DATA MINING ON CORPORATE FILLING BASED ON BAYESIAN LEARNING APPROACH

Open Access
Author:
Hu, Xiaocheng
Graduate Program:
Industrial Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
None
Committee Members:
  • Tao Yao, Thesis Advisor
Keywords:
  • data mining
  • text analysis
  • Perl
Abstract:
Most of the researches on corporate filling mainly focus on qualitative analysis. This thesis used quantitative method-Bayesian Learning Machine in analyzing the information content of future prediction statements (FPS) in the Management Discussion and Analysis section of 10-Q fillings. The thesis proposed a new approach that involved a combination of mathematical methods and text analysis. Naïve Bayesian machine learning approach was used to examine the prediction of future performance of the company. In conclusion, the average profit tone of FPS is negatively associated with profit predict and negatively associated with other predict which includes the prediction related to employees, regulations, accounting and other. The liquidity tone is negatively associated with other predict. The overall tone is negatively associated with other predict.