PinnedXwang·Feb 2, 2025SHAP Value Dilution: How XGBoost Feature Sampling MisleadsInterpreting machine learning models is critical in high-stakes applications like finance, healthcare, and fraud detection. SHAP (SHapley…A response icon3A response icon3
Xwang·Jul 17DS Snapshot: Making Data Science Work Easier to Archive and ReviewI recently built DS Snapshot, a tool for archiving and managing evidence from analytical and technical work.A response icon1A response icon1
Xwang·Jun 24I Built a Local Context Generator to Help AI Agents Save TokensAI agents are powerful, but their context is not free.
Xwang·May 3, 2025Why More Control Can Hurt: Understanding Bad Controls Through Bayesian NetworksWhether evaluating a new product or a regulatory change, getting the causal structure right is essential. To satisfy the conditional…
Xwang·Feb 28, 2025Demystifying Gibbs Sampling: Ergodicity, Detailed Balance, and ConvergenceMarkov Chain Monte Carlo (MCMC) methods, such as Gibbs sampling, are widely used in Bayesian inference and probabilistic modeling. However…
Xwang·Feb 8, 2025Feature Importance Dilution in XGBoost: Just Like SHAP or Different?Gain-based feature importance is another popular metric to evaluate in machine learning models for both variable reduction and model…
Xwang·Jan 28, 2025No Hype, Just Results: DeepSeek Educated Me Better on Correlation and VIFIn the world of data science, having too many variables can be both a blessing and a curse. Doing variable reduction helps to improve model…
Xwang·Jan 26, 2025How XGBoost Handles Missing Values? A Proof of ConceptIntroduction:A response icon1A response icon1
Xwang·Jan 6, 2023Causal Inference 102 EP04: Implementation of Inverse Probability WeightingIn this post, I will demonstrate how to implement Inverse Probability Weighting (IPW) from scratch. For simplicity, I use the well-known…