CODEWITHSUNDEEP
pythonscikit-learn
MheadHero
MheadHero

Reputation: 153

AttributeError: 'TfidfVectorizer' object has no attribute 'get_feature_names_out'

Why do I keep on getting this error? I try other codes too, but once it uses the get_feature_names_out function it will pop out this error.

Below is my code:

from sklearn.datasets._twenty_newsgroups import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB  # fast to train and achieves a decent F-score
from sklearn import metrics
import numpy as np

def show_top10(classifier, vectorizer, categories):
    feature_names = vectorizer.get_feature_names_out()
    for i, category in enumerate(categories):
        top10 = np.argsort(classifier.coef_[i])[-10:]
        print("%s: %s" % (category, " ".join(feature_names[top10])))

newsgroups_train = fetch_20newsgroups(subset='train')
print(list(newsgroups_train.target_names))

cats = ['alt.atheism', 'sci.space', 'rec.sport.baseball', 'rec.sport.hockey']
newsgroups_train = fetch_20newsgroups(subset='train', categories=cats)
print(list(newsgroups_train.target_names))
print(newsgroups_train.filenames.shape)

vectorizer = TfidfVectorizer()
vectors = vectorizer.fit_transform(newsgroups_train.data)
print(vectors.shape)

Upvotes: 15

Views: 59062

Answers (2)

afghani
afghani

Reputation: 676

When sklearn.__version__ <= 0.24.x use following method

get_feature_names()

When sklearn.__version__ >= 1.0.x use following method

get_feature_names_out()

Ref:

  1. https://github.com/scikit-learn/scikit-learn/blob/0.24.X/sklearn/feature_extraction/text.py

Upvotes: 40

Arne
Arne

Reputation: 10545

This is probably because you are using an older scikit-learn version than the one this code was written for.

get_feature_names_out is a method of the class sklearn.feature_extraction.text.TfidfVectorizer since scikit-learn 1.0. Previously, there was a similar method called get_feature_names.

So you should update your scikit-learn package, or use the old method (not recommended).

Upvotes: 14

Related Questions