Regression Analysis Skill for Holdout RMSE Comparisons
Fit a simple linear model to synthetic delivery data, compare it against a mean baseline, and review held-out residuals.
中文Install in Claude Code
Use Python, pandas, and scikit-learn with synthetic route_km and delivery_minutes columns and a fixed split.
- Get the complete `plugins/ai-ml/regression-analysis-tool/skills/performing-regression-analysis` directory from the pinned GitHub commit.
- Copy it into the target project's `.claude/skills/`. Check `.claude/skills/performing-regression-analysis/SKILL.md`, start Claude Code at the project root, and call `/performing-regression-analysis`.
Example task
Generate 100 samples with seed 7 using roughly 10 + 2 times distance plus small random noise. Hold out 20 rows and compare the mean baseline with linear regression.What to check
The report records both RMSE values, the fitted coefficient, and the test sample count. The same fixture and settings reproduce the comparison.
Original SKILL.md on GitHub · Updated
Frequently asked question
Does a positive coefficient establish the only cause of delivery time?
It describes an association in the fitted model. A causal claim requires an appropriate experiment or identification design.