martinakaduc / xNeuSM

Explainable Neural Subgraph Matching with Graph Learnable Multi-hop Attention Networks
MIT License
13 stars 0 forks source link

The data in this degradation and performance were obtained through what kind of experiment? #3

Open wujiaaaaa opened 3 weeks ago

wujiaaaaa commented 3 weeks ago

481dac9c5a4197d3f5b2dca6e0dabd9d c499aba21f421ab6c8a6cb85e20e0309

I'm glad to hear that you find my writing helpful! I'd be more than happy to assist you with any questions you have regarding the code. Please feel free to ask, and I'll do my best to help you out. I appreciate your gratitude in advance!

martinakaduc commented 3 weeks ago

I have a note on how to create data in https://github.com/martinakaduc/xNeuSM/blob/main/data_real/README.md Hope that helps.

wujiaaaaa commented 3 weeks ago

Hello, what I want to ask is how the data in the performance file and the degradation file under the results file were obtained. For example, xNeuSM's ROU AUC is 0.983, PR AUC is 0.974, F1 is 0.983 and so on. By what experiment were these data obtained? Thank you

---- Replied Message ---- | From | Duc Quang @.> | | Date | 11/07/2024 19:12 | | To | martinakaduc/xNeuSM @.> | | Cc | wujiaaaaa @.>, Author @.> | | Subject | Re: [martinakaduc/xNeuSM] The data in this degradation and performance were obtained through what kind of experiment? (Issue #3) |

I have a note on how to create data in https://github.com/martinakaduc/xNeuSM/blob/main/data_real/README.md Hope that helps.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

martinakaduc commented 3 weeks ago

Hi, sorry for late reply, I am a little bit busy these days.

Firstly, you might need to re-train the models using train.py. Next, you can run evaluate.py, which will save all metrics for performance and degradation results. After that, I manually copy values to these files. Though it is not too convenient, it would be better if you could help me by writing a script to read the results and construct these files.

You can find all training and evaluation scripts in the scripts folder.