Yield Prediction Modeling Using Data Envelopment Analysis Methodology for Direct Seeding, Wetland Paddy Cultivation

A yield prediction model was developed using measured energy inputs data from 40 wetland paddy farms in Malaysia. Energy inputs from six sources were optimized using benchmarking and data envelopment analysis. The model developed based on Cobb-Douglas production function has coefficient of determination (R2) of 91%. Analysis on the model revealed direct relationship of yield with machinery, fuel, fertilizer and pesticides energy expenditures and an inverse relationship of yield with seeding rate. Hence, suggesting the need to reduce farmers’ present seeding rate of 149 kg/ha to the optimum seeding rate of 128 kg/ha for higher yield and cost reduction.