Financial stock market forecast using data mining in Palestine

Abstract EN

This research involves building a model for forecasting stock price movements using data mining techniques with Artificial Intelligence classifier.

Concerning the forecasting problem, we tackle the challenge of choosing Palestinian Stock Exchange (PSX) as an input data in the period between 2005-2017 and processing it accordingly to improve the predicting accuracy for the Palestinian stock market prices.

We also address the issue of evaluating the prediction performance.

Our results show that the best configuration for ANN is 6-5-1, which means six inputs, five neurons in hidden layer and one output.

ANN classifier is feed-forward multilayer perceptron neural network that is trained with back propagation algorithm.

In addition, the results show we have a high degree of accuracy for prediction (0.010 +/- 0.000 for Root Mean Squared Error), this accuracy is higher than other related researches