Similar domains

At Kodex we have created our own vectorization model that uses AI to find the domains that have most similarities in the name and categories. In the similar section you are shown first 8 similar domains, but if you want to discover all 100 similar domains, that the vectorization model has chose, press the "see more" card at the end of the list.

How it works

The vectorization modal we use at kodex uses a lightweight library, that allows users to learn text representations and text classifiers, fastText.

Using this library we at kodex have spent a large amount of time creating a model from all the domains we have stored in our indexer. It currently finds similar domains just by trying to find best possible match to the domain of the name, though in the future we are planning to also include the category of the domain to optimize the search performance.

To learn more about how fastText works under the hood and what logic they use in their documentation.