MODEL REGRESI NONPARAMETRIK KERNEL MENGGUNAKAN ESTIMASI NADARAYA-WATSON UNTUK DATA HARGA INDEKS SAHAM GABUNGAN DI INDONESIA
No Thumbnail Available
Date
2022-05
Journal Title
Journal ISSN
Volume Title
Publisher
Elfitra
Abstract
The composite index is one of the stock price indexes in Indonesia. In this study the
kernel nonparametric regression model with the Nadaraya-Watson estimator for the
composite index data. The kernel regression method is one of the methods in
nonparametric regression used to estimate conditional expectations using kernel
functions. The kernel function used in this study is the Gaussian kernel function. The
data used is the composite stock price index in Indonesia in 2018-2019. The first step is
to determine in advance the optimal bandwidth with the Generalized Cross Validation
(GCV) method. Kernel regression using gaussian kernel functions obtained a bandwidth
value of 82,03 with an optimal GCV of 269,42. Based on the results of the analysis to
measure the goodness of the model using Mean Absolute Square Error (MAPE) of 2,71,
which means that the MAPE value is in the first category which is very good.
Description
Keywords
Nonparametric regression, Nadaraya-Watson estimation, Generalized Cross Validation, bandwidth, Mean Absolute Square Error
Citation
Perpustakaan