Doubt in some lines of code

[Sid10 (SIDDHARTH NAIR)]

Hello.Why are these lines of code used in the decision tree implementation code
if data_left.shape[0] == 0 or data_right.shape[0] ==0:
if X_train.Survived.mean() >= 0.5:
self.target = “Survive”
else:
self.target = “Dead”
return
#Stop earyly when depth >=max depth
especially why is if X_train.Survived.mean() >= 0.5: used?

[prashant_ml (Prashant Arora)]

hey @Sid10 ,
here i provide you the explanation of each line in code you provided :

@Sid10

if data_left.shape[0] == 0 or data_right.shape[0] ==0:

after splitting each node depending , if we see that on any side sub branch we get no node ,so that side branch will result in self.target value depending on the mean value.

@Sid10

if X_train.Survived.mean() >= 0.5:

Like in logistic regression we provide a threshold to classify , similarly in binary classification task we have taken this threshold to be as 0.5 and if the mean value is less than 0.5 we take target as dead else survived.

@Sid10

Stop earyly when depth >=max depth

This is used to control over fitting. Stopping the model when our validation accuracy starts to decrease.

I hope this would have resolved your doubt.
Thank You and Happy Learning .