Classification of handwritten Javanese script using random forest algorithm

The ability to read and write Javanese scripts is one of the most important competencies for students have in order preserve language as Indonesian cultures. In this study, we developed a predictive model 20 characters using random forest algorithm basis developing script learning media students. building model, used an extensive handwritten image dataset experimented with several different preprocessing methods, including conversion black-and-white, cropping, resizing, thinning, feature extraction histogram oriented gradients. From experiment, it can be seen that resulting able classify very accurately accuracy, precision, recall 97.7%.