Supervised multi-label text analytics in R

I want to implement supervised multi-label text analytics (for use in opinion mining) using a review dataset.

Which package should I use for this?
Can someone suggest step-wise guide/case-study/code in R?

I tried using 'mlr' which gave following error:
*Error in (function (cn, x) : *

  • Unsupported feature type (character) in column 'text'.*

Hi, @Arv welcome to the forums.
As dar as I know, I would recommend you to use Keras integration with R, or Rweka.
In addition, below I am sharing very related resources with you, hope this helps.

https://stackabuse.com/python-for-nlp-multi-label-text-classification-with-keras
Hope this helps.

Thanks @AndyR .

The tutorials helped me a lot.
I worked around mldr objects using the RWeka package.
I am a new user, hence, it took me some time to solve the java related issues. But, I was able to do it.

I was trying to evaluate the model performance. As per my learning, mldr uses predict() and mldr_evaluate() function for this task. The predict() function here is giving me labels-output as a single factor level value instead of dividing it into a data frame with labels as different columns. While the mldr_evaluate() take data frame as input.

Example: For the five-label dataset, the row-wise output is '01001', '10010', etc., instead of dividing into 5 columns.

Is there any way to get label output in different columns or should I go for manual rule-based functions?

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.