Modeling of composite stock price index (CSPI) using semiparametric regression truncated spline based on GUI R

Abstract

Abstract The Composite Stock Price Index (CSPI) is widely used as the beginning of consideration for investors to invest, because from the CSPI it can be known the general situation of market conditions is whether stock prices are experiencing an increase or decrease. This condition is characterized by a decrease or increase in the price of the CSPI. The Amount of Money Supply and the NASDAQ Index is thought to affect the price movements of the CSPI. The Amount of Money Supply has a nonparametric relationship pattern to the CSPI and the NASDAQ Index has a parametric relationship pattern to the CSPI. The correct method for conducting modeling is use the semiparametric spline truncated regression. The semiparametric regression spline truncated coefficient is estimated using the Ordinary Least Square (OLS) method which is determined based on the polynomial degree, much and the location of the optimum knot point is seen from the criteria of Mean Square Error (MSE). This study uses (Graphical User Interface) GUI R with the intention of facilitating the analysis process. The data used are monthly data from June 2014 to March 2019. Based on the results of the analysis that has been done, the best semiparametric spline truncated regression model with order 3 with the optimal three knots is 4246.361, 4443.078, and 4730.38. Evaluation results the in-sample data model produces a coefficient of determination of 90.25%. The results of the performance evaluation of the out sample data model resulted in a MAPE value of 3770204% indicating the performance of the model was very good.