Dissertation > Excellent graduate degree dissertation topics show

A Study of Algorithms for Text Categorization Based on Reducing Class and Fuzzy Theory

Author: ZengHongBo
Tutor: YangTianQi
School: Jinan University
Course: Applied Computer Technology
Keywords: Category cutting Fuzzy Theory Text Classification KNN Class skew
CLC: TP391.1
Type: Master's thesis
Year: 2011
Downloads: 37
Quote: 0
Read: Download Dissertation

Abstract


Text classification has become a key technology for processing large amounts of text information , and has become an important research directions in the field of data mining . KNN text classification algorithm is one of the many automatic text classification techniques in performance relative prominence , at the same time , it also has its own inadequacies , the first classification slow KNN classification algorithm to All calculations are deferred to the classification , and every must calculate the test text classification similarity of text with all the training , classification slow ; second , the class skew the processing of the data set is not ideal , is much larger than the number of another class training documents some categories , the number of text , the classification results would tend to text a large number of categories . For these two shortcomings , this paper proposes the corresponding solutions . Proposed a method based on class cutting , improve the speed of KNN classification , the program text classification is divided into multi-step , every step through the rapid classification cropped some categories until the last remaining category , classified as the final results. Second, the fuzzy theory is applied to the KNN algorithm for training text set and set of categories to establish a membership degree matrix , then this matrix to calculate the class attribute of the test text . And a balanced membership matrix factor , thereby reducing the negative impact of the difference in the number of category text classification results . The experiments show that the proposed scheme achieved the desired results , the former can greatly reduce the classification time does not affect the classification results , ; latter KNN classification performance can be significantly improved , especially in dealing with the class - skewed data set The performance improvement is more significant .

Related Dissertations

  1. Research on Dam-Break Probability Analysis Method for Dangerous Reservoir,TV697
  2. Research on Text Classification Based on Biomimetic Pattern Recongnition,TP391.1
  3. Tourism Comments on the Internet’s Semantic Analysis and Usefulness Research,TP391.1
  4. Based on Data Distribution Characteristics of Text Classification,TP391.1
  5. Research on Improved K Neighbor Support Vector Machine Algorithm Faced Text Classification,TP391.1
  6. Research of Image Enhancing of Vein Based on Fuzzy Theory,TP391.41
  7. Research for Event Extraction Method in Specific Domain Based on Tree Conditional Random Field,TP391.1
  8. Online Education News Text Categorization System Design and Implementation,TP391.1
  9. One kind of empirical data on the workload of a software bug fixes Prediction Model,TP311.53
  10. Ontology-based medicine named entity recognition technology research,TP391.1
  11. Research on cross-language text categorization,TP391.1
  12. Based on Fuzzy Clustering of Web users,TP311.13
  13. Classification model based monitoring of e-commerce Prohibited Research and Implementation,TP393.09
  14. Based on semantic analysis of text mining research,TP391.1
  15. Analysis on Stability and Construction Mechanical Behavior of Karst Tunnel Surrounding Rock,U451.2
  16. Planning and Schedling of Production Transferring Projects of International Company,F276.7;F273
  17. Based on the associated technology Chinese Text Classification,TP391.1
  18. Automatic distribution of software defects,TP311.52
  19. Research of Web Text Classification Based on Decision Tree Classification Algorithm,TP391.1
  20. Study on Chinese Text Classification Combined with Ontology,TP391.1
  21. Spam messages based on ensemble learning multi-level classification technology research,TN929.53

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Text Processing
© 2012 www.DissertationTopic.Net  Mobile