Comparison of Apple Inc Stock Forecasting Accuracy Using Hybrid TSR Linear-ARIMA Model and ARIMA Model

Abstract

This study aims to compare the accuracy of Apple Inc. stock price forcasting using two time series models, namely the hybrid TSR Linear-ARIMA model and the ARIMA model. The background of this research is the need for more accurate forcasting methods in a dynamic stock market, especially for technology stocks such as Apple which have high volatility. The research methodology uses the quantitative approach with daily Apple stock price time series data for the period 2023. The hybrid TSR Linear-ARIMA model incorporates trend and residual components, while the ARIMA model uses the Box-Jenkins approach. Both models were implemented using statistical software R Studio and Minitab. The results that the ARIMA model provided better forcasting accuracy compared to the hybrid TSR Linear-ARIMA model. Comparative analysis using the MAPE shows the ARIMA model has a lowwer error rate. Specifically, the ARIMA model produces a MAPE of 2.909%, while the hybrid TSR Linear-ARIMA model produces a MAPE of 3.780%. in conclusion, the ARIMA model proved to be more effective in forecasting the stock price of Apple Inc. compared to the hybrid TSR Linear-ARIMA model. This research contributes to the development of forecasting techniques in finance and investment, especially for technology stock.