Enhancing GDP Forecasting in India Through a Hybrid ARIMA-LSTM Model

In the last few years, accurate GDP forecasting has been growing crucial for financial forecasting and governance in recent years. While traditional techniques, like as ARIMA, have proven effective in modelling linear trends,They frequently struggle to recognise the complicated dynamic relationships built into economic data. This research presents a new hybrid model that blends ARIMA and LSTM networks, which are both used to increase GDP prediction precision in India. Using GDP data from 1960 to 2022, the study applies inferential statistical tests to determine data stationarity and preprocess the dataset effectively. The ARIMA component detects linear trends, whilst the LSTM network solves non-linear patterns by analyzing residuals created by the ARIMA model. This combined strategy leverages the qualities of both approaches, resulting in better prediction performance. The hybrid model’s efficacy is measured against standard forecasting methodologies using measures such as MAE and MSE. Results show this combination method improves forecast accuracy while also providing useful insights into India’s economic dynamics, allowing for more informed planning.