Whitepaper: Collection Management
Collection Manager Demo
Yoogli Patent Information
- Strategic White Papers:
ENABLING TECHNOLOGY
COLLECTION MANAGEMENT
DIGITAL MUSIC APPLICATION
CERTIFIED CLICKS
RESEARCH
Collection Manager
Yoogli’s Collection Manager is designed for a community of users that assemble separate collections of data, each on a single topic. A big problem with the search models now being used to assemble these collections is that the context and content of the collection being built cannot be used to locate new resources that might belong - only keywords can be used in the user’s search query. No matter what the user is actually looking for, if a keyword that the user thinks will bring results appears on a completely unrelated page, that page will be returned anyway. The user is forced to manually classify pages by semantic domain before he can even begin the task of seeing if the page is related to the current collection of interest and useful to add to that collection.
Even when the collection is complete, it stays resident on the user’s computer, being of no use to anyone else. It doesn’t matter if multiple users are researching the exact same topic - all of the work needed to complete that research must be done by each user independently. For each user that successfully builds a collection, there are many users who would be interested in having the same collection. Unfortunately, those users lack either the ability or the discipline to complete the required research tasks.
Yoogli’s semantic matching technology has enabled the creation of a Collection Manager application that possesses none of these flaws. Users can quickly locate URL resources that are semantically aligned with their current collection, and that collection can be effective in such a search even if it only contains one URL! Collections are stored and manipulated on a central server, so they can be made accessible (at the user’s option) to other people that might be interested in the same topic. Yoogli's algorithms provide a measure of user similarity, allowing users to contact other people or make themselves available for such contacts from people who have created specific types of collections. This semantic partitioning of a social network is quite revolutionary when compared to the unfiltered (or attribute-filtered) social networking applications of today. Even anonymously constructed collections can be returned as the result of a standard keyword query by linking the context of each returned URL with the most similar collections in the database.