Hybrid Regression Model for Soil Moisture Quantity Prediction

Predicting quantity of soil moisture would help the farmers who are involved in agriculture. Recently researchers have used various machine learning algorithms to predict the quantity of soil moisture. In this study a hybrid method of different types of regression algorithms have been employed for prediction. The proposed method first clusters the dataset using K-Means Clustering, then for each clusters an individual regression model is trained. The proposed method is compared with other regression models like DT (Decision Tree), MLP (Multilayer Perceptron)-Regressor, LR (Linear Regression), KNN-Regressor and SVM (Support Vector Machine Regression). Experimental results have shown that Hybrid model using Decision Tree Regression achieved an average RMSE of 0.002924 and outperformed other models.