python 2.7 - How to combine multiple feature sets in bag of words -

i have text classification data predictions depending on categories, 'descriptions' , 'components'. classification using bag of words in python scikit on 'descriptions'. want predictions using both categories in bag of words weights individual feature sets x = descriptions + 2* components how should proceed?

you can train individual classifiers descriptions , merchants, , obtain final score using score = w1 * predictions + w2 * components.

the values of w1 , w2 should obtained using cross validation.

alternatively, can train single multiclass classifier combining training dataset.

you have 4 classes:

1. neither 'predictions' nor 'components'
2. 'predictions' not 'components'
3. not 'predictions' 'components'
4. 'predictions' , 'components'

and can go ahead , train usual.