An investigation into predictive modelling of Malaysian stock prices using an optimum feature set

The stock market is often seen as a sentiment indicator which can indirectly impact the GDP (gross domestic product) of the nation either negatively or positively. At the personal level, many may have invested their life savings in retirement funds which are managed by professional fund managers. Thus if fund managers performed badly in the stock market, the performance of stock market will indirectly impact the returns of the retirement fund under their care. However, due to the huge amount of data that they are dealing with, it can be very challenging if not humanly impossible to analyze all the data to come up with an accurate model of the performance as well as the potential of the stocks. Artificial intelligence can help improve the effectiveness of fund managers. This can then allow a fund manager to manage multiple funds at a time, which in turn lowers the fund management fees, effectively giving the client a bigger capital return.. This paper describes a novel approach of using financial news sentiment and genetic algorithm to predict the performance of Malaysian stock prices. The performance of the final optimized model is also compared with other popular predictive models.