COMPARATIVE ANALYSIS OF ARIMA AND LSTM MODELS IN FORECASTING EGG PRODUCTION TRENDS IN INDIA

Abstract

Egg production forecasting is crucial for managing the agricultural supply chain, as it impacts operational planning, market stability, and environmental sustainability. Efficient egg production ensures dietary needs are met and economic stability is maintained in the agricultural sector. Accurate forecasts aid resource allocation, market demand prediction, pricing optimization, and mitigation of supply chain disruptions.This research compares the efficacy of ARIMA (Autoregressive Integrated Moving Average) and LSTM (Long Short-Term Memory) models in predicting trends in egg production. Historical data spanning several decades were utilized, and the models’ performance was evaluated using the Root Mean Squared Error (RMSE). ARIMA, a traditional statistical model, and LSTM, a recurrent neural network, were applied to the data. The LSTM model demonstrated a significantly lower RMSE (2,299.54) compared to ARIMA (13,394.96), indicating superior performance in capturing the underlying patterns in egg production data. The LSTM model provided forecasts that were closer to the actual figures, particularly reflecting the upward trend in production from 2019 to 2021. The findings suggest that LSTM models are more effective for forecasting in agricultural contexts due to their ability to handle complex, nonlinear data. This can lead to better resource allocation, optimised supply chain management, and more stable market conditions. The study highlights the potential for integrating advanced machine learning models into agricultural forecasting to enhance decision-making processes. Future studies could explore hybrid models that incorporate external variables to further enhance forecast accuracy and reliability.