Forecasting of Retirement Insurance Filled Forecasting of Retirement Insurance Filled via Internet by ARIMA Modelsvia Internet by ARIMA Models

Abstract

Pension fund insurance is critical for everyone because it can guarantee a good life during retirement because retirement is a period when someone no longer gets a steady income. Technological advances make it easier for retirement insurance applications. By using ARIMA Models, we can predict the number of internet users who apply for retirement insurance via the internet, using the monthly data of the Social Security Administration from January 2008 to October 2020. The data used has a steady increasing trend with some seasonal components, so it needs to be removed first. ARIMA models use the assumption that the data is stationary, so the data must be tested using the ADF test command in R. After seeing the plotting of ACF and PACF, 9 ARIMA models are formed. ARIMA model is selected based on the smallest AIC. By using 95% confidence it can be concluded that ARIMA (9,1,9) is the best model for forecasting.