Advanced Predictive Modeling for Peanut Production Estimation in Mozambique: Addressing Food Security Challenges through ARIMA and LSTM Approaches Aligned with SDG 2

This study investigates the application of advanced predictive modeling techniques, specifically ARIMA models and LSTM neural networks, to estimate peanut production in Mozambique from 1961 to 2022, with projections extending to 2030. The research aims to provide accurate forecasts that are essential for addressing food security challenges, in alignment with Sustainable Development Goal 2 (SDG 2). The ARIMA models were calibrated using data from 1961 to 2009 and validated with actual production data from 2010 to 2020. Meanwhile, the LSTM model was trained on data from 1961 to 2013 and evaluated with data from 2014 to 2020. The analysis highlighted the relative stability of peanut production in Mozambique, alongside moderate variability, indicating the sector’s susceptibility to external factors such as climate change and agricultural practices. The ARIMA (1,1,1) model was identified as the most robust among the ARIMA models, effectively capturing temporal dependencies and smoothing abrupt fluctuations. However, the LSTM model’s ability to model nonlinear sequences and long-term dependencies resulted in significantly more accurate forecasts, with a mean absolute percentage error (MAPE) much lower than that of the ARIMA models. This study underscores the superiority of LSTM neural networks over ARIMA models for forecasting peanut production in Mozambique, demonstrating their effectiveness in capturing complex temporal patterns and generating more precise predictions. The projections indicate relatively stable production, albeit with a slight declining trend by 2030. These findings highlight the importance of investing in advanced agricultural technologies and climate change mitigation strategies to ensure the resilience of agricultural production in the country. In summary, the adoption of more sophisticated modeling approaches, such as LSTM, is crucial for addressing food security challenges in Mozambique, directly contributing to the achievement of SDG 2.