MuheCC

This is a repliacation package for Paper Boosting Commit Classification Based on Multivariate Mixed Features and Heterogeneous Classifier Selection.

Content

1. Get Started
     1.1 Requirements
     1.2 Dataset
   
2. Run
   

1 GetStarted

1.1 Requirements

• Packages:

  • bayesian_optimization==1.4.2
  • hyperopt==0.2.7
  • lightgbm==3.2.1
  • matplotlib==3.5.2
  • numpy==1.21.5
  • optuna==3.2.0
  • pandas==1.3.5
  • scikit_learn==1.2.2
  • torch==1.12.1
  • transformers==4.18.0

  You can see these in EnsembleLearning/requirements.txt.

1.2 Dataset

The dataset can be found in MuheCC_Dataset_with_refactoring_final.xlsx.

2 Run

1. Train and test three baselines

python EnsembleLearning/BERT+lightGBM.py

2. Train and test MuheCC

python new.py