Image Classification of Starlings Using Artificial Neural Network and Decision Tree

Abstract

Starlings are famous animals in Indonesia. Therefore, many in Indonesia maintain and cultivate starlings. Almost every region in Indonesia has different types of starlings. Therefore, the researchers used Artificial Neural Networks and Decision Trees to classify starlings. Both methods are useful for obtaining the accuracy values generated in the classification of the starlings. In this comparison, the Artificial Neural Network has a precision of 0.870, the highest recall value is 0.600, the f-measure is 0.865, and the accuracy is 93% at a split ratio of 90:10. The Decision Tree has resulted in the classification of starlings on features, shapes, and colours with the highest texture value at a precision of 1,000, recall reaching 1,000, f-measure reaching 1,000, and accuracy reaching 100% at a split ratio 90:10. The tests carried out show that the Decision Tree can classify starling images based on 3 feature levels. And in this case, it can be proven that the Decision Tree is more accurate in classifying starlings images. The method of this Decision Tree can make it easier to find the right accuracy value in classifying starling species.