Modeling of NASDAQ-GEM stock price relationship using neural network

It is believed that the NASDAQ index has been one of the major “news” affecting the GEM stock prices. In order to understand the complex relationship between this index and the GEM stock prices, the time-series models using this index as exogenous input are studied. In addition, the correlation between this index and the GEM stock prices, and the reduction of error variance using neural networks are investigated. Based on the significance of this NASDAQ effect, seven GEM stocks are extracted and examined using different neural networks. In-sample and out-of-sample tests are performed using this index or the change of this index as the exogenous input. A comparison among different neural networks is given