MARS Evaluation Report

Feature risk and stability analysis with source-aware tables, grouped bin comparisons, and native risk trend charts.

Features: 1
PSI > 0.25 的特征（任一标签触发）: 0
Trend Metrics: 7
Group By: dataset_flag
Jumps to the matching row in Summary. Charts are also searchable.
Data Source
Export uses Summary expression + Data Source only. Global and local searches only affect display.
Dataset ContextBusiness MetadataSummaryMissing By DayTrend TablesGrouped PivotCharts
Dataset Context
Dataset context, grouping setup, and target-level baseline stats.
Rows
356255
Features
1
Profile By
dataset_flag
Groups
5
Time Range
2026-01-01 ~ 2026-07-31
Targets
TARGET
Event Rate
TARGET: 8.07%
Business Metadata

FeatureMetadata

feature display_name dtype data_source description unit
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv How many days before the application the person started current employment 天（相对申请日，保留原始符号）

BusinessContext

key value part parts
label_definition "unknown" 1 1
observation_window "unknown" 1 1
currency "unknown" 1 1
dataset_id "home-credit-default-risk-2018" 1 1
source "https://www.kaggle.com/competitions/home-credit-default-risk/data" 1 1
real_features {"source":"application_train.csv/application_test.csv","simulated":false,"derived_features_added":false} 1 1
target {"field":"TARGET","definition":"Target variable (1 - client with payment difficulties: he/she had late payment more than X days on at least one of the first Y installments of the loan in our sample, 0 - all other cases)","simulated":false,"available_in":"application_train.csv","unavailable_in":"application_test.csv","unobserved_value":null,"performance_window":"X/Y 未公开；不推断 bad30/bad60"} 1 1
date {"field":"APPLY_DATE","simulated":true,"train_range":["2026-01-01","2026-06-30"],"test_range":["2026-07-01","2026-07-31"],"limitation":"功能演示日期，不支持真实跨期泛化、季节性或线上漂移结论。"} 1 1
partitions {"field":"dataset_flag","simulated":true,"0train":"拟合、候选发现、开发筛选与 PSI 参考分布","1val":"验证或筛选；参与选择后不是最终独立评估","2test":"有标签独立评估，不等同于原始 application_test","3oot":"模拟时间外独立评估，不是真实业务 OOT","4_oot_undue":"原始比赛测试表，标签未公开；名称不代表真实贷款未到期"} 1 1
amount {"field":"AMT_CREDIT","unit":"原始金额单位","currency":"未公开","meaning":"贷款金额，不等于损失或利润","included_as_feature":false} 1 1
weights null 1 1
sample_scope "两个 application 文件的全部样本" 1 1
preparation_version "home-credit-2018-v1" 1 1
special_values {"DAYS_EMPLOYED":{"value":365243,"count":64648,"handling":"保留原值；该异常正值作为本字段的特殊值配置，不应用于其他字段。","meaning":"官方字段说明未解释此值，不推断实际就业时长。"}} 1 1
labels {"TARGET":{"definition":"Target variable (1 - client with payment difficulties: he/she had late payment more than X days on at least one of the first Y installments of the loan in our sample, 0 - all other cases)","positive_class":"unknown","negative_class":"unknown","performance_window":"X/Y 未公开；不推断 bad30/bad60","field":"TARGET","simulated":false,"available_in":"application_train.csv","unavailable_in":"application_test.csv","unobserved_value":null}} 1 1
special_value_processing {"DAYS_EMPLOYED":[365243]} 1 1

