Long Term Stock Prediction based on Financial Statements

Code

• github.com/sugia/tradeX
  In papertf★ 9

Abstract

This paper proposes a model with LSTM and fully connected layers to predict long term stock trendings based on financial statements. Two data augmentation techniques are applied on structured data: 1) adding random noise to data fields; 2) erasing partial information from training examples. The performance of the proposed approach is demonstrated on real-world data of about 7,000 stocks listed on Nasdaq.

Tasks

Reproductions