Title page for etd-0708117-200339


[Back to Results | New Search]

URN etd-0708117-200339
Author Ya-ting Jhang
Author's Email Address No Public.
Statistics This thesis had been viewed 5373 times. Download 0 times.
Department Information Management
Year 2016
Semester 2
Degree Master
Type of Document
Language English
Title A forecasting model to predict business performance trend by combining textual information and financial ratios
Date of Defense 2017-07-28
Page Count 62
Keyword
  • Support vector machine
  • Logistic regression
  • Business performance trend
  • Naïve Bayes
  • Forecasting models
  • Abstract The annual report is a complete financial report of the company. It contains textual and financial information such as balance sheet, operating conditions and financial status to help investors have better understanding of the company’s operating status and the future policy. Compared with traditional analysis method based on financial ratios, the textual information derived from annual reports can supply much more immediate and helpful clues related with the company’s operating status and future direction. Therefore, textual contents are very necessary information for investors to make decisions.
    In the previous studies, we found that few researchers employed textual information to predict the business performance trend. Most of them estimated corporate performance only with financial ratios. Therefore, this study combines textual information and financial ratios to predict business performance trend. To analyze the textual information of annual reports, we examine the explanatory contents extracted from annul reports to obtain the text information. The number of variables are reduced by exploratory factors analysis (EFA) into more accurate variables. Afterwards, we adopt Synthetic Minority Over-sampling Technique (SMOTE) to address imbalanced data problem.
    To examine the performance of combing textual information and financial ratios, we apply three classifiers including Naïve Bayes, SVM, logistic regression. According to the results of experiment, the textual information can strengthen the model’s forecasting performance. The investors and shareholders can take this model to support them managing their investment strategies.
    Advisory Committee
  • Yi-Ling Lin - chair
  • Ming-Fu Hsu - co-chair
  • Te-Min Chang - advisor
  • Files
  • etd-0708117-200339.pdf
  • Indicate in-campus at 5 year and off-campus access at 5 year.
    Date of Submission 2017-08-24

    [Back to Results | New Search]


    Browse | Search All Available ETDs

    If you have more questions or technical problems, please contact eThesys