/Stock-Prediction-Using-Data-Science

Using Long short-term memory (RNN) to predict Stocks

Primary LanguageJupyter Notebook

Stock Prediction with Recurrent Neural Network

Stock price prediction with RNN. The data we used is from the Chinese stock.

Requirements

  • Python 3.5
  • Pandas 0.19.2
  • Keras 1.2.2
  • Numpy 1.12.0
  • scikit-learn 0.18.1
  • TensorFlow 1.0 (GPU version recommended)

Demo

Data

Data used is Chinese Stocks data, and here is a glimpse of the DataFrame

Data

Training

I've trained with 5 epochs, for a basic visualization.

Training Demo

Results

Training Result Demo

Reference