User Modelling and Feature Selection for Personalized Local Search
Search engines, such as Google, have revolutionized the way we search for electronic information, providing a user with a ranked list of documents most relevant for a particular query. This project with GenieKnows R&D, a search engine company, concerns an extension of this basic technology, in which the goal is to incorporate geographic constraints into the search (e.g. find coffee shops near the Halifax Citadel). An additional challenge is to learn and incorporate user preferences when ranking the results returned. Thus a search for “restaurants in downtown Halifax” might return Thai, Sushi, and Chinese restaurants if the user was known to prefer Asian food.