SemetonBug - Next-Generation Machine Learning-Powered Code Analyzer for Precision Bug Detection and Dynamic Error Localization

Aug 3, 2026

Upload by Ophelia

SemetonBug delivers an automated approach to Python bug detection using machine learning and dynamic error localization. The system relies on a Random Forest Classifier trained on features derived from code syntax via Abstract Syntax Tree (AST) representations. A dataset of 200 Python files (100 buggy, 100 non-buggy) is used, and the model is tuned with Grid Search Cross Validation, reaching 85% accuracy, 0.84 precision, 0.87 recall, and 0.86 F1-score. Detected bugs are exported for further analysis and quality improvement.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Abstract

Introduction

Background and challenges in Python bug detection

Prior work and related tools

SemetonBug: Next-Generation Machine Learning-Powered Code Analyzer for Precision Bug Detection and Dynamic Error Localization Surni Erniwati 1*, Bahtiar Imran 2**, Zumratul Muahidin 3***, Zaeniah 4***, Juhartini 5**** * Manajemen Informatika, Universitas Teknologi Mataram ** Rekayasa Sistem Komputer, Universitas Teknologi Mataram *** Sistem Informasi, Universitas Teknologi Mataram **** Teknik Informatika, Universitas Teknologi Mataram [mentari1990@gmail.com](mentari1990@gmail.com1)[1](mentari1990@gmail.com1), [bahtiarimranlombok@gmail.com](bahtiarimranlombok@gmail.com2)[2](bahtiarimranlombok@gmail.com2), [muahidinzumratul@gmail.com](muahidinzumratul@gmail.com3)[3](muahidinzumratul@gmail.com3), [zaen1989@gmail.com](zaen1989@gmail.com4)[4](zaen1989@gmail.com4), [j](juhartini8815@gmail.com5)[uhartini8815@gmail.com](juhartini8815@gmail.com5)[5](juhartini8815@gmail.com5) | Article history:
Received 2025-11-24 Revised 2025-12-22 Accepted 2026-01-07 | Bug detection in Python programming is a crucial challenge in software development. This research proposes SemetonBug, a machine learning-based system for automatically detecting bugs in Python code. The system utilizes a Random Forest Classifier as the main model, with features extracted from the syntactic structure of the code using an Abstract Syntax Tree (AST). The dataset consists of 200 Python files, divided into 100 files with bugs and 100 files without bugs. The model is optimized using Grid Search Cross Validation, with the best combination of n_estimators = 300, max_depth = 20, min_samples_split = 5, and min_samples_leaf = 2. Evaluation results show that the model achieves 85% accuracy, 0.84 precision, 0.87 recall, and 0.86 F1-score. The detected bugs are stored in an Excel file for further analysis. By leveraging machine learning, SemetonBug enhances efficiency and accuracy in bug identification compared to traditional rulebased methods. These findings highlight the potential of machine learning models in improving software quality and reducing coding errors automatically.

This is an open access article under the CC–BY-SA license. | | --- | --- | | Keyword:
Bug Detection, Machine Learning, Python,
Random Forest, Abstract Syntax Tree. | | Article Info ABSTRACT I. INTRODUCTION Bug detection in Python programming is one of the most critical challenges in software development. Various studies have demonstrated that bugs can be minimized through appropriate tools and techniques. For instance, PYBUGLAB, an implementation designed for the Python language, is developed to detect and fix various types of simple bugs that significantly impact code accuracy. By focusing on simple bugs, this tool illustrates that minor corrections can lead to substantial improvements in overall code quality, reducing errors and enhancing software reliability [1] . Empirical analysis of code modifications after bug fixes in Python reveals specific patterns, indicating that changes are not made randomly but occur within a particular coding context [2]. This finding underscores the importance of understanding code context for further advancements in bug detection tools. In the modern era, advanced techniques are increasingly employed, including machine learning models for bug detection and duplicate bug report identification. Research indicates that attention-based models can improve accuracy in detecting duplicate bug reports, which often pose challenges in large-scale software management [3] . In software development, manual bug identification is often timeconsuming, prone to human error, and inefficient, particularly when dealing with large-scale codebases. This manual approach relies on programmers' expertise to review and test code directly, which may result in undetected bugs or issues only discovered during the final testing phase. Therefore, automated methods that enable faster, more accurate, and consistent bug detection are essential. Several previous stu

FAQs

What is SemetonBug and what problem does it address?

SemetonBug is a machine learning-based system that automatically detects bugs in Python code. It targets the limitations of manual and rule-based bug identification by improving speed and detection quality.

How does SemetonBug represent Python code for model training?

SemetonBug extracts features from the syntactic structure of the code using an Abstract Syntax Tree (AST). These AST-derived features feed a Random Forest Classifier for classification.

What dataset and evaluation results does SemetonBug report?

The study uses 200 Python files split into 100 buggy and 100 non-buggy samples. The optimized model achieves 85% accuracy, with 0.84 precision, 0.87 recall, and 0.86 F1-score.

Aug 3, 2026

Upload by Ophelia