KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#dsh #dsh-plugin #developer-tools #ai-agent #claude-code #codex #typescript #ai-agents #deepseek-harness #mcp #cli #rust

WaterQuality

★ 55 Open GitHub ↗

An explainable water quality classification model

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# WaterQuality

This study aims to classify water resources, predict their safety, and explain the trained model.
XGBoost is used to construct the model and predict the water datasets. 
SHAP is used to explain the model.
  
    
![image](https://github.com/yuntech-bdrc/WaterQuality/blob/main/image/Screenshot%202024-09-22%20134306.png)

## Requirement  
 
``` shell
pip install -r requirement.txt
```

## Performance
| Dataset | Accuracy | Precision | Recall | F1-score |
| :-- | :--: | :--: | :--: | :--: |
| water-potability | 0.77 | 0.73 | 0.66 | 0.7 |
| water-quality | 0.96 | 0.93 | 0.78 | 0.85 |

## Datasets
[water-potability](https://www.kaggle.com/datasets/adityakadiwal/water-potability "kaggle_water_potability")  
[water-quality](https://www.kaggle.com/datasets/mssmartypants/water-quality "kaggle_water_quality")

## References

[SHAP](https://arxiv.org/abs/1705.07874 "A Unified Approach to Interpreting Model Predictions")  
[TreeSHAP](https://arxiv.org/abs/1802.03888 "Consistent Individualized Feature Attribution for Tree Ensembles")  
[XGBoost](https://arxiv.org/abs/1603.02754 "XGBoost: A Scalable Tree Boosting System")

WaterQuality

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README