On-Demand Forecasting-Based Crop Yield Prediction and ‎Recommendation Using Deep Ensemble Swarm Intelligence ‎with Multi-Perceptron Neural Network

On-Demand Forecasting-Based Crop Yield Prediction and ‎Recommendation Using Deep Ensemble Swarm Intelligence ‎with Multi-Perceptron Neural Network Kuppan. P Dr. Vishwa Priya. V

Agriculture is a fast-growing industrial resource that has developed the Indian economy in recent years to achieve more crop production ‎worldwide. The traditional methodologies don't carry the multi-feature constraints and forecasting rate to degrade the recommendation ‎accuracy because of a lower precision rate and accurate positive margins. To resolve this problem, we propose an on-demand forecasting-based ‎Crop Yield Prediction and Recommendation Using Optimal Spider Swarm Intelligence Technique (OSSIT) with a Perceptron Neural ‎Network (MPNN). The multi-constraints data are based on metrological seasonal data, crop production rate, and demand forecasting rate to ‎augment the collective dataset. C-score Min-max normalizer preprocesses to process the data and formalizes the feature limits to scale the ‎actual and ideal margin variations. Then, the Crop Subjectivity Impact Rate (CSIR) is analyzed with decision tree margins (DTM) to identify ‎the actual support crop production for feature relations. Further, absolute demand forecasting variation feature limits are observed with ‎the ARIMA moving index rate to identify the findings in feature scaling. Also, the feature selection is carried out by the Optimal Spider Swarm ‎Intelligence Technique (OSSIT) to determine the relational features by considering the multi-concern feature relation. The non-relation ‎features are reduced accordingly by the inequality relation to avoid the feature dimension. By intention, the Multi Perceptron Neural ‎Network (MPNN) takes the multi-constraint feature margins. The proposed system produces high performance by selecting the correct ‎feature dependencies for selecting the seasonal crop and absolute mean growth rate with demand-level margins. The proposed system ‎produces a higher precision agriculture rate, recall rate, F1 measure, and lower false rate with redundant time complexity‎.
11 01 2025 538 584 10.14419/rv2aqw26 https://www.sciencepubco.com/index.php/IJBAS/article/view/34991 https://www.sciencepubco.com/index.php/IJBAS/article/download/34991/19405 https://www.sciencepubco.com/index.php/IJBAS/article/download/34991/19405