Open AccessPEMODELAN INDEKS HARGA SAHAM GABUNGAN MENGGUNAKAN REGRESI SPLINE MULTIVARIABEL
Author(s) -
Ihdayani Banun Afa,
Suparti Suparti,
Rita Rahmawati
Publication year - 2018
Publication title -
jurnal gaussian : jurnal statistika undip
Language(s) - English
Resource type - Journals
ISSN - 2339-2541
DOI - 10.14710/j.gauss.v7i3.26659
Subject(s) - nonparametric statistics , mathematics , knot (papermaking) , econometrics , statistics , stock (firearms) , mean squared error , regression , spline (mechanical) , geography , engineering , archaeology , structural engineering , chemical engineering
The Composite Stock Price Index (CSPI) is a composite index of all types of shares listed on the stock exchange and their movements indicate the conditions occurring in the stock market. CSPI movement is an important indicator for investors to determine whether they will sell, hold, or buy a stock. One of the factors that influence the movement of CSPI is Inflation (X1), Exchange Rate (X2) and SBI rate (X3). This study aims to obtain the best CSPI model using a multivariable nonparametric spline regression approach. The approach is done by nonparametric regression because the regression curve obtained does not show a certain relationship pattern. Spline is very dependent on the order and location of the knot point. The best spline model is the model that has the minimum MSE (Mean Square Error) value. In this study, the best spline regression model is when X1 is 4 order, X2 is 2 order and X3 is 2 order. The number of knots on X1 is 1 knot at 8.22, X2 is 2 knots at 13066.82 and 13781.75 While X3 is 2 knots at 6.6 and 6.67 with value of MSE equal to 6686.85.Keywords: Composite Stock Price Index, Multivariable Spline Regression, MSE