Predicting the City and Districts Consumer Price Index in East Java with the Gaussian-Radial Basis Function Kernel
Mimin Fatchiyatur Rohmah1, I Ketut Gede Darma Putra2, Rukmi Sari Hartati3,Luki Ardiantoro4

1,4 Informatics Engineering Universitas Islam Majapahit
2 Informatics Engineering Udayana,
3 Electronics Engineering,Udayana

Email: 1 miminfr[at]unim.com, 2 ikgdarmaputra[at]unud.ac.id, 3 rukmisari[at]unud.ac.id, 4 ipan.ardianto[at]unim.com

Abstract

An economic indicators regarding information on prices of goods and services paid by consumers are known as the Consumer Price Index (CPI). In this study the researchers predicted the Consumer Price Index for Foodstuffs in Cities and Districts in East Java using the Gaussian-Radial Basis Function Kernel for 2019. As a comparison, Foodstuff Type CPI issued by the Central Statistics Agency and as an input variable taken from the prices of basic commodities in three districts namely Banyuwangi District, Jember District and Sumenep District and five cities namely Kediri City, Madiun City, Malang City, Probolinggo City and Surabaya City from 2016 to 2018. The SVR method aims to find a function as a regression function hyperplane that matches the input data by making the least error possible. Forecasting data using the SVR method, the data is divided into training data and testing data. With the RBF kernel where the function is to produce CPI predictions with the smallest Mean Squared Error (MSE) in the City of Kediri of 0.0067 and the Mean Absolute Percentage Error (MAPE) of 0.0191. The average MSE and MAPE for the three districts and five cities are 0.011275 and the average MAPE is 0.0322125.

Keywords: Consumer Price Index, Support Vector Regression, Mean Absolute Percentage Error, Mean Square Error, Gaussian-Radial Basis Function.

Topic: Computer Science and Engineering

Link: https://ifory.id/abstract-plain/JMm4Vp6KfaeC

Web Format | Corresponding Author (Mimin Fatchiyatur Rohmah)