Seagull optimization based deep belief network model for biofuel production

Abstract

Biofuels have emerged as a promising alternative to conventional fossil fuels due to their potential to decrease greenhouse gas emissions and reliance on non-renewable resources. Fluctuating energy costs and policy interventions have substantially increased global interest in biofuel production, imperative for population growth and accelerated economic development. High computation complexity, low accuracy, and other factors limited earlier works in biological production, which were overcome by predictive modeling, a promising approach to enhance efficiency and sustainability through precise forecasting and process optimization. This article introduces an innovative biofuel production prediction model named the Seagull optimization based deep belief network model for biofuel production (SGO-DBN), comprising four major stages: data pre-processing, reconstruction, prediction, and SGO optimization. The proposed model initially performs data pre-processing using the empirical mode decomposition (EMD) technique. A DBN model is used to predict biofuel production, which is further optimized by a seagull optimization algorithm-based hyperparameter optimizer. The biofuel production rate consistently increased over six years with minimal divergence between the predicted and actual outcome. A comparative analysis showed the computation time of the proposed SGO-DBN model was lower than that of existing techniques, while the rate of production analysis emphasized the model's robust predictive performance. Results of numerous simulations conducted to evaluate the model's performance based on various metrics showed that the SGO-DBN model surpassed the performance of recent state-of-the-art techniques.