Loading...

Semantic Analysis and Event Detection Using Deep Learning for Stock Prediction

Basirian Jahromi, Ali | 2018

496 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 51656 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Sameti, Hossein; Bokaei, Mohammad Hadi
  7. Abstract:
  8. News plays a very important role in stock market trading. Nowadays news from a different part of the world and about different fields can be accessed easily, and for a successful trade, it is necessary to analyze accurately and use this big data and information as soon as possible. For this reason, this thesis tries to present and study models based on Deep Learning networks and Natural Language Processing for financial news analysis and predicting stock indices movement. This research takes advantage of a language model for learning and representing news text, and beside this language model it uses deep learning networks at multiple levels to extract proper features from each news in a day which are related to a targeted index and then in next step extract better and more fine-grained features from all of extracted news features related to that index in that day, and use those to predict that specific index movement. In this thesis, multiple models are proposed and implemented based on Dynamic Memory, Hierarchal Attention, a combination of Inception and Residual architectures repeating in two levels and the combination of Inception and linear repeating in two levels. Previously best accuracy reported in the similar dataset for predicting an index representing a number of companies was 66.93%. But model based on Inception-Residual was able to achieve 68.11% accuracy for the dataset tagged on full texts and 69.40% accuracy for dataset tagged on titles. The model based on Inception-Linear was able to achieve the accuracy of 69.78% for dataset tagged on titles
  9. Keywords:
  10. Stock Market ; Big Data ; Deep Learning ; News ; Natural Language Processing

 Digital Object List

 Bookmark

No TOC