I have a set of a few thousand organisations (org) with some attributes (name, location, and type) and I want to identify the official URL for each organisation from a set of 10 URLs retrieved from Google.

The features to characterise org/URL pairs may include a few different string-matching scores, URL length, top-level domain, etc. Given the project constraints, I can generate a training/test set of correct pairs up to 100 or 150.

This can be seen as a binary classification problem between an organisation and a set of URL (true=official website). Another approach could be learning to rank, where the algorithm ranks the URLs and I pick the top one.

What algorithm/method would you recommend for this scenario, considering the small training set?


Your Answer

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

Browse other questions tagged or ask your own question.