PREDICTING THE STOCK PRICE OF APPLE, INC. (AAPL) USING THE ARIMA MODEL AND CROSS-CORRELATION TECHNIQUE WITH THE STOCK PRICE OF ADVANCED MICRO DEVICES, INC. (AMD)

• Main Author: Wan Kedana, Tabina
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81513
• Institution: Institut Teknologi Bandung

This study aims to model and predict the stock price movements of Apple, Inc. (AAPL) using the ARIMA model and cross-correlation techniques. The historical stock price data for AAPL used in this study covers the period from July 3, 2023, to December 3, 2023, chosen for its relevance to the analysis conducted. The ARIMA(1,1, 0) model was identified as the best model for this data compared to the ARIMA(1,1, 1), ARIMA(12,1, 0), ARIMA(12,1, 1), and ARIMA(16,1, 0) models. The selection of the ARIMA(1,1, 0) model was based on the lowest AIC and BIC values, as well as the Ljung-Box test results that did not reject the null hypothesis, indicating model suitability. Further evaluation demonstrated the model’s performance with an RMSE of 1.897, MAPE of 0.848, and MAE of 1.562. Based on these results, the ARIMA(1,1, 0) model was concluded to be the most appropriate for predicting AAPL stock price movements during the selected period. The ARIMA(1,1, 0) model also showed consistency and reliability in predicting AAPL stock prices for the periods July 4, 2023 - December 4, 2023, July 5, 2023 - December 5, 2024, July 6, 2023 - December 2023, July 7, 2023 - December 7, 2023, July 8, 2023 - December 8, 2023, July 9, 2023 - December 9, 2023, and July 10, 2023 - December 10, 2023. Cross-correlation analysis was also applied to predict the stock price movements of Advanced Micro Devices, Inc. (AMD). The model evaluation using RMSE, MAPE, and MAE metrics showed very satisfactory results with very small values. This study implemented the prediction of AAPL stock prices based onAMDstock prices through a cross-correlation value diagram between AAPL stock price predictions and their actual values, as well as numerical analysis of the cross-correlation technique. The results indicated that short-term predictions of AAPL stock prices can be made based on AMD stock prices one or two days prior, considering the differencing process conducted in the fitted ARIMA model. These findings suggest that the cross-correlation technique with ARIMA(1,1, 0) provides a strong basis for predicting AAPL stock prices based on AMD stock price data, given that the data has proven to be non-heteroskedastic.