Stock Index Forecasting Using PSO Based Selective Neural Network Ensemble
Xueqin Zhang, Yuehui Chen, Jack Y. Yang · 2015
Abstract — Stock market analysis is one of the most important and hard problems in finance analysis field. Recently, the usage of intelligent systems for stock market prediction has been widely established. In this paper, a PSO based selective neural network ensemble (PSOSEN) algorithm is proposed, which is used for the Nasdaq-100 index of Nasdaq Stock MarketSM and the S&P CNX NIFTY stock index analysis. In the algorithm, each neural network is obtained by bagging and is trained by PSO algorithm, and then the networks selected according to the pre-set threshold are combined. Experimental results show that the improved algorithm is effective and outperforms GA based selective ensemble (GASEN) algorithm for the stock index forecasting problems. I.