tree

# 决策树分析 确认目标变量和任务类型，排除泄漏字段，合理处理缺失值与类别变量。执行训练验证并报告性能、重要特征和可解释规则；避免把相关性表述为因果。 ## Tool routing 1. Use `get_schema` to identify the target, candidate features, task type, and source table. 2. Use `query_data` to verify field names, target distribution, missingness, and leakage risks. 3. Use `run_analysis` with `analysis_name="Decision_Tree"` for the actual tree computation. 4. Use `generate_chart` on tree, feature importance, or validation 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/Decision_Tree/analyze.py` - Chart implementation: `Function/Charts_generation/chart_generate.py`