Reverse Engineering of Stock Selection Algorithms in the Capital Market

Please enable javascript in your browser.

Rajabzadeh Estahbanati, Mohyaddin | 2022

134 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 55183 (19)
University: Sharif University of Technolog
Department: Computer Engineering
Advisor(s): Habibi, Jafar
Abstract:
Capital market is one of the important ways to invest. Investors use different ways to analyze this market. One of the most important ways is to use stock filters that people need to learn both programming and capital market concepts. Automatic programming tools emerged to solve the need to learn a lot of programming languages in different fields but in stock filters have not been used until now in ower knowledge. In this research, the goal is to create automatic programming techniques to recognize stock filters from input-output examples. For this task, in the first step, the history of stock market have been used and a lot of stock filters with different length have been created. After preprocessing this information, the problem of creating filters has been considered to be a classification problem and techniques like SVM with different kernels, Decision Tree, Random Forest and Neural Networks have been used to solve this problem. The best accuracy of this model reached 88 percent with data that models have not seen before.
Keywords:
Data Analysis ; Stock Market ; Reverse Engineering ; Stock Filter ; Capital Market Analysis ; Programming by Example ; Automatic Programming

No TOC