Earticle

초록

영어

Due to the complex market environment, it is very difficult to get the accurate predict of the stake only by the data analysis method. This paper uses the text categorization method to predict the trend of the stock. We divide the text categorization method into the following three steps: Text representation, Feature selection and Text Categorization. By comparing several categorization methods including feature selections and feature spaces, etc., the results show that the SVM method with Information Gain and 1000 feature spaces can get the better performance for the predict of the stock with the news.

목차

Abstract
 1. Introduction
 2. Text Representation
  2.1. Related Works
  2.2. Text Preprocessing
  2.3. Feature Representation
 3. Feature Selection Methods
  3.1. IG (Information Gain)
  3.2. CHI Square Statistics
  3.3. Mutual Information (MI)
  3.4. Expected Cross Entropy (ECE)
 4. Text Categorization Methods
  4.1. K Nearest Neighbors (KNN)
  4.2. Support Vector Machine (SVM)
 5. Analysis and Result
  5.1. Text Collections
  5.2. Evaluation Measures
  5.3. Analysis
 5. Conclusion
 Acknowledgements
 References

키워드

저자

• Bozhao Li [ School of Computer and Information Engineering, Beijing Technology and Business University, Beijing, 100048, China ]
• Na Chen [ School of Computer and Information Engineering, Beijing Technology and Business University, Beijing, 100048, China ]
• Jing Wen [ School of Computer and Information Engineering, Beijing Technology and Business University, Beijing, 100048, China ]
• Xuebo Jin [ School of Computer and Information Engineering, Beijing Technology and Business University, Beijing, 100048, China ]
• Yan Shi [ School of Computer and Information Engineering, Beijing Technology and Business University, Beijing, 100048, China ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of u- and e- Service, Science and Technology

• 간기

  격월간

• pISSN

  2005-4246

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605