ReportSemantics

key value part parts
report_type "MarsBinningReport" 1 1
report_id "55ac7f0c-2773-4f52-9085-75415347e4be" 1 1
format_version 1 1 1
source {"kind":"analysis","producer":"mars","input_provenance":"caller_provided_dataframe"} 1 1
context_source {"business_context":{"label_definition":"unknown","observation_window":"unknown","currency":"unknown","dataset_id":"user_provided","source":"user_provided","real_features":"user_provided","target":"user_provided","date":"user_provided","partitions":"user_provided","amount":"user_provided","weights":"user_provided","sample_scope":"user_provided","preparation_version":"user_provided","special_values":"user_provided","labels":"user_provided","special_value_processing":"user_provided"},"parameters":"calculation"} 1 1
serialization {"non_finite":"tagged float: {$mars: float, value: nan/inf/-inf}","date":"ISO-8601","null":"missing; calculation state is separate"} 1 1
tables {"summary":{"rows":1,"index":null,"grain":"feature","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"iv":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"ks":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"auc":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"psi_max":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数；跨有效分组最大值"},"rc_min":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数；跨有效分组最小值"},"lift_min":{"dtype":"Float64","unit":"dimensionless","meaning":"Total 正常箱 Lift 的最小值"},"lift_max":{"dtype":"Float64","unit":"dimensionless","meaning":"Total 正常箱 Lift 的最大值"},"missing":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"missing_min":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值；跨有效分组最小值"},"missing_max":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值；跨有效分组最大值"},"mono":{"dtype":"Float64","unit":"correlation","meaning":"Total 正常箱索引与坏率的 Spearman 相关系数；单箱或坏率恒定时无法定义"}}},"detail":{"rows":36,"index":null,"grain":"target/feature/group/bin","fields":{"y":{"dtype":"String","unit":"identifier","meaning":"目标列标识；真实目标及无标签状态见参数"},"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"trend":{"dtype":"String","unit":"category","meaning":"Total 正常箱 WOE 形态；无标签时未计算"},"mars_group":{"dtype":"String","unit":"dimension","meaning":"实际计算分组字段；取值保留原始分组标识"},"bin_index":{"dtype":"Int16","unit":"index","meaning":"分箱规则索引；缺失和特殊箱为负，汇总行由 bin_type 标识"},"bin_label":{"dtype":"String","unit":"category","meaning":"来自拟合分箱规则的箱标签"},"count":{"dtype":"Float64","unit":"count","meaning":"全样本数；配置 weights_col 时为样本权重和"},"observed_count":{"dtype":"Float64","unit":"count","meaning":"已表现标签样本数或权重和；无标签时未计算"},"bad":{"dtype":"Float64","unit":"count","meaning":"已表现坏样本数或权重和"},"good":{"dtype":"Float64","unit":"count","meaning":"observed_count-bad"},"pct":{"dtype":"Float64","unit":"ratio","meaning":"箱样本量/分组全样本量，配置权重时为权重和之比"},"bad_rate":{"dtype":"Float64","unit":"ratio","meaning":"坏人数/有表现人数；加权时为坏权重/有表现权重，见 parameters"},"lift":{"dtype":"Float64","unit":"dimensionless","meaning":"箱坏率/总体坏率，复用稳定化规则"},"cum_count":{"dtype":"Float64","unit":"count","meaning":"按分箱索引展示顺序的累计全样本量或权重和"},"cum_observed_count":{"dtype":"Float64","unit":"count","meaning":"按分箱索引展示顺序的累计已表现样本量或权重和"},"cum_bad":{"dtype":"Float64","unit":"count","meaning":"按分箱索引展示顺序的累计坏样本量或权重和"},"cum_bad_rate":{"dtype":"Float64","unit":"ratio","meaning":"累计 bad/累计 observed_count"},"psi_bin":{"dtype":"Float64","unit":"dimensionless","meaning":"当前箱 PSI 贡献；缺失箱和特殊箱由参数控制"},"ks_bin":{"dtype":"Float64","unit":"points_0_100","meaning":"当前箱累计坏样本与好样本分布绝对差 ×100"},"auc_bin":{"dtype":"Float64","unit":"ratio","meaning":"分箱 ROC 梯形面积贡献；汇总行进行 AUC 归一"},"iv_bin":{"dtype":"Float32","unit":"dimensionless","meaning":"(bad_dist-good_dist)×WOE"},"total_count":{"dtype":"Float64","unit":"count","meaning":"分组全样本数或权重和"},"bin_type":{"dtype":"String","unit":"category","meaning":"正常、缺失、特殊或汇总箱类型"},"tot_amt":{"dtype":"Float64","unit":"unknown_business_unit","meaning":"全样本金金额和；业务币种未知"},"good_amt":{"dtype":"Float64","unit":"unknown_business_unit","meaning":"好样本金金额和；业务币种未知"},"bad_amt":{"dtype":"Float64","unit":"unknown_business_unit","meaning":"坏样本金金额和；业务币种未知"},"avg_amt":{"dtype":"Float64","unit":"unknown_business_unit","meaning":"tot_amt/count；配置样本权重时 count 为权重和；业务币种未知"},"amt_bad_rate":{"dtype":"Float64","unit":"ratio","meaning":"bad_amt/observed_amt；未表现金额不计入分母，分母非正时为空"},"lift_amt":{"dtype":"Float64","unit":"dimensionless","meaning":"箱金额坏率/总体金额坏率，复用稳定化规则"}}},"trend.psi":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"},"1val":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"},"2test":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"},"3oot":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"},"4_oot_undue":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"},"Total":{"dtype":"Float64","unit":"dimensionless","meaning":"各箱 (actual-expected)×ln(actual/expected) 之和；参考及箱范围见参数"}},"target":"TARGET"},"trend.auc":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"1val":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"2test":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"3oot":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"4_oot_undue":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"},"Total":{"dtype":"Float64","unit":"ratio","meaning":"按配置排序的分箱 ROC 梯形面积；报告归一到不小于 0.5"}},"target":"TARGET"},"trend.ks":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"1val":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"2test":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"3oot":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"4_oot_undue":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"},"Total":{"dtype":"Float64","unit":"points_0_100","meaning":"按实际配置排序的分箱累计坏样本与好样本分布最大绝对差 ×100"}},"target":"TARGET"},"trend.iv":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"1val":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"2test":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"3oot":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"4_oot_undue":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"},"Total":{"dtype":"Float32","unit":"dimensionless","meaning":"各箱 (bad_dist-good_dist)×WOE 之和，复用稳定化规则"}},"target":"TARGET"},"trend.missing":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"1val":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"2test":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"3oot":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"4_oot_undue":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"},"Total":{"dtype":"Float64","unit":"ratio","meaning":"缺失箱样本量/全样本量；画像包括 Null、NaN 和配置缺失值"}},"target":"TARGET"},"trend.lift":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"},"1val":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"},"2test":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"},"3oot":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"},"4_oot_undue":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"},"Total":{"dtype":"Float64","unit":"dimensionless","meaning":"当前分组箱 Lift 的最大值；未表现分组为空"}},"target":"TARGET"},"trend.risk_corr":{"rows":1,"index":null,"grain":"feature (groups in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"0train":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"},"1val":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"},"2test":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"},"3oot":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"},"4_oot_undue":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"},"Total":{"dtype":"Float64","unit":"correlation","meaning":"正常箱坏率与 RC 参考坏率的 Spearman 相关系数，参考来源见参数"}},"target":"TARGET"},"calculation_status":{"rows":7,"index":null,"grain":"target/feature/group/metric","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"target":{"dtype":"String","unit":"identifier","meaning":"真实目标列标识"},"group":{"dtype":"String","unit":"dimension","meaning":"状态所属分组；Total 为全量"},"metric":{"dtype":"String","unit":"identifier","meaning":"计算状态适用的指标或指标族"},"status":{"dtype":"String","unit":"state","meaning":"computed 已计算（有效零仍为零）；not_computed 未计算；unobserved 无观测标签；undefined 无法定义；insufficient_samples 样本不足；failed 计算失败；skipped 跳过；结合 reason"},"reason":{"dtype":"String","unit":"state_reason","meaning":"no_target 未提供目标；no_observed_labels 无已观测标签；single_observed_class 仅一种观测类别；其他失败、跳过、样本不足或无法定义原因；空值不替代有效零"}}},"missing_by_day":{"rows":1,"index":null,"grain":"feature (days in columns)","fields":{"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"2026-01-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-30":{"dtype":"Float64","unit":"ratio","meani 1 3
tables ng":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-01-31":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-02-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-30":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-03-31":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-04-30":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-30":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-05-31":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不 2 3
