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
*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.
*Train Score: 0.00006 MSE (0.01 RMSE)
*Test Score: 0.00029 MSE (0.02 RMSE)
- Include more technical indicators from *https://github.com/saradindusengupta/technical_indicators_stock-market
- Use tweets for sentiment analysis more effectively
- More data
- More Indexes and better optimized hyperparameter
-
Bernal, A., Fok, S., & Pidaparthi, R. (2012). Financial Market Time Series Prediction with Recurrent Neural Networks.
-
A hybrid SOFM-SVR with a filter-based feature selection for stock market forecasting CL Huang, CY Tsai - Expert Systems with Applications, 2009 - Elsevier
-
Twitter mood predicts the stock market J Bollen, H Mao, X Zeng - Journal of computational science, 2011 - Elsevier
-
Evaluating the impact of technical indicators on stock forecasting IEEE
-
A hybrid stock trading framework integrating technical analysis with machine learning techniques Rajashree Dash Pradipta Kishore Dash
Language - Python 3.5
keras : https://keras.io/
tensorflow : https://www.tensorflow.org/
sklearn: http://scikit-learn.org/stable
numpy : http://www.numpy.org/