| 摘要 |
In one embodiment, a personal information profile is constructed for a user having a collection of information items. The personal information profile comprises one or more information topics, and each information topic comprises one or more information items. Each information topic has an interesting score, which indicates a level of interest the user has in the information topic. In one embodiment, a social network is inferred among a group of users, each having a personal information profile. A connection exists between every two users that share at least one common-interest information topic, indicated by their respective personal information profiles. |
| 主权项 |
1. A method comprising: by one or more computing devices,
for each information item of a plurality of information items in a collection of information associated with a first user:
constructing a vector space model; andadjusting the vector space model by an importance score of the information item, wherein the importance score indicates the level of importance the information item is to the first user; calculating a degree of similarity for every two information items in the collection of information; based on the degree of similarity calculated for every two information items in the collection of information, clustering the plurality of information items in the collection of information associated with the first user to generate a plurality of information topics using the adjusted vector space model of each information item; for each information topic of the plurality of information topics generated from the collection of information associated with the first user, determining an interest score, which indicates a level of interest the first user has in the information topic relative to other information topics within the plurality of information topics; for each information topic of the plurality of information topics, determining a second interest score, which indicates a level of interest a second user has in the information topic relative to other information topics within the plurality of information topics; determining that the first user shares a common interest in a plurality of common-interest information topics with the second user, wherein the determining that the interest is a shared common interest is based on the difference between the first user's level of interest in an information topic and the second user's level of interest in an information topic, as determined by the interest score and the second interest score respectively, being less than a threshold difference; determining a common-interest strength between the first user and the second user, wherein the common-interest strength is calculated based on:
the number of common-interest information topics shared between the first user and the second user,a respective level of interest, as determined by the interest score and the second interest score, by each of the first user and the second users in the plurality of common-interest information topics, anda correlation and difference between the respective level of interest in the plurality of common-interest information topics by the first user and the second user, the common-interest strength measuring the strength of connection between the first user and the second user; and establishing a connection in a social network between the first and second users based on the common-interest strength. |
