Modelling of Artificial Intelligence (AI) Algorithms for Analysis of the Effects of Agricultural Production in Nigeria

Abstract

Agriculture is a cornerstone of Nigeria's economy, particularly in regions like Anambra State. However, its potential is hindered by fluctuating yields and the complex interplay of environmental factors such as temperature, rainfall, and flooding. This study explores the application of Artificial Intelligence (AI) models to analyze and optimize agricultural productivity for four key crops—cassava, yam, maize, and rice—in the Aguata Local Government Area. Using thirteen years of historical data on environmental inputs and crop yields, the study developed and compared three predictive models: an Artificial Neural Network (ANN), a Response Surface Method (RSM) model, and a Linear Regression model. Sensitivity analysis revealed that temperature was the most influential factor for all crops, while the impact of rainfall, relative humidity, pressure, and flood severity varied significantly by crop type. Permutation importance analysis revealed temperature as the most influential positive factor for all crops, while rainfall and pressure often exhibited negative or insignificant effects, highlighting the complex, non-linear relationships between climatic factors and crop productivity. Model adequacy metrics showed the ANN modelto be superior overall (average R2=0.852), and the lowest mean squared error (0.038), outperforming both RSM and Linear Regression in capturing complex patterns. Subsequent multi-objective optimization using Differential Evolution confirmed the ANN model's superiority, achieving the highest total predicted yield (34.2) and the lowest fitness score (0.068) under optimized environmental conditions (e.g., Temp=29.22∘C, Rainfall=1895.01 mm). The findings underscore the significant potential of AI, particularly ANNs, to inform data-driven decision-making in agriculture, enabling farmers and policymakers to enhance productivity and resource allocation in the face of environmental variability to boost food security and economic stability.