GEM-example

This repository provides the code and example data of the paper "Gradient Boosting-accelerated Evolution for Multiple-Fault Diagnosis"

Usage

1. To run the first stage of GEM , you need to go to the work folder, with circuit.bench、circuit.faults、circuit.test and the corresponding fault simulation results. Then simply execute

./GEM-stage1 circuit begin end

where begin and end are used to limit the range of fault simulation results. Subsequently, the folder train_and_test was generated, which contains the csv files extracted from the fail log for training the HGB model.

2. To train HGB model and make prediction, execute

python training.py
python predicting.py

3. To run the second stage of GEM, execute

./GEM-stage2 circuit begin end

where begin and end are used to limit the range of fault simulation results. The folder x_ans is then generated, where the .out file is the final diagnostic result.

Environment

· gcc 11.4.0 Ubuntu 22.04

· x86-64

· python 3.10.12