A Software Framework for Predicting the Maize Yield Using Modified Multi-Layer Perceptron

Predicting crop yields is one of agriculture’s most challenging issues. It crucial in making national, provincial, and regional choices estimates the government to meet food demands its citizens. Crop production anticipated based on various factors such as soil conditions meteorological, environmental, variables. This study intends develop an effective model that can accurately anticipate agricultural advance, assisting farmers better planning. In current study, Yield Prediction Dataset normalized initially, then feature engineering performed determine significance assessing yield. yield forecasting using Multi-Layer Perceptron Spider Monkey Optimization method. The technique efficient dealing with non-linear relations among features data, would assist optimizing corresponding weights. uses data from Food Agriculture Organization World Data Bank forecast maize Saudi Arabia region average temperature, rainfall, Hg/Ha past years. suggested MLP-SMO model’s prediction effectiveness being evaluated several evaluation metrics Root-Mean-Square Error, R-Squared, Mean Absolute Bias where has outperformed process a Error value 0.11, which lowest all techniques are considered statical analysis study.