An Analysis of Classification Method Performance on Handwritten Lontara Numerals

The research investigates the performance of various classification methods on handwritten Lontara digits, a script used by the Bugis and Makassar communities in South Sulawesi, Indonesia. The dataset comprises 10,890 samples from 99 individuals, categorized into 10 classes (digits 0-9). The study employs the K-Nearest Neighbors (KNN), Gaussian Naive Bayes (GNB), and Nu-Support Vector Classifier (NuSVC) algorithms, implementing cross-validation to assess accuracy, precision, recall, and F1 score. The results indicate varying performance across classifiers, with GNB showing the highest recall, while KNN and NuSVC display moderate effectiveness. The study concludes with recommendations for further improving classification accuracy through enhanced feature extraction and algorithm optimization.