Back to results
US 8,266,157 B2Granted
Method and system for using social bookmarks
AssigneeInternational Business Machines Corporation
InventorsDavid Carmel, Haggai Roitman, Elad Yom-Tov
Life of this patent
Estimating the term…
Abstract
A method and system for using social bookmarks wherein a social bookmark is a triplet of the entities of user, document, and tag. The method including: collecting multiple bookmarks; representing the bookmarks as a three-dimensional space or matrix of the number of times a user u, used tag t to bookmark document d; measuring the similarity of two entities of the same type; and using the similarity to weight bookmarks or entities. The weightings may be used to provide a measure of a usefulness of a bookmark for describing a document for retrieval purposes. Two-dimensions of the bookmark space may also be used to predict the third-dimension.
Claims24
A method for using social bookmarks, the method comprising:
- collecting multiple social bookmarks, wherein each social bookmark is a triplet of entities, the triplet having three entity types comprising user, document, and tag;
- representing the collected social bookmarks as a three-dimensional space or matrix of a number of times a user u, used tag t to bookmark document d;
- measuring respective similarities of two entities having a common entity type in the collected social bookmarks;
- using the similarities to weight the entities by using two-dimensions of the three-dimensional space to predict a third dimension thereof, wherein predicting the third dimension comprises defining two-dimensional matrices for each pair of (u,t), (d,t) and (d,u);
- computing values of the entities in one of the two-dimensional matrices as respective vectors of weights of the other two-dimensional matrices;
- ranking the computed values;
- and
- recommending of at least a portion of the entities in the one of the two-dimensional matrices according to the ranking, wherein any of said steps are implemented in either of computer hardware or computer software embodied in a non-transitory computer-readable medium.
5 more elements
A computer program product
- including a non-transitory computer-readable storage medium in which computer program instructions are stored, which instructions, when executed by a computer, cause the computer to:
- collect multiple social bookmarks, wherein each social bookmark is a triplet of entities, the triplet having three entity types comprising user, document, and tag;
- represent the collected social bookmarks as a three-dimensional space or matrix of a number of times a user u, used tag t to bookmark document d;
- measure respective similarities of two entities having a common entity type in the collected social bookmarks;
- use the similarities to weight the entities by using two-dimensions of the three-dimensional space to predict a third dimension thereof, wherein predicting the third dimension comprises defining two-dimensional matrices for each pair of (u,t), (d,t) and (d,u);
- computing values of the entities in one of the two-dimensional matrices as respective vectors of weights of the other two-dimensional matrices;
- ranking the computed values;
- and
- recommending of at least a portion of the entities in the one of the two-dimensional matrices according to the ranking, wherein said program instructions are stored on said computer readable medium.
6 more elements
A system for using social bookmarks comprising:
- a collector for collecting multiple social bookmarks from one or more bookmarking servers, wherein each social bookmark is a triplet of entities, the triplet having three entity types comprising user, document, and tag;
- a storage medium for storing the collected social bookmarks as a three-dimensional space or matrix of a number of times a user u, used tag t to bookmark document d;
- a measuring component for measuring respective similarities of two entities having a common entity type in the collected social bookmarks;
- a weighting component for weighting the entities by using two-dimensions of the three-dimensional space to predict a third dimension thereof, wherein predicting the third dimension comprises defining two-dimensional matrices for each pair of (u,t), (d,t) and (d,u);
- a scoring component for computing values of the entities in one of the two-dimensional matrices as respective vectors of weights of the other two-dimensional matrices;
- ranking the computed values;
- and
- a recommendation component operative for recommending of at least a portion of the entities in the one of the two-dimensional matrices according to the ranking,;
- wherein any of said collector, storage medium, measuring component, scoring component and weighting component are implemented in either of computer hardware or computer software embodied in a non-transitory computer readable medium.
6 more elements
Dependencies are inferred from the claim text.