System and method for recommending items of interest to a user

摘要: A system and method is disclosed for recommending items to individual users using a combination of clustering decision trees frequency-based term mapping. The the present invention configured receive data based on user action, such as television remote control activity, or computer keyboard entry, when new item made available from sources program guides, movie databases, deliverers advertising data, on-line auction web sites, electronic mail servers, analytically breaks down compares it ascertained attributes that liked in past, produces numeric ranking dynamically, without subsequent input, manipulation by deliverers, tailored each user. embodiment learning interests actions then applying learned knowledge rank, recommend, and/or filter items, e-mail spam, level interest may be used automated personalized information learning, recommendation, filtering systems applications programming, web-based auctions, targeted advertising, filtering. structured generate descriptions, learn interest, terms effectively describe cluster similar compact structure, use structure rank offerings. Embodiments include, way non-limiting example: allowing assignment scores candidate so one can recommended over another, building incrementally unsupervised examples into categories automatically, consistency with 'edge' (thick client) computing whereby certain structures most processing are localized set-top box local PC, ability content automatically on-the-fly, store preferences opaque not easily traceable users.