Can LLM interpret system logs? - Part B
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The goal of this project is to develop an Intelligent Log Analysis Platform employing Large Language Model (LLM) capabilities to automate the interpretation and processing of system logs. The implemented solution addresses the limitations of traditional manual parsing — specifically low efficiency and high labor costs — by leveraging the Deepseek API to perform advanced semantic analysis, log summarization, and anomaly detection. The system architecture is established on a Browser/Server (B/S) model, utilizing the Flask framework for backend logic, MySQL for data persistence, and Bootstrap for the frontend interface. Deepseek LLM integration serves as the core intelligence engine, enabling the platform to extract key operational information and identify irregular patterns from unstructured text. Additionally, secure file management and interactive data visualization modules are implemented to create a cohesive and user-centric analytical environment. The proposed solution significantly enhances operational efficiency and reliability for system maintenance. It transforms raw log data into actionable insights, thereby streamlining troubleshooting processes and reducing the workload on operations personnel. This will be illustrated in the report through system architecture diagrams, use case figures, and implementation flowcharts, detailing the modular design and data interaction processes. The report will also include the functional testing and validation results carried out to demonstrate the feasibility and effectiveness of the solution in real world scenarios.
Fields of Study & Subjects
Software System Performance and Reliability
Software Engineering Research
Topic Modeling
Publication details
- Journal / source
- DR-NTU (Nanyang Technological University)
- Publication date
- 1 January 2026
- Language
- EN