The proliferation of computers as handheld devices with Internet connectivity along with ecommerce and social networking sites allow the generation of huge amount of data. This data empowers the corporations and other organizations to produce meaningful business patterns from consumers' behavior. Also, they can build recommender systems to predict future social trends which can enhance their services and improve their products. For example, The recommendation systems used by companies such as Amazon, Google News, and Netflix utilize Collaborative Filtering techniques such as k-nearest neighbor (kNN) to discover what their users like and dislike. Using kNN, the system compares a primary user with all others and determines how similar their interests are to the primary user's. Doing so creates a neighborhood list, consisting of every user's similarity to the primary user. Using this list, it is easy to determine the primary user's most similar, or nearest neighbor. This nearest neighbor will then provide the basis for the primary user's recommendations. In this research, we present a realistic method to process large data sets collected from Internet for recommending bookmarks by using kNN in a variation of Collaborative Filtering called One-Class Collaborative Filtering (OCCF).
展开▼