diagnose_regression
Diagnose regression model issues like multicollinearity, heteroscedasticity, and autocorrelation to validate statistical assumptions and improve analysis accuracy.
Instructions
회귀분석 진단 가이드 (다중공선성, 이분산, 자기상관 등)
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| issues | Yes | 의심되는 문제 (multicollinearity, heteroscedasticity, autocorrelation, outliers) | |
| model_type | No | 모형 유형 |
Implementation Reference
- src/tools/index.ts:1324-1351 (handler)Handler function that takes issues array and returns diagnostic tests, code snippets, and solutions for common regression problems like multicollinearity, heteroscedasticity, and autocorrelation.function handleDiagnoseRegression(args: Record<string, unknown>) { const issues = args.issues as string[]; const diagnostics: Record<string, any> = { multicollinearity: { tests: ["VIF (>10 문제)", "Condition Number (>30 문제)", "Correlation Matrix"], r_code: "car::vif(model)", stata_code: "vif", solution: "변수 제거, PCA, Ridge regression" }, heteroscedasticity: { tests: ["Breusch-Pagan test", "White test", "Residual plot"], r_code: "lmtest::bptest(model)", stata_code: "hettest", solution: "Robust SE, WLS, Log transformation" }, autocorrelation: { tests: ["Durbin-Watson", "Breusch-Godfrey", "ACF plot"], r_code: "lmtest::dwtest(model)", stata_code: "dwstat", solution: "HAC SE, GLS, ARIMA errors" } }; return { issues, diagnostics: issues.map(issue => diagnostics[issue] || { issue, message: "진단 정보 검색 필요" }) }; }
- src/tools/index.ts:201-215 (registration)Registration of the 'diagnose_regression' tool in the main tools array, including name, description, and input schema.name: "diagnose_regression", description: "회귀분석 진단 가이드 (다중공선성, 이분산, 자기상관 등)", inputSchema: { type: "object", properties: { issues: { type: "array", items: { type: "string" }, description: "의심되는 문제 (multicollinearity, heteroscedasticity, autocorrelation, outliers)" }, model_type: { type: "string", description: "모형 유형" }, }, required: ["issues"], }, },
- src/tools/index.ts:203-214 (schema)Input schema defining the expected parameters: issues (array of strings) required, model_type optional.inputSchema: { type: "object", properties: { issues: { type: "array", items: { type: "string" }, description: "의심되는 문제 (multicollinearity, heteroscedasticity, autocorrelation, outliers)" }, model_type: { type: "string", description: "모형 유형" }, }, required: ["issues"], },
- src/tools/index.ts:808-809 (registration)Switch case in handleToolCall that routes calls to the diagnose_regression handler.case "diagnose_regression": return handleDiagnoseRegression(args);