# Regression

> 执行线性回归并解释关键变量影响（linear regression 回归分析）

- Skill: `zafer-liu/regression` (Agent Skill)
- Install (CLI): `npx skillmds add zafer-liu/regression`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zafer-liu/regression/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: Zafer-Liu (https://skillmd.com/u/zafer-liu)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/zafer-liu/regression

---

# 线性回归

确认连续目标、解释变量和分析口径，检查缺失、异常、共线性与残差假设。报告拟合与验证指标、系数方向和不确定性；明确相关关系不等于因果关系。

## Tool routing

1. Use `get_schema` to identify the continuous target, candidate predictors, and source table.
2. Use `query_data` to verify field names, missingness, outliers, and whether the target is numeric.
3. Use `run_analysis` with `analysis_name="Regression"` for the actual regression computation.
4. Use `generate_chart` on coefficient, residual, or prediction result tables after `run_analysis` succeeds.

## Implementation reference

- Tool entry: `agent/tools/business/data.py::_tool_run_analysis`
- Analysis registry: `Function/Analyze/registry.py`
- Analysis implementation: `Function/Analyze/Regression/analyze.py`
- Chart implementation: `Function/Charts_generation/chart_generate.py`