Skip to content

MengqinShen/log_classification_nlp_LLMs_vs_hybrid_models

Repository files navigation

Comparing Fine-tuned BERT and Hybrid Classification Model for Log Classification

This project was to implements a log classification system to category log information and send out alert if needed. A Bert model was first fine-tuned to do the task and another approach was done by combining three complementary approaches (Regular Expression (Regex), Sentence Transformer + Logistic Regression and LLM (Large Language Models)) to handle varying levels of complexity in log patterns. The focus was about the accuracy and latency of this two models regarding flexibility and effectiveness in processing predictable, complex, and poorly-labeled data patterns.

  1. data : check data patterns (grouping/clustering) - DB scan algorithm
  2. model : fine-tuned BERT vs hybrid system
  3. model evaluation comparision

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages