snci/stock-forecasting

Stock closing and opening forecasting using Deep neural network and LSTM(technical indicators included)

Stock closing and opening forecasting using Deep neural network and LSTM(technical indicators included)

Details about the indicators are here *https://github.com/saradindusengupta/technical_indicators_stock-market

Data :

*https://www.kaggle.com/souravroy1/stock-market-data/data The two py files

stock-forecast-lstm.py
stock-forecast-tweet.py

are for forecasting stock opening and closing prices from twitter and NYtimes using deep neural network and lstm. The notebook directory contains the results and ipynb files.

Results

![Alt text](

*Train Score: 0.00006 MSE (0.01 RMSE)

*Test Score: 0.00029 MSE (0.02 RMSE)

Future Improvements :

1. Include more technical indicators from *https://github.com/saradindusengupta/technical_indicators_stock-market
2. Use tweets for sentiment analysis more effectively
3. More data
4. More Indexes and better optimized hyperparameter

References:

1. Bernal, A., Fok, S., & Pidaparthi, R. (2012). Financial Market Time Series Prediction with Recurrent Neural Networks.

2. A hybrid SOFM-SVR with a filter-based feature selection for stock market forecasting CL Huang, CY Tsai - Expert Systems with Applications, 2009 - Elsevier

3. Twitter mood predicts the stock market J Bollen, H Mao, X Zeng - Journal of computational science, 2011 - Elsevier

4. Evaluating the impact of technical indicators on stock forecasting IEEE

5. A hybrid stock trading framework integrating technical analysis with machine learning techniques Rajashree Dash Pradipta Kishore Dash

Dependencies :

Language - Python 3.5

keras : https://keras.io/

tensorflow : https://www.tensorflow.org/

sklearn: http://scikit-learn.org/stable

numpy : http://www.numpy.org/

pickle: https://docs.python.org/2/library/pickle.html