Analisis Nilai Inflasi Bulanan Indonesia Menggunakan Regresi Nonparametrik Estimator Kernel

High levels of inflation are plaguing Indonesian society. Inflation occurs due to price increases as indicated by the increase in most expenditure group indices. This can lead to a higher poverty rate in Indonesia. This study aims to identify the best method that can be used to estimate Indonesia's monthly inflation value based on a nonparametric regression approach with a kernel estimator and analyze the results of predicting Indonesia's monthly inflation value for the next four months. The data used in this study is secondary data sourced from Bank Indonesia, with the variable used is the value of Indonesian inflation during the period January 2019 to July 2024. The collected data were analyzed using descriptive statistics and analytical statistics in the form of nonparametric regression with kernel estimators and predictions using kernel estimator and non-seasonal ARIMA methods. The results showed that triweight kernel regression was the best kernel function model with a minimum bandwidth value of 1.214, value of 99.990, MSE of 0.00016, and MAPE of 0.348%. The results of data prediction for the next thirteen months provide that triweight kernel estimator was better than non seasonal ARIMA method, with a MAPE value of 10.92%, so that the nonparametric regression method with the triweight kernel function is good or accurate in predicting data, which also can be used to analyze and predict Indonesia's monthly inflation data.