DecisionTreeClassifier Question
16 views
Skip to first unread message
Cristiano Coelho
May 13, 2017, 5:00:38 PM5/13/17
to nltk-users
Hello there,
When using a DecisionTreeClassifier, I'm wondering if there's a way to override the default label when a feature value is not found. Right now I believe it tries to guess based on frequency, but that's not ideal.
For example, I'm using DecisionTreeClassifier to classify some data that a user marks as valid or invalid, one of the features can be a "user name", so at some point of the tree there are a bunch of fixed rules (if user == ... return True/False) but since the data is mostly classified as valid, when data with an unknown user is classified it is always treaten as a True rule, but it should be false.
When using a DecisionTreeClassifier, I'm wondering if there's a way to override the default label when a feature value is not found. Right now I believe it tries to guess based on frequency, but that's not ideal.
For example, I'm using DecisionTreeClassifier to classify some data that a user marks as valid or invalid, one of the features can be a "user name", so at some point of the tree there are a bunch of fixed rules (if user == ... return True/False) but since the data is mostly classified as valid, when data with an unknown user is classified it is always treaten as a True rule, but it should be false.
Reply all
Reply to author
Forward
0 new messages