Piotr Kozikowski was one of our star interns during the summer of 2008 and we are excited to welcome him back as a full-time employee starting next week. Piotr recently published his Master’s thesis on data inference, in conjunction with Rapleaf. Here is the abstract of his very well-written paper. Congratulations Piotr!
Although social networking websites have existed for well over a decade, it is only recently (since approx. 2005) that they have become immensely popular. Today, social networking websites play a very important role in the lives of millions of people.
As the user base of social networking sites grows at a rapid pace, so do the opportunities to take advantage of the vast amount of data that becomes available as a result.
The purpose of this thesis is to use data gathered from social networking sites to develop techniques to infer unknown attributes of individuals based on other, known attributes, as well as attributes of people with whom the individuals have a relationship.
The main focus of the thesis is the inference of gender, age, and location. Other attributes are also analyzed, but in much less detail. Additionally, the possibility of assigning weights to the relations among individuals to provide a measure of friendship closeness is also explored.
The data analyzed does not come from one specific social networking website, but rather is composed of publicly available information coming from a number of such websites. It is hoped that this data is representative of the social web in general and, therefore, any findings will also be applicable to the social web in general.
One Comment
great addition to the Rapleaf team!