Selection of most frequent features for processed mtl dataset

ccsasuke / man

Multinomial Adversarial Networks for Multi-Domain Text Classification (NAACL 2018)

https://arxiv.org/abs/1802.05694

MIT License

80 stars 28 forks source link

Selection of most frequent features for processed mtl dataset #4

Closed Humanity123 closed 5 years ago

Humanity123 commented 5 years ago

The paper mentions selecting 5000 most common features for the processed amazon mtl dataset. However, the code selects the first 5000 features from the pickle files. Are the columns of the feature matrix sorted?

ccsasuke commented 5 years ago

Yes, if I remember correctly.