Stock Price Prediction Using ARIMA: A Study on AAPL, MSFT, NFLX, and GOOGL Stocks

It is widely acknowledged that being able to predict stock prices with accuracy remains the crux of classical financial statements, and has a multipurpose function including benefits for individual investors, portfolio managers, and policymakers. This paper explores the efficacy of the Autoregressive Moving Average (ARMA) model in predicting the stock prices of four leading technology companies: For example, Apple Inc. (AAPL), Microsoft Corp. (MSFT), Netflix Inc. (NFLX), and Alphabet Inc. (GOOGL). Using the original daily stock prices from January 2010 to December 2023, this study is going to trace out the good and bad of applying the ARMA model. Research utilizes a systematic approach to data preprocessing that makes stationarity features available and makes optimal lag selection based on the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) possible. The performance of the ARMA model is rigorously analyzed using three primary metrics: Mean absolute errors (MAE), Root mean square Error (RMSE), and Mean absolute percentage error (MAPE). The results value stands out in that the ARMA model can achieve good accuracy, with the MAPE values ranging from 1.10% to 1.50%. In such a manner, as half the cases were correct, the method can be considered a reliable forecasting instrument. It should be added that comparative study with more complicated models, such as ARIMA, LSTM, and GARCH models, gives the ARMA model a competitive advantage, primarily in terms of the simplicity of its code and the shortness of its execution time. The study puts ARMA models in a strong proactive position and ends with ARMA model further research into the hybrid model that combines ARMA with advanced machine learning techniques aimed at improving predictive accuracy.