Iraqi Stock Market Prediction Using Hybrid LSTM and CNN Algorithm

Predicting a stock price index is an important method that contributes to increasing confidence between the investor and the market, and helps him to make the right investment decisions, to achieve profits and avoid losses, because the wrong decisions carry heavy losses that may lead to bankruptcy often due to the sharp volatility of the index up and down, and provide opportunities to develop the right strategic plans in the long and short term, which in turn leads to the prosperity of the market, and attract investors to it by relying on future expectations that mimic the reality. In this paper, a hybrid model was developed for the purpose of increasing the accuracy of prediction the Iraqi stock markets index. Hybridization of the CNN Convolution Neural Network model with the most important long-term time series forecasting model, Long Short-Term Memory (LSTM), has been created to increase its accuracy by predicting. English translation. The study community is represented by the 6 sectors in Iraq's financial markets. The sectors studied are (banking, telecommunications, and tourism, agriculture, industry and services sectors) and a weekly data series has been taken for the previous five years from 2017 to 2021. Through many experiments on the proposed model, it has been concluded that the proposed system has reached a minimum error rate is 0.000404.