| Summary: | With the rise of Web 2.0, and a more interactive Internet, a greater emphasis has been placed on making services better tailored for the end user; "local" searches for places near the user, or "profiled" searches for music the user likes. At the same time, people have embraced the possibility of geo-tagging their posts and images: associating geographic information with them to convey a meaning or opinion about that specific place. While it is already possible to search through the posts and images organized such, there has been a distinct lack of a Web 2.0-style personalized search, capable of finding the geo-tagged posts and images that a specific user is interested in. This thesis investigates how to identify a user’s interests, using a combination of traditional techniques such as folksonomy tagging and profiling, as well as making use of the geo-location data now available in mobile phones. This data is then used to locate and present to the user the posts that they want to view, both based on their interests, the current location, and the combination of interests at the current location. Along with the proposed solution, a prototype of it has been developed, testing the feasibility of the solution from an implementation perspective. Furthermore, this prototype has permitted the evaluation of the solution’s ability to power mashups – combining two or more existing services, in turn creating a new or unique way of presenting the services’ input or output.
|
|---|
