Decision trees to multiclass prediction for analysis of arecanut data

Currently, the grading of arecanuts is done manually by trained experts, and no work has been attempted towards an automated classification. This paper discusses an automated technique for the classification of arecanuts, based on their texture. In our previous work [1], we made a classification using the Mean Around features, Gray Level Co-occurrence Matrix (GLCM) features, and a combination of both (Mean-around and GLCM). In this paper, we formulate different tree splitting rules and discuss the comparative results. The decision tree classifiers have been used for the classification of arecanuts into six classes. The splitting rules used for constructing the decision tree are the Gini Diversity Index (GDI), Twoing and Entropy. The results obtained from the proposed method are in good agreement with the observations of the agricultural experts, and the solutions proposed have been well received by them. The experiments on the proposed approach were carried out on a dataset of size 2214 using cross validation, with a good success rate.