explain

app.py

@@ -232,9 +232,10 @@ Full dataset at the bottom of this tab
 
 Explain by Context
 ===============
-Below are explaination in typical background E[f(x)]
+- Below are explanation in typical background E[f(x)]
 
-Sometime it is useful to switch to credit healthy background, to explain why a certain person default by changing the baseline E[f(x) | credit healthy] with interventional feature perturbation
+- Sometime it is useful to switch to credit healthy background, to explain why a certain person default by changing the baseline E[f(x) | credit healthy] with interventional feature perturbation
+https://arxiv.org/pdf/2006.16234.pdf
  
 Explain by Dataset
 ===============
@@ -281,19 +282,20 @@ Explain by Top 5 Error Example
 ![Top 5 Error Data](file=./xgb/error_data.png)
 
 **Top Features for Errors:**
-- **Age** stands out as the top feature impacting the top 5 errors negatively (for young ages).
+- **dist_subway, age** stands out as the top feature impacting the top 5 errors negatively (for young ages).
 
 ![Error Record](file=./xgb/error_record.png)
 **Top 1 Error:**
 - Notably, young age has a negative impact on pricing (top 1 error).
+- lat has positive impact
 
 ![Error Feature](file=./xgb/error_feature.png)
 **Insight from Errors:**
-- Further distance from the subway might positively impact pricing for the top 5 errors.
+- Further distance from the subway might positively impact pricing for the top 5 errors at around 700
 
 ![Error Instance](file=./xgb/error_instance.png)
 **Error Instances:**
-- Younger age negatively impacts price, while older age positively impacts it for the top 5 errors.
+- Younger age negatively impacts price, while older age positively impacts price for the top 5 errors.
 
 ML Observability
 ===============