Classification of traditional ulos of Batak Toba ssing probabilistic neural network

Abstract

Ulos are commonly used in traditional Batak events. Ulos has many variations of pattern and the pattern almost similar to one another. Ulos also have different functions in its use. So, people are often wrong to call the name of Ulos because the similarity of its pattern. Therefore, the research is doing Ulos classification to increase the introduction of Batak Toba’s Ulos. In this research, the method used in classification is Probabilistic Neural Network method. The stages before classification are pre-processing, extraction feature using Gray Level Co-Occurrence Matrix, and then classified using Probabilistic Neural Network. The research uses 650 data to be used as training data and test data. After testing in this research, it was concluded that the Probabilistic Neural Network method had the ability to classify Ulos into five patterns namely Bintang Maratur, Ragi Hidup, Ragi Hotang, Mangiring and Sibolang. The result showed that the accuracy level is 80%.