Loading...

News-based Stock Forecasting Using Text Mining Methods

Ashtiani, Mohammad Hossein | 2020

368 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53336 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Rafiee, Majid
  7. Abstract:
  8. Predicting the trend of stock prices is always one of the concerns that stock market analysts and investors face with, which plays a critical role in maximizing the profit from investing in stocks. Past stock price charts, raw material prices, the value of the company's assets, the impact of global markets, the company's products, the organization's development plans, and the same other factors influencing the stock price. News reports are an important source of information for people. Recently, a lot of research has been done to examine the impact of news on stock price trends. This research has two research phases. In the first phase, a text mining method is presented which is using various text mining techniques and pre-processing news texts to become documents containing significant information. Then, using the TF-IDF method, the words in the document are weighted, and after that we analyze the emotions of news articles with the VADER natural language tool and classifies articles into positive or negative. In the last step of the first phase, we predict the proposed method with four algorithms which the support vector machine algorithm with 98% precision creates better performance than the others and achieves higher accuracy than previous researches. The second phase combines news polarities obtained in first phase and some fundamental indicators of companies together and use them in a resilient supplier selection and optimal order allocation under disruption risks model. This selected the suppliers by minimizing the costs of the system and delivery time
  9. Keywords:
  10. Text Mining ; Sentiment Analysis ; Supplier Selection ; Resilient Supply Chain ; Stock Market ; Price Forecasting

 Digital Object List

 Bookmark

No TOC