tables 使用样本权重，total 为全量比例"},"2026-06-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-06-30":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-01":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-02":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-03":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-04":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-05":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-06":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-07":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-08":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-09":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-10":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-11":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-12":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-13":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-14":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-15":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-16":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-17":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-18":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-19":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-20":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-21":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-22":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-23":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-24":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-25":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-26":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-27":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-28":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-29":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-30":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"2026-07-31":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"},"total":{"dtype":"Float64","unit":"ratio","meaning":"自然日原始缺失值数/当日样本数，含 Null、NaN 和配置缺失值；不使用样本权重，total 为全量比例"}}},"risk_corr_reference":{"rows":4,"index":null,"grain":"target/feature/bin/reference","fields":{"y":{"dtype":"String","unit":"identifier","meaning":"目标列标识；真实目标及无标签状态见参数"},"feature":{"dtype":"String","unit":"identifier","meaning":"原输入特征名称"},"display_name":{"dtype":"String","unit":"category","meaning":"调用方校验后的特征可读名称；缺失可在展示时回退原始 ID"},"dtype":{"dtype":"String","unit":"type","meaning":"原输入字段的数据类型"},"data_source":{"dtype":"String","unit":"category","meaning":"调用方登记的特征来源；未映射字段可能标为 UNMAPPED"},"bin_index":{"dtype":"Int16","unit":"index","meaning":"分箱规则索引；缺失和特殊箱为负，汇总行由 bin_type 标识"},"base_br":{"dtype":"Float64","unit":"ratio","meaning":"RC 参考箱的 bad/observed_count"},"source":{"dtype":"String","unit":"category","meaning":"已保存的参考来源，例如 total、first_group 或 benchmark_df"}}}} 3 3
calculation_state {"table":"calculation_status","diagnostics":"parameters.diagnostics/fit_failures/raw_ks_diagnostics","scope":"feature/target/group/metric; profile status contains exceptional outcomes only, not all trend cells"} 1 1
parameters {"row_count":356255,"benchmark_row_count":215374,"binning_fit_source":"benchmark_df","features":["DAYS_EMPLOYED"],"weights_col":null,"batch_size":100,"ordered_metric_sort_by":"woe","binning_config":{"cart_params":{},"join_threshold":100,"merge_small_bins":false,"method":"quantile","min_bin_size":0.05,"missing_values":[],"n_bins":4,"n_jobs":1,"remove_empty_bins":false,"special_values":[365243]},"fit_failures":{},"binning_fit_report":[{"feature":"DAYS_EMPLOYED","dtype":"Int64","feature_type":"numeric","status":"fitted","usable":true,"n_bins":4,"reason":null}],"feature_data_source":{"DAYS_EMPLOYED":"Home Credit 2018 application_{train|test}.csv"},"time_grain":"month","target_requested":"TARGET","has_target":true,"target_available":true,"feature_count":1,"profile_by_input":"dataset_flag","group_col":"mars_group","group_count":5,"feature_dtypes":{"DAYS_EMPLOYED":"Int64"},"dt_col":"APPLY_DATE","start_dt":"2026-01-01","end_dt":"2026-07-31","targets":["TARGET"],"event_rate_by_target":{"TARGET":0.08072881945686496},"feature_start_aware_reference":false,"psi_include_missing":false,"psi_include_special":false,"risk_corr_baseline":"total","risk_corr_reference_source":"total","amount_col":"AMT_CREDIT","feature_start_reference_features":[],"feature_start_reference_dates":{},"group_context":{"input_group_col":"dataset_flag","internal_group_col":"mars_group","groups":[{"internal_value":"0train","display_value":"0train","start_dt":"2026-01-01","end_dt":"2026-05-07"},{"internal_value":"1val","display_value":"1val","start_dt":"2026-05-08","end_dt":"2026-05-25"},{"internal_value":"2test","display_value":"2test","start_dt":"2026-05-26","end_dt":"2026-06-12"},{"internal_value":"3oot","display_value":"3oot","start_dt":"2026-06-13","end_dt":"2026-06-30"},{"internal_value":"4_oot_undue","display_value":"4_oot_undue","start_dt":"2026-07-01","end_dt":"2026-07-31"}],"total_range":{"start_dt":"2026-01-01","end_dt":"2026-07-31"}}} 1 1
limitations ["只组织已有统计，未计算表不在目录中；失败和跳过见 parameters.diagnostics / fit_failures。","Null 不是数值 0；风险指标需结合标签表现状态，样本分布使用全样本。","比较前核对样本范围、权重、缺失/特殊箱、拟合来源、排序口径及参考来源；关联不代表因果。"] 1 1

