0
$\begingroup$

I need to apply data binning on a set of reviews, I have searched for some data binning methods for reviews and long-texts and couldn't find anything other than classification. Is NLP or classification the only method to bin long-text data?

$\endgroup$
0
$\begingroup$

You can bin on Len of the data. you can also bin on Keywords you are interested in. You can use Countvectorizer with YellowBricks library to arrive at Frequency Plot. You can tokenize into N grams and find the frequencies. Depends on your problem statement that you are trying to solve. You can do Sentiment Analysis on each of the reviews and bin them into positive negative and neutral. There are many ways you can think off.

I hope this helps.

| improve this answer | |
$\endgroup$
0
$\begingroup$

Another option is doing some topic modeling on the text (can work, based on their length) to extract topics, and use these as one hot features. If you use LDA or similar methods you will have to set up the number of topics yourself, which could work if you already have an idea of what you want. Alternatively, you can use Hierarchical Dirichelet Allocation, which will generate as many topics as it finds - and if you manage to "understand" them you can cluster them manually.
Hope it helps!

| improve this answer | |
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.