A hybrid model of least squares support vector regression and particle swarm optimization for Vietnam stock market analysis

Abstract

Machine learning techniques have been successfully applied in time series data prediction. In this paper, a combination of least squares support vector regression (LSSVR) and particle swarm optimization (PSO) is used to forecast stock prices in Vietnam stock market. The PSO is adopted to optimize the hyperparameters of the LSSVR for improving the forecast accuracy. The proposed model (PSO-LSSVR) is validated by two financial time series data, including the daily VN-INDEX 100 and the daily stock price of Gas Joint Stock Company (PGC). The forecast accuracy of the PSO-LSSVR is compared with that of the autoregressive integrated moving average (ARIMA) and the LSSVR via performance measures, including mean absolute error (MAE) and mean absolute percentage error (MAPE). The experimental results show that the predictive ability of the PSO-LSSVR outperformed that of the ARIMA and the LSSVR both datasets. This finding provides a promising solution in predicting financial time series data.