How to train a classifier using list of list feature Python

Question

I have two variables called entity and label. The entity variable store list of word, each element in this list contain list as well. So it is a list of list variable. This list actually a bi-gram feature so I need to keep it.

I try to train a classifier using this two variables. This my code so far:

from sklearn import svm
from sklearn.feature_extraction.text import TfidfVectorizer

entity = [[['Prabowo Subianto']], [['Muhtar Ependi']], [['Nina Zatulini']], [['Partai Gerindra']], [['Persiba']], [['Partai Kebangkitan Bangsa (PKB)'], ['Partai Kebangkitan'], ['Kebangkitan Bangsa'], ['Bangsa ('], ['( PKB'], ['PKB )']], [['Sman 3 Kabupaten Tangerang'], ['Sman 3'], ['3 Kabupaten'], ['Kabupaten Tangerang']], [['Bandara Changi Singapura'], ['Bandara Changi'], ['Changi Singapura']], [['Warung Kopi Kita'], ['Warung Kopi'], ['Kopi Kita']]]
label = ['PERSON', 'PERSON', 'PERSON', 'ORGANIZATION', 'ORGANIZATION', 'ORGANIZATION', 'LOCATION', 'LOCATION', 'LOCATION']

vectorizer = TfidfVectorizer(min_df=1)
train_vector_entity = vectorizer.fit_transform(entity)
train_vector_label = label

classifier = svm.SVC()
classifier_word = classifier.fit(train_vector_entity,train_vector_label)

The error result:

AttributeError: 'list' object has no attribute 'lower'

What is the best way to train the classifier? Thanks

What does the list of list represent? I mean does it have any significance of keeping in list? Cant you iterate through it to make it a single list which can then be passed to TfidfVectorizer — Vivek Kumar
– Vivek Kumar, Commented Aug 15, 2017 at 7:03

Stergios · Accepted Answer · 2017-08-16 07:09:02Z

0

Just change this row:

train_vector_entity = vectorizer.fit_transform([i[0][0] for i in entity])

answered Aug 16, 2017 at 7:09

Stergios

3,2166 gold badges36 silver badges55 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

How to train a classifier using list of list feature Python

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related