31-08-2017, 09:12 AM
Collaborative filtering (CF) is a technique used by recommendation systems. Collaborative filtering has two meanings, one narrow and one more general. In the more recent, narrower sense, collaborative filtering is a method for making automatic predictions (filtering) on a user's interests by collecting preferences or tasting information from many users. collaborating). The underlying assumption of the collaborative filtering approach is that if a person A has the same opinion as a person B in a subject, A is more likely to have the opinion of B on a different subject than a randomly chosen person. For example, a collaborative filtering recommendation system for television tastes could make predictions about which television program a user should want a partial list of the tastes of that user (like it or not). Keep in mind that these predictions are user-specific, but use information obtained from many users. This differs from the simpler approach of giving an average (non-specific) score for each item of interest, for example based on its number of votes.
In the most general sense, collaborative filtering is the process of filtering information or patterns using techniques that involve collaboration between multiple agents, viewpoints, data sources, and so on. Collaborative filtering applications typically involve very large datasets. Collaborative filtering methods have been applied to many different types of data, including: detection and monitoring data, such as mineral exploration, large-area or multi-sensor environmental detection; financial data, such as financial services institutions that integrate many financial sources; or in e-commerce and web applications, where the focus is on user data, etc. The rest of this discussion focuses on collaborative filtering of user data, although some of the methods and approaches can also be applied to other important applications.
In the most general sense, collaborative filtering is the process of filtering information or patterns using techniques that involve collaboration between multiple agents, viewpoints, data sources, and so on. Collaborative filtering applications typically involve very large datasets. Collaborative filtering methods have been applied to many different types of data, including: detection and monitoring data, such as mineral exploration, large-area or multi-sensor environmental detection; financial data, such as financial services institutions that integrate many financial sources; or in e-commerce and web applications, where the focus is on user data, etc. The rest of this discussion focuses on collaborative filtering of user data, although some of the methods and approaches can also be applied to other important applications.