PREDIKSI TINGKAT PELANGGARAN LALU LINTAS MENGGUNAKAN METODE LVQ
Traffic violations remain a concern in maintaining road safety and ensuring compliance with traffic regulations. Differences in the number of violations based on traffic enforcement operations, violation types, and vehicle categories indicate the need for data processing methods that can provide information about traffic violation levels. This study aims to apply the Learning Vector Quantization (LVQ) method to predict traffic violation levels using available historical data. The dataset consists of 10 aggregated observations processed through several stages, including data collection, categorical attribute encoding, normalization, data splitting, model training, and testing. Seven observations were used for training, while the remaining three were used for testing. The LVQ method determines the predicted class by calculating the distance between input data and the reference weight vectors representing each class. The testing results showed that two out of three observations were correctly classified, resulting in an accuracy of 66.67%. This study also developed a web-based application that supports data management, model training, testing, prediction, visualization, and reporting. The findings indicate that LVQ can be implemented to classify traffic violation levels into low, medium, and high categories. However, the limited dataset restricts the interpretation of the model's performance and does not provide sufficient evidence of its generalizability. This study is expected to serve as a foundation for further development of traffic violation analysis systems using larger datasets.
Authors
- Windy Winata
- Muhammad Yasin
Publication Details
- Journal
- JOURNAL OF SCIENCE AND SOCIAL RESEARCH
- Published
- 2026-11-13
- DOI
- https://doi.org/10.54314/madhda06
- Primary Topic
- Data Mining and Machine Learning Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00