Graciello, Manuel Tanbica (2025) Optimizing Nglegena Javanese Script Recognition with Zoning and Feature Normalization using Machine Learning / Manuel Tanbica Graciello</p>. Diploma thesis, Universitas Negeri Malang.
Full text not available from this repository.Abstract
p This study investigates the application of supervised machine learning algorithms mdash namely K-Nearest Neighbors (KNN) Na iuml ve Bayes Decision Tree and Random Forest mdash for the classification of handwritten Javanese Nglegena script which is vital for the preservation of Indonesia rsquo s cultural heritage. Feature extraction was conducted using a zoning technique wherein each character image was partitioned into multiple zones (16 25 36 and 64) to capture localized structural information. Subsequently the extracted features were normalized utilizing Min-Max Z-Score and Binary normalization methods to ensure uniform data distribution and enhance model performance. The dataset comprised 600 images of handwritten Nglegena characters which were divided into training and testing subsets under various partition ratios. Experimental results indicate that the Na iuml ve Bayes classifier combined with 36-zone feature extraction and either Min-Max or Z-Score normalization achieved the highest classification accuracy of 65%. These findings underscore the significance of optimizing both zoning granularity and normalization techniques to improve the efficacy of machine learning models in recognizing Javanese script. This research contributes to the advancement of Optical Character Recognition (OCR) technologies specific to Javanese script thereby supporting the digital preservation and accessibility of Indonesia rsquo s historical manuscripts and cultural artifacts. The study provides a foundation for future work in computational heritage conservation and the development of culturally informed machine learning applications. /p
| Item Type: | Thesis (Diploma) |
|---|---|
| Divisions: | Fakultas Teknik (FT) > Departemen Teknik Elektro (TE) > S1 Teknik Informatika |
| Depositing User: | library UM |
| Date Deposited: | 08 Aug 2025 04:29 |
| Last Modified: | 09 Sep 2025 03:00 |
| URI: | http://repository.um.ac.id/id/eprint/400153 |
Actions (login required)
 |
View Item |