Firefly Algorithm with Deep Belief Network for Crop Classification Model in Agricultural Sector

Data mining (DM) is an effective way to extract essential and meaningful data to determine patterns from a large dataset. Data mining finds its applications in diverse areas of healthcare, agriculture, transport, e-commerce, etc. In the agricultural sector, the recently developed DM and machine learning (ML) methods can be leveraged for effective decision-making. Crop classification is the process employed to properly categorize various kinds of crops in an automated and timely way. At the same time, hyperparameters perform a decisive role in affecting the overall classification performance of the deep learning (DL) models. Therefore, this study develops an optimal deep belief network (OPTDBN) approach for crop classification in the agricultural sector. The proposed OPTDBN model intends to properly classify distinct types of crops that exist in the agricultural data. To attain this, the OPTDBN model performs an effective classification process using the DBN model. To enhance the performance of the DBN based predictive technique, the firefly (FF) algorithm is utilized and it results in enhanced classification results. The application of the FF algorithm assists in optimal hyperparameter selection of the DBN model. To exhibit the superior outcomes of the OPTDBN methodology, a sequence of simulations were effectuated and the outcomes are reviewed under soybeans and mushroom datasets. The simulation outcomes stated the enhanced performance of the OPTDBN methodology than the other current approaches interms of different measures.