making neural Network for stock price prediction
Hole project is written in Python programming language. Beside basic libraries, I used: pandas - for csv file manipulation Theano - library for basic mathematical operations inside neural network (its only for usage of Keras) Keras - wrapper for Theano library (it does all calculations for me)
The idea behinde this project is to use LSTM (Long - Short - Term - Memory) to try and find connection between time series of our data. As data source i used Apple historical data prices downlaoded from Yahoo Finance. I used only closing prices to try and predict , next prices. Every 19 elements are taken into prediction, and we tried to predict next price.
We used 70% of data set for training , and 30% for testing purposes.