User logs for query disambiguation

This lab sample is really a master’s project that involves solving (beginning to anyway) the problem of understanding user search queries. What does it actually mean when someone searches for java, is it coffee or code they want?

By looking at user search logs, one can analyze user behavior and begin to understand user intention. For instance, we can see which results and filters were clicked and opened, whilst connecting these to a certain query. Once user search logs have been analyzed, we have a basis to interpret vague queries. By adding words or substituting words in the query, we are more likely to present users with what they were looking for. This approach was tested and evaluated by Karnov Group (a large provider of legal content) and was proven very successful.

Made by Studio Interact