کنفرانس مهندسی برق ایران

صفحه اصلی / سی و دومین کنفرانس بین المللی مهندسی برق

Machine Learning-based Fundamental Stock Prediction Using Companies’ Financial Reports

نویسندگان :

Hossein Rezaei¹ Kamran Abdi² Mohsen Hooshmand³

1- دانشگاه تحصیلات تکمیلی علوم پایه زنجان 2- دانشگاه تحصیلات تکمیلی علوم پایه زنجان 3- دانشگاه تحصیلات تکمیلی علوم پایه زنجان

کلمات کلیدی :

Stock market prediction،Fundamental feature analysis،Dataset generation،Machine learning،Time series،Statistical conversion

چکیده :

Portfolio management is a significant and crucial goal in investments. Therefore, the efficient prediction of the shares and securities plays an important role in highly profitable incomes. The fundamental stock analysis is a powerful and fruitful method of predicting the long-term of shares. However, further investigation in this area suffers from the lack of online fundamental features. This paper provides a fundamental feature dataset from TSE. Its features are extracted from the financial reports of companies and its prediction targets are both the return and risk values of each share. Moreover, this work investigates the effect of utilizing time dependency, absent in a vast majority of machine learning methods, on the prediction performance. Last, we propose machine-learning methods, i.e., logistic regression and gradient boosting to analyze the performance of fundamental stock prediction.