I'm doing a student project where I construct a model predicting the number of languages that a given Wikipedia article is translated into (for example, the article TOYOTA is translated into 93 languages). I've tried extracting basic info (article length, number of links, etc.) to create a simple regression model, but can't get the $R^2$ value above $0.25$ or so.

What's the most appropriate NLP algorithm for regression problems? Almost all examples I find online are classification problems. FYI I'm aware of the basics of NLP preprocessing (tokenization, lemmatization, bag of words, etc).

  • $\begingroup$ Random Forest Classifier, RNN, etc. Check out this NLP collection. I am sure you will get your answers $\endgroup$
    – Arpit-Gole
    Oct 5 '20 at 11:35
  • $\begingroup$ One thing that might not be apparent at the first glance when starting with ML is that many algorithms could be used for more than just classification tasks. As you say, many deep learning models, for example, are used for classification tasks most of the time. However, when changing the loss function, (possibly) the set-up of the output layer, and training a network to predict continuous values, you can turn a deep learning model into a regression model as well. So, often what you get out of a model only depends on how you configure it. $\endgroup$
    – Daniel B.
    Oct 6 '20 at 12:23

I think it's difficult to tell wich algorithm is "the best" or "the simplest".

I had the same issue of choosing the suited NLP algorithm for my dataset and I used :


Then I recommanded you to test many algorithms as you can to find the best for your needed.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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