Performance Evaluation of Support Vector Machine with Random Forest and Multilayer Perceptron for the Quality Assessment of Rice

Rice is one of the most important item included in our daily diet. In market rice is being traded as cheaper quality one irrespective of various regions which might affect the customer's health and may lead to new diseases. In order to overcome this issue, there are various techniques that helps to check the rice quality using different ML techniques. Hence there is high time necessary to calculate the performance and to check quality of rice using SVM-support vector machine, random forest and MLP-multilayer perceptron. The accuracy achieved for SVM ranges about 59.00 to 67.00, random forest achieved is 70.00 to 74.00. Thus, we conclude that neural networks found with highest accuracy compared to other techniques such as SVM and random forest and multi- layer perceptron are better. The work states that the findings such as compared to CNN, SVM, KNN HOG and random forest trees and MLP holds better accuracies than SVM.