Abstract
MULTI-MODEL ENSEMBLE-MODIFIED DEPTH ADAPTIVE DEEP NEURAL NETWORK FOR CROP YIELD PREDICTION
Dr. M. Saranya*, Dr. S. Sathappan
ABSTRACT
Our recent study using crop yield prediction associated with climate, weather and soil for crop yield prediction. By combining regression models with Neural Networks (NN), can able to release highly satisfactory forecasting of crop yield. Prediction of crop yield accurately for tracking crop production is a trendy issue and it is a main area of research for agriculture studies. Multi-Model Ensemble Modified Depth Adaptive Deep Neural Network (MME-MDADNN) was an effective crop yield prediction method where the variation of climate, weather and soil parameters were learned through DNN. The existing Support Vector Regression (SVR) model with Deep Neural Network, the slow convergence speed, possibility of stuck in local minima and risk of over-fitting problems are resolved. Here the Ridge Regression (RR) model was applied to analyze multicollinearity in multiple regression data which enrich the better solution. Hence by applying ridge regression along with DNN, the crop yield prediction Accuracy is improved. The effectiveness of the proposed MME-MDADNN is tested in terms of Accuracy, Precision, Recall and F-measure.
[Full Text Article] [Download Certificate]