Lobang Forever

You gotta trust the algorithm. After all, it what gets you those views, like and the sweet sweet ad revenue. But how do you trust something you barely understand? Its hard to make a system fair when you don't understand it.

This hackathon, we thought to ourselves: How might we make more transparent, fair revenue sharing models?

We trained ensemble regression models to score the quality of tiktok videos and evaluated the weight of the each video features on the score through SHAP and LIME analysis. This allows us to make more explainable and fair systems.

Built With

• fastapi
• ml
• scikit-learn