California Papaya Fruit Maturity Classification Uses Learning Vector Quantization

Abstract

This research aims to build a system for the classification of papaya maturity level using Learning Vector Quantization. The classification process is done by the colour feature extraction value. Forty-five images consist of 30 images for training data and 15 images for test data were used. The images were divided into 3 classes: rip, mature and raw. The parameters for classification are mean, skewness, and kurtosis. Test results 1 obtained an accuracy of 60% consisting of 9 true images and 6 incorrect images with hidden layer 5 and learning rate 0,1. Test results 2 obtained an accuracy of 66,67% consisting of 10 true images and 5 incorrect images with hidden layer 10 and learning rate 0,5. Test image data are 15 papaya images consisting of 5 mature images, 5 imperfect images, and 5 raw images.