The Autoregresiive Integrated Moving Average and Fuzzy Time Series Cheng Hybrid for Predicting Stock Price

Abstract

Background: PT Telkom Indonesia Tbk is the largest company in the telecommunications sector in Indonesia. PT Telkom's share price always rises every year, attracting investors to invest. In investing, it is very important to analyze shares in order to know the situation and condition of the shares. Objective: This research aims to predict the share price of PT Telkom Indonesia Tbk. Methods: The method used is the Autoregressive Integrated Moving Average (ARIMA)-Fuzzy Time Series Cheng hybrid method. Cheng's FTS model is able to overcome nonlinearity problems in ARIMA model residuals. In this research, the first modeling uses the ARIMA model, where the data is divided into two, namely January to November 2019 data used as training data, and December 2019 data used as testing data. Next, residual modeling was carried out with FTS Cheng. Hybrid forecasting is obtained by adding up the results of ARIMA and FTS Cheng forecasts. Result: Model evaluation is based on MAPE values and in this study the MAPE value of the ARIMA-FTS Cheng hybrid model was obtained at 1.03\% for training data and 1.09\% for testing data. Conclusion: The hybrid model has a MAPE value of less than 10\%, so it can be concluded that the ARIMA-FTS Cheng hybrid model can predict PT Telkom Indonesia Tbk stock closing price data accurately.