and disadvantages of various technologies. Which utilize the principles of probability and statistics, Bayesian filtering technology is widely used, it can be calculated according to the e-mail priori probability and the conditional probability category of mail to be measured with better performance and stability, achieve better filter effect, the Bayesian spam filtering is the focus of this study. Spam filtering for Chinese e-mail text preprocessing is still blank. The text preprocessing technology in text classification is introduced into the spam filter, and improved pretreatment method for the unique characteristics of spam. Spam filtering there are certain differences with ordinary text classification, major differences: First, the spam filtering is a binary classification problem, the only distinction between the two kinds of spam and normal mail, not a multi-classification problem. Second, the spammers will deliberately creating syntax, syntax error avoidance mail filtering, text classification and no problem. Solve these problems, a new Chinese text preprocessing module, it can effectively remove stop words, replace mail deformation word, to be refined segmentation results, such as the proposed POS tagging synonym substitution and so on. Provide excellent and efficient feature set as the next feature selection, and lay a good foundation for the Chinese spam filtering. Bayesian algorithm, in-depth research to introduce the principle of Bayesian techniques, the core idea of ??the Naive Bayes algorithm and two derivations of the model, and multivariate Bernoulli event model and polynomial event model carried out a detailed analysis and comparison. Analyzes the advantages and disadvantages of the Na?ve Bayes algorithm reflects the spam filtering. Two improved methods for its shortcomings, a probability multiples comparison method is to compare the resulting probability values, the introduction of a multiple of a λ relationship, and then determines the category of the message should be. Another classification model is to distinguish the size of the message to take different discriminant. Last mentioned in the paper, based on improved Bayesian spam filtering module of the overall design, respectively from the two aspects of the training process and the process of learning Bayesian filter module to elaborate. This article using CCERT provided Chinese the the mail sample set CDSCE, download mail from CSDN sets and personal messages collected training set and a test set used in the experiments produce. Tested by four groups of different number of mail collection improved method taken to be experimental validation Finally spam filtering effect of commonly used evaluation to validate the experimental results using data confirm the improved Bayesian spam filtering The module has a good filtering effect.

Related Dissertations

  1. The Study on Algorithm of the Minimum Risk Bayes Spam Filtering,TP393.098
  2. Chinese Spam Filtering Based on SVM,TP393.098
  3. Design and Implementation of the Email Spam Detection System Based on Naive Bayes and SVM,TP393.098
  4. The corporate intranet spam monitoring software design,TP393.098
  5. The Application Research on IB Method in Spam Filtering,TP393.098
  6. Research on Technology of Spam Filtering Based on Markov Logic Network,TP393.098
  7. Research and Application of Bayesian algorithms in spam filtering technology,TP393.098
  8. The Research of the Spam Filtering Method Based on the Behavior Identifying,TP393.098
  9. Realization of the Image Spam Filtering System,TP393.098
  10. Research of Spam Filtering System Based on Logistic Regression Model,TP393.098
  11. Classification based on kernel partial least squares spam filtering research,TP393.098
  12. Research on the RFID Indoor Location Approach Based on the Bayesian Learning,TN92
  13. Research and Implementation Ofthe Junk Sms Filtering,TN929.53
  14. A Research of Spam Filtering Based on Text Mining,TP393.098
  15. Design and Implementation of Call Act Analysis Subsystem in Call Monitor Platform,TN915.09
  16. Research and Implementation based the Bayesian algorithm spam filtering system,TP393.098
  17. The Research on Data Mining in the Telephone Harassment Surveillance System,TN916.2
  18. Research on Behavior-based Recognition of Spam Filtering Techniques,TP393.098
  19. Research and Implementation based on minimum risk Bayesian multi-level mail filtering system,TP393.098
  20. Research and Application of Transfer Learn- ing Algorithm for Text Classification Based on Clustering,TP181

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Computer network > General issues > The application of computer network > E-mail ( E -mail )
© 2012 www.DissertationTopic.Net  Mobile