NAV of Equity Fund Forecasting Using ARIMA and RNN

Abstract

This study aims to predict NAV of equity fund by comparing two commonly used methods, namely ARIMA and RNN. Both of these methods have their own characteristics which will be calculated based on their accuracy based on the MSE and RMSE values from the forecasting results. This study uses the CRIPS-DM framework as a reference, for the data used are Sucorinvest Equity Fund mutual fund data from 2017 to 2019. For ARIMA, the variable used is only NAV. For RNN, besides NAVs, stock indexes are also used that affect the value of NAV. Based on the results of the study, the ARIMA and RNN error rates are not too much different, but RNN is better.

Keywords: Equity Fund, NAV, ARIMA, RNN, CRIPS-DM