I have created a pipeline object which uses countvectorizer and MultinomialNB. MultinomialNB instances have an attributes named intercept_. If I try to access this attribute using the Pipeline object I am getting the error that 'Pipeline' object has no attribute 'intercept_'. I am wondering if there is a way to access attributes of the final estimator class in a Pipeline instance. I would like to thank in advance.
Asked
Active
Viewed 582 times
1 Answers
0
Yes, using either the steps
or named_steps
attributes, see this question: Getting model attributes from scikit-learn pipeline

Community
- 1
- 1

Andreas Mueller
- 27,470
- 8
- 62
- 74