300961 Lecture Notes - Lecture 11: Social Computing
Document Summary
Also known as social matching (the term is proposed by terveen and mcdonald), people recommender systems deal with recommending people to people on social media. Those factors are related to types of relationships among people on social networking sites, such as symmetric vs asymmetric, ad-hoc vs long-term, and con rmed vs noncon rmed relationships. The scope of people recommender systems can be categorized into three: recommending familiar people to connect with, recommending people to follow and recommending strangers. Recommending strangers is seen as valuable as recommending familiar people because of leading to chances such as exchanging ideas, obtaining new opportunities, and increasing one"s reputation. Handling with social streams is one of the challenges social recommender systems face with. social stream can be described as the user activity data pooled on newsfeed on social media websites. Social stream data has unique characteristics such as rapid ow, variety of data (only text content vs heterogenous content), and requiring freshness.
