Sensitivity based linear learning method for forecasting the saudi arabia stock prices

In this paper, we have proposed Sensitivity Based Linear Learning Method (SBLLM) for the prediction of Saudi stock market. The proposed SBLLM prediction model, which is a variant of the classical ANN, has unique attributes of high speed, accuracy and stability, and thus could be used as investment advisor for the investors and traders in the Saudi stock market. We have only used the closing price of the stock as the variable considered for input to the system. The number of windows gap to determine the numbers of previous data to be used in predicting the next day closing price data has been choosing based on heuristics. The proposed model is based mainly on Saudi Stock market historical data covering a large span of time. The obtained results indicated that the proposed SBLLM model predicts the next day closing price stock market value with a low RMSE and high correlation coefficient of up to 99.9% for the test set, which is an indication that the model adequately mimics the trend of the market in its prediction. In comparison with ANN model, the proposed SBLLM achieve better performance in most cases, while having a competitive performance with ANN in few cases. In term of speed of operation, the SBLLM model achieved better aped of training throughout the reported experiments. The outcome of this work indicated that the proposed system will impact positively on the analysis and prediction of Saudi stock market in general.