Automatic detection of navigational queries according to Behavioural Characteristics

This document summarizes research on automatically detecting navigational queries based on user behavioral characteristics. The researchers analyzed the AOL query log to develop navigational coefficients (NC) based on the number of visits to the most popular result, number of distinct results, and percentage of navigational sessions. The NC were highest for queries with most traffic to one result, few total results, and isolated to one session. Comparing the coefficients showed they identify navigational queries when combined but not individually. Future work could refine the session detection algorithm and combine NCs with other methods to better identify user intentions.