Rankers

A ranker defines documents in a search result through the use of learning-to-rank algorithms, which are a form of machine learning. Document ranking requires training data to learn where documents should be placed in the the search result. A ranker with a trained model can be deployed as a ranker instance, and the deployed ranker instance can be associated with one or more collections to modify their search results. There is one of type of ranker that you can create, the Similar Document Ranker ranker.

The Similar Document Ranker ranker accepts a document as a search query and returns a list of documents ranked by their expected similarities to the queried document. For more information, see Similar Document Ranker.

To create a ranker, click Add ranker and follow the instructions in Add a ranker.

After you create a ranker, you need to train it. For more information, see Training a ranker.

Supported languages

The following languages are supported by rankers.

• Arabic, Czech, Danish, German, English, Spanish, French, Hebrew, Italian, Japanese, Korean, Dutch, Polish, Portuguese, Russian, Slovak, Turkish, Chinese