Performance Analysis of Pistachio Species Classification using Support Vector Machine and Ensemble Classifiers

The effectiveness of post-harvest industrial operations is crucial for preserving the economic value of pistachio nuts, which play a significant role in the agricultural economy. To achieve higher efficiency, new methods and technology are required for pistachio separation and categorization. Different pistachio species cater to distinct markets, which enhances the necessity for species categorization. The objective of this paper is to construct different classification models that differs from existing separation approaches and is capable of providing the requisite classification based on machine learning (ML) techniques. In this paper, a multi-level system comprising feature extraction, feature selection, and classification is developed. To analyze the performance of variety of Support Vector Machines (SVM) and Ensemble Classifiers (ECs), initially a set of 28 attributes are extracted from a dataset of 2148 images. Further, with the help of Principle Component Analysis (PCA) 16 effective features are identified to train the models. Experimental findings revealed that the suggested method had a classification success rate of 93.9 percent. From the analysis, it is observed the proposed models are superior to that of existing approaches.