Open AccessModeling of Indonesia Composite Index using Artificial Neural Network and Multivariate Adaptive Regression Spline
Author(s) -
Mutia Yollanda,
Dodi Devianto,
Putri Permathasari
Publication year - 2019
Publication title -
mantik
Language(s) - English
Resource type - Journals
eISSN - 2527-3167
pISSN - 2527-3159
DOI - 10.15642/mantik.2019.5.2.112-122
Subject(s) - multivariate adaptive regression splines , multicollinearity , econometrics , composite index , statistics , artificial neural network , stock exchange , multivariate statistics , mars exploration program , mathematics , stock market index , economics , regression analysis , computer science , bayesian multivariate linear regression , artificial intelligence , geography , stock market , finance , composite indicator , physics , context (archaeology) , archaeology , astronomy
The Indonesian Composite Stock Price Index is an indicator of changes in stock prices are a guide for investors to invest in reducing risk. Fluctuations in stock data tend to violate the assumptions of normality, homoscedasticity, autocorrelation, and multicollinearity. This problem can be overcome by modelling the Composite Stock Price Index uses an artificial neural network (ANN) and multivariate adaptive regression spline (MARS). In this study, the time-series data from the Composite Stock Price Index starting in April 2003 to March 2018 with its predictor variables are crude oil prices, interest rates, inflation, exchange rates, gold prices, Down Jones, and Nikkei 225. Based on the coefficient of determination, the determination coefficient of ANN is 0.98925, and the MARS determination coefficient is 0.99427. While based on the MAPE value, MAPE value of ANN was obtained, namely 6.16383 and MAPE value of MARS, which was 4.51372. This means that the ANN method and the good MARS method are used to forecast the value of the Indonesian Composite Stock Index in the future, but the MARS method shows the accuracy of the model is slightly better than ANN.