Klasifikasi Logo Mobil Menggunakan Jaringan Syaraf Tiruan Backpropagation dan Decision Tree

Abstract

The car logo itself is very distinguishing in a car is a vehicle logo that serves to introduce to the public about their brand. This will also create its own appeal to the public in knowing the existing car logo. Types in study aims to classify for car logos in Indonesia. So that later people will understand in choosing a car with a quality brand. From the results obtained that the Decision Tree at split ratio 50:50 precision gets a value of 0.604, recall gets a value of 0.611, f-measure gets a value of 0.598 and accuracy gets a value of 95.70% at comparing data testing and training 50:50. Then the tests carried out by ANNbackpropagation resulted in a split ratio of 50:50 texture and shape features with a precision value reaching 0.680, recall getting a value of 0.521, f-measurement getting a value of 0.600 and accuracy also having the highest value generated by ANNbackpropagation reaching 92.50% at comparing data testing and training 50:50. The results prove that the classification using Decision Tree produces the highest accuracy, precision, recall, and f-measure compared to the decision tree.