2
$\begingroup$

My dataset looks like this

Sport_Type       City         Report_Text                                                 Labels
Ball             Toronto      Messi has been announced the best soccer player...          Soccer
Swimming         London       Todays new records in Butterfly Stroke & Backstroke...      Butterfly Swimming, Backstroke Swimming, Front Crawl
Ball             Chicago      Tennis and basketball along with football has...            Tennis, basketball, Soccer
Fighting         Sydney       Todays matches include boxing, judo, and...                 Boxing, Judo, Karate
Horse            Melbourne    Melbourne Cup is the race that stops the nation...          Horse Racing

I can build multi label model to identify labels in each Report_Text field.

but is there a way I can consider Sport_Type and City field in my model as it will help in improving results.

How can I use other features such as Sport_Type and City in NLP multi label model?

$\endgroup$
3
  • 1
    $\begingroup$ Sure, you can add these two features to the features you obtain for the text. $\endgroup$
    – Erwan
    Commented Aug 9, 2019 at 12:02
  • $\begingroup$ @Erwan to the LSTM Neural Network? $\endgroup$
    – asmgx
    Commented Aug 9, 2019 at 13:16
  • $\begingroup$ If you're working with LSTM specifically you should edit your question to mention it. Hopefully somebody more expert than me can answer. $\endgroup$
    – Erwan
    Commented Aug 9, 2019 at 14:50

1 Answer 1

0
$\begingroup$

Sport_Type and City features are categorical features so they need to be encoded into a numeric format (e.g., one-hot encoding or feature hashing). Those numerical features can be added to any machine learning model, including Long Short Term Memory (LSTM).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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