Experimental Study on Zoning, Histogram, and Structural Methods to Classify Sundanese Characters from Handwriting

Abstract

Sundanese characters are one of the original Sundanese historical relics that have existed since the 5th century and have become the writing language at that time. Classification of handwriting characters is a challenge because the results of handwriting are very diverse, including the characters of handwritten characters. The number of feature extraction methods that can be used in the classification process, but not all feature extraction methods are in accordance with the characteristics of the Sundanese characters. Therefore, the focus of this research is to find the optimal feature extraction method to classify the character of Sundanese characters, in order to get better accuracy by running some experiments. Feature extraction methods proposed in this research are zoning, histograms and structural approaches. Then, some following classifier methods are used for constructing models and prediction over new data: Random Forest (RF), K-Nearest Neighbor (KNN), Artificial Neural Network (ANN), and Support Vector Machine (SVM). Based on the experiments, we can state that RF provided the best results (i.e., 89.84% in average) while the optimal feature-constructing method is by using the structural approach.