Comparison of ARIMA and Artificial Neural Network in Forecasting of Stock Price in Bursa Malaysia

Kah Chun Chai, Maria Elena Binti Nor

Stock market volatility has gained much attention by many researchers and practitioners as it reflects the degree to which price moves. Past volatility can be used to foresee the future volatility as this is an important approach for making better investment decisions and selecting a portfolio. The aim of this paper is to determine the best model to predict the stock price listed in Bursa Malaysia using two different approaches which are Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Network (ANN). Meanwhile, the training algorithm used in ANN is the Levenberg-Marquardt backpropagation training algorithm and is tested with different number of neurons in the hidden layer. The best ANN model is selected based on the accuracy error. Lastly, both ARIMA model and ANN model are evaluated in term of forecast performance. The result showed that the performance of ANN outperformed the ARIMA. It is proven that the nonlinear characteristic of ANN is more accurate in predicting the stock price.

Volume 11 | 12-Special Issue

Pages: 89-97