0
$\begingroup$

The problem is to identify the fake news detection, As this is text classification problem . Constraints are basically that we cannot use traditional machine learning and deep learning approaches.

If we move towards Machine learning then we can easily sort out this problem, using Naive Bayes or Logistic Regression etc but we cannot use this. I want to take yours suggestion that using cosine similarity can we done this , Take out the text, apply the feature embedding techniques like Glove or Word2vec or Doc2vec and returns it similarity score. Is it possible , if not kindly suggest your solutions ?

$\endgroup$
4
  • $\begingroup$ Fake news detection is a very hard problem, see this related question. The constraint not to use any ML is strange, btw the techniques you mention belong to ML. But the most important question is: do you have any labelled data? I don't think you can achieve anything good with unsupervised methods on this. $\endgroup$
    – Erwan
    Nov 17, 2021 at 13:28
  • $\begingroup$ @Erwan Thanks , Can you suggest some approach or technique . Is it possible with taking out of embeddings using Word2Vec etc and then apply cosine similarity etc ?. Suggest your perceptions on this problem as this is the client requirements to not use supervised machine learning ? I do not have labelled data , if labelled then it is supervised and we cannot use supervise machine learning $\endgroup$
    – Hamza
    Nov 23, 2021 at 10:29
  • $\begingroup$ Honestly without any supervision I doubt this can be done. Similarity measures can only help if you know whether you're comparing to a fake or legitimate text, in other words it requires supervision as well. Practically your only option seems to be clustering or ranking, but you'd have to guess which features are useful for the task and that's impossible for a complex task like this. Additionally the lack of labelled data means that you have nothing to evaluate your method quantitatively. $\endgroup$
    – Erwan
    Nov 23, 2021 at 16:42
  • $\begingroup$ ML is not magic, it can't work well if there's no available information for the task. In my opinion this is what you should tell this client ;) $\endgroup$
    – Erwan
    Nov 23, 2021 at 16:42

0

Your Answer

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