1 year ago
#293079
Sara Ghaderi
Catboost Python: How to interpret the MAE and R2 values
I have fitted a regression model using Catboost. I want to predict tax values. The target variable which I am trying to predict is:
count 192687.00000
mean 94280.45840
std 1546991.34422
min -514999.00000
25% 7159.00000
50% 19601.00000
75% 50589.50000
max 353125221.00000
Name: tax_value, dtype: float64I have fitted two models. Model 1 has fewer variables than model 2:
Model 1:
MSE: 2355128439744.5
MAE: 81851.6
Train r2 score: 0.26
Test r2 score: 0.23
Train RMSE : 1553001.7
Test RMSE : 1534642.7
Mode 2:
MSE: 5374477074472.9
MAE: 110471.2
Train r2 score: 0.56
Test r2 score: 0.50
Train RMSE: 1988204.3
Test RMSE: 2318291.8 I can see that the R2 in model 2 is much better than in model 1. However I am unsure how to interpret the MAE values. Shouldn't the MAE in Model 2 be smaller?
I appreciate any help or input!
python
regression
metrics
evaluation
catboost
0 Answers
Your Answer