Stock price prediction, ARIMA-LSTM hybrid models, time series analysis
Abstract
In the context of highly volatile financial markets, accurately predicting stock prices remains an important and difficult task, particularly for major technology firms such as Apple Inc. (AAPL). Forecasting methods that are more traditional, like the Autoregressive Integrated Moving Average (ARIMA) model, perform well in linear analysis but have trouble handling the non-linear patterns that are common in financial information. In order to handle both linear and non-linear dynamics, this study presents a novel hybrid model that combines ARIMA with Long Short-Term Memory (LSTM) networks. This model efficiently leverages the advantages of both approaches. Utilizing historical closing price data for AAPL obtained from Yahoo Finance, the study demonstrates that the ARIMA-LSTM model significantly enhances predictive accuracy and adapts proficiently to the complexities inherent in stock market fluctuations. As seen by the data, the model’s predictive ability has significantly improved, accounting for 91.1% of the fluctuation in stock prices. In addition to providing a more trustworthy prediction framework, this hybrid approach offers analysts and investors helpful information to assist them negotiate the turbulent nature of financial markets.
