bayesian_guide
Guide Bayesian analysis by selecting priors, running convergence diagnostics, and interpreting results for statistical modeling.
Instructions
베이지안 분석 가이드 (사전분포, 수렴진단, 해석)
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| analysis_type | Yes | 분석 유형 | |
| prior_type | No | 사전분포 | |
| convergence | No | 수렴진단 포함 |
Implementation Reference
- src/tools/index.ts:1915-1922 (handler)The handler function that executes the bayesian_guide tool. It returns guidance on priors, convergence checks, and sample R code using brms.
function handleBayesianGuide(args: Record<string, unknown>) { return { analysis_type: args.analysis_type, prior_type: args.prior_type, convergence_checks: ["Rhat < 1.01", "ESS > 400", "Trace plots", "Posterior predictive checks"], r_code: "library(brms)\nfit <- brm(y ~ x1 + x2, data = df, family = gaussian())" }; } - src/tools/index.ts:705-716 (registration)The tool registration in the exported tools array, including name, description, and input schema.
name: "bayesian_guide", description: "베이지안 분석 가이드 (사전분포, 수렴진단, 해석)", inputSchema: { type: "object", properties: { analysis_type: { type: "string", description: "분석 유형" }, prior_type: { type: "string", enum: ["uninformative", "weakly_informative", "informative"], description: "사전분포" }, convergence: { type: "boolean", description: "수렴진단 포함" }, }, required: ["analysis_type"], }, }, - src/tools/index.ts:707-715 (schema)The input schema defining parameters for the bayesian_guide tool: analysis_type (required), prior_type, convergence.
inputSchema: { type: "object", properties: { analysis_type: { type: "string", description: "분석 유형" }, prior_type: { type: "string", enum: ["uninformative", "weakly_informative", "informative"], description: "사전분포" }, convergence: { type: "boolean", description: "수렴진단 포함" }, }, required: ["analysis_type"], }, - src/tools/index.ts:889-890 (registration)The switch case in handleToolCall that routes calls to the bayesian_guide handler.
return handleBayesianGuide(args); case "ml_for_causal":