Abstract:
Capital Investments in the stock market is the easiest and fastest way of building the healthy financial
foundation for future life. In the past few decades, stock markets around the world have become more
institutionalized and advanced as the main forms of investments for making profit investments in
numerous organizations as well as individuals to arrange their large investment funds to the general
public. As a result, the stock market prediction has become one of the great challenges caused by its
complexity and eruptive nature. Generally, stock prices are chaotic and show both linear and nonlinear
behaviors. Therefore, the accuracy of the forecast might be enhanced by modeling the non-linear
behaviors of the series as well. The main purpose of this study is to take an attempt to understand the
behavioral patterns as well as seek to develop a new hybrid forecasting approach based on Geometric
Brownian Motion (GBM) for estimating price indices in Colombo Stock Exchange (CSE), Sri Lanka.
Indeed, the Autoregressive integrated moving average (ARIMA) approach is used as a comparison
mode. The current study was carried out on the basis of CSE daily trading data from January 2010 to
May 2018 were extracted and tabulated for calculations. Because of the nonlinear behavioral patterns
in the CSE, the mean absolute percentage error analysis results suggested that new proposed hybrid
model (HGBM) is highly accurate than traditional ARIMA (HGBM (0.521%) < ARIMA (7.18%)) for
forecasting one day ahead predictions. Furthermore, the results reveal that, the new proposed model is
more significant for investors to make their investment decisions precisely