CalculationStatus

feature display_name dtype data_source target group metric status reason
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET Total risk_metrics computed NaN
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET 0train risk_metrics computed NaN
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET 1val risk_metrics computed NaN
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET 2test risk_metrics computed NaN
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET 3oot risk_metrics computed NaN
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET 4_oot_undue risk_metrics unobserved no_observed_labels
DAYS_EMPLOYED 开始当前工作相对申请的天数 Int64 Home Credit 2018 application_{train|test}.csv TARGET Total mono computed NaN
Summary
Feature-level ranking and monitoring summary.
Available metrics: iv, ks, auc, psi_max, rc_min, lift_min, lift_max, missing, missing_min, missing_max.
Supported operators: >, >=, <, <=, ==, !=, &, |, ( ). Use & for AND and | for OR.
IV 0.01 / 0.05 / 0.10, purple above 0.20KS 4 / 8 / 12, purple above 16AUC 0.525 / 0.55 / 0.575, purple above 0.625PSI 0 / 0.1 / 0.25Missing 0 / 0.5 / 1Lift Min 0.5 / 0.6 / 0.7 / 0.8, purple at or below 0.5Lift Max 1.2 / 1.3 / 1.4, purple above 1.5
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.1114.480.590.011.000.671.390.00%0.00%0.00%
Missing Trend By Day
Daily missing-rate trend derived from dt_col=APPLY_DATE.
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%0.00%
Trend Tables
Trend tables with local search, default Total ranking, and Excel-like color scales.
Missing 0 / 0.5 / 1PSI 0 / 0.1 / 0.25IV 0.01 / 0.05 / 0.10, purple above 0.20KS 4 / 8 / 12, purple above 16Lift 1.2 / 1.3 / 1.4, purple above 1.5AUC 0.525 / 0.55 / 0.575, purple above 0.625Risk Corr 0.2 / 0.5 / 0.8
MISSING
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.00%0.00%0.00%0.00%0.00%0.00%
PSI
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.000.000.000.000.010.00
IV
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.110.100.100.120.11
KS
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv14.5713.8713.7615.2414.48
LIFT
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv1.391.421.341.421.39
AUC
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv0.590.590.580.600.59
RISK_CORR
DAYS_EMPLOYED开始当前工作相对申请的天数Int64Home Credit 2018 application_{train|test}.csv1.001.001.001.001.00
Grouped Pivot
Binned distribution and risk comparison across groups.
TARGET
FeatureBinpctriskliftbad counttotal countiv
Total0train1val2test3oot4_oot_undueTotal0train1val2test3oot4_oot_undueTotal0train1val2test3oot4_oot_undueTotal0train1val2test3oot4_oot_undueTotal0train1val2test3oot4_oot_undueTotal0train1val2test3oot4_oot_undue
DAYS_EMPLOYED00_[-inf, -3173)25.32%24.98%25.24%24.88%25.22%27.21%4.63%5.49%5.46%5.25%5.08%0.00%0.690.680.660.690.633419.002422.00346.00328.00323.000.0073832.0044138.006339.006253.006362.0010740.000.130.030.030.030.040.00
01_[-3173, -1646)25.08%24.99%25.52%24.79%24.74%25.62%6.75%7.81%8.24%7.48%7.96%0.00%0.990.960.990.980.994938.003447.00528.00466.00497.000.0073146.0044155.006409.006229.006241.0010112.000.000.000.000.000.000.00
02_[-1646, -765)25.06%25.02%24.77%25.17%25.19%25.30%8.79%10.27%10.05%9.69%10.18%0.00%1.271.271.211.271.266426.004541.00625.00613.00647.000.0073089.0044206.006221.006324.006353.009985.000.060.020.010.020.020.00
03_[-765, inf)24.53%25.00%24.46%25.17%24.85%21.87%9.86%11.24%11.80%10.20%11.46%0.00%1.421.391.421.341.427052.004964.00725.00645.00718.000.0071540.0044173.006143.006324.006267.008633.000.130.030.040.020.040.00
Charts
Risk trend charts rendered from the shared plotting path.

DAYS_EMPLOYED

HTML export is self-contained. detail_table remains available in Python and Excel workflows.