0
$\begingroup$

Which among KNN, Logistic and Naive Bayes would yield best results for SMS spam
detection? Is there any other efficient approach worth exploring.
I am planning to make a python application for SMS spam detection.
Any suggestions or resources would be great.

$\endgroup$
0
$\begingroup$

It totally depends on what sort of feature engineering you use. Except for the case of KNN which is useless. Naive Bayes will work well with Bag of Words and TF-IDF while Logistic regression will perform well on all including Word2Vec.

$\endgroup$
3
  • $\begingroup$ Could you give me some insight into feature engineering. Any resources. What it seems to me is that you have to decide the features of an SMS. I am not sure if there are any existing data sets like that. Also I am unable to think of any good features apart from the SMS itself or SMS containing a specific word $\endgroup$
    – nanu
    Jul 14 at 18:42
  • $\begingroup$ are there any existing libraries or github repo which does SMS spam detection? $\endgroup$
    – nanu
    Jul 14 at 18:43
  • $\begingroup$ Although you are new but try to get your head around this - huggingface.co/transformers/custom_datasets.html. This is almost State of the Art for Sequence Classification. OR just google it, look on youtube. you will find tons of tutorails. Upvote if it helped. $\endgroup$ Jul 14 at 19:41

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.