Optimasi Klasifikasi Parasit Malaria Dengan Metode LVQ, SVM dan Backpropagation

Abstract

The use of the classification method affects the accuracy of the test results. The accuracy of the classification method is affected by the number of classes in the image. The number of classes and the amount of data should be considered when making decisions in choosing a classification method. This study used 600 data, which were divided into 510 training data and 90 test data. The number of classes tested is 12 classes with the number of initial features used by 22 features. The characteristics used in the test consist of shape characteristics and texture characteristics. The classification methods used in this study are LVQ, Backpropagation, and SVM. The data has 22 features or attributes that are the result of texture and shape feature extraction. Texture features are energy 0^o, energy 45^o, energy 90^o, energy 135^o, entropy 0^o, entropy 45^o, entropy 90^o, entropy 135^o, contrast 0^o, contrast 45^o, contrast 90^o, contrast 135^o, homogeneity 0⁰, homogeneity 45^o, homogeneity 90^o, homogeneity 135^o, correlation 0^o, Correlation 45^o, correlation 90^o, correlation 135^o, features of área and perimeter shape. The test results using the Backpropagation method obtained 89.7% results, using the LVQ method obtained 77.78% results, and the SVM method obtained 99.1% results.