Grain Classification Using Different Machine Learning (ML) Classifier Along With Feature Extraction Techniques PCA & LDA

Abstract

The paper presents a model which does identification of multi grains. To design and test such model, datasets of raisin, dry-beans and rice are used. For the said purpose commonly preferred classifiers Decision Tree, Naïve Bayes & Support Vector Machine (SVM) are opted. The classifiers may accompanied by the feature extraction/reduction techniques (like PCA & LDA) to have better result. This possibility is explored in this paper. The performance of the model designed with Machine Learning (ML) classifiers only and along with feature reduction techniques are analyzed. It is observed that Decision Tree classifier serves better amongst opted ML classifiers but involving PCA/LDA in addition to ML classifier gives the best result. Additionally, the simulation results advocate the proposed model for high efficiency, accuracy and requirement of less computation power.