The present disclosure describes computer-implemented systems and methods for improving image-based search results by using result click data and/or similar metadata to rank or order results so as to reduce search time and provide more relevant search results. In one particular embodiment, the count of clicks for a result URL may be used to train an search result ranking model for ranking image search results. For a set of search results, the trained model can be used to determine a probability of a user selecting a particular search result in response to a query. The search results may be ranked or ordered based on the probability of a user selecting each search result.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.