Forecasting using Statistical Modeling and Deep Learning Techniques

Abstract

Objective: Many financial institutions are interested in forecasting the conditions of financial markets by explaining a limited number of variables that affect the financial markets. This is done by designing statistical models using neural network applications to help obtain more accurate forecasting of changes in the financial markets. Material and Method: Statistical analysis of real data and compare the different modeling using time series modeling and deep learning, for the daily closing price index over the period of 1st January 2013 to 31st  January 2018 data in the Kingdom of Saudi Arabia. Results: The prediction results using  the proposed Neural Network are presented and discussed through the data of  a daily closing price index. The prediction results are more acceptable with normalization of input, that the sum square error is lower. Conclusion: For this studies we establishing the attitudes towards forecasting: It was established that these methods or models, that best fitted available data for the data of  a daily closing price index.