System and method for context-based knowledge search, tagging, collaboration, management, and advertisement
Comprehensive methods and systems are described for creating, managing, searching, personalizing, and monetizing a knowledge system defined over a corpus of digital content. Systems and methods are described in which a user can initiate in-depth searches of subject matter and can browse, navigate, pinpoint, and select relevant contexts, concepts, and documents to gain knowledge. Systems and methods are described in which knowledge can be personalized through tagged, personalized context, and personalized context can be shared within social and professional networks, securely and confidentially and with the desired access control. Systems and methods are described in which products and services can be advertised in context and advertising can be selected through a bidding process. Systems and methods are described by which a user can navigate contexts and concepts to obtain relevant information, products and services.
1. A computer implemented method for providing relevant advertising, the method comprising:
receiving a query from a user;
determining a context of the query using a representation of a structured web community stored in a database, the structured web community comprising a plurality of contexts, each context comprising a set of content, the set of content in each context determined using a community finding method based on structural closeness and semantic closeness of the content, each context including an array of predefined concepts generated based on terms extracted from the set content in the context and statistical frequency of occurrence of terms within the set of content in the context, each predefined concept characterized by a pattern of terms derived from the set of content in the context, wherein a plurality of the plurality of contexts are associated together to form a hierarchy of overlapping contexts;
identifying at least one predefined concept associated with the determined context;
mapping advertising content to the at least one context of the plurality of contexts based on the at least one predefined concept; and
providing the advertising content to the user in response to the query;
wherein determining the set of content in each context using a community finding method based on structural closeness and semantic closeness of the content comprises:
selecting a node in a network as a source node; computing a set of local communities for the source node;
identifying a set of nodes in the set of local communities having a weight greater than a threshold;
generating a strong local community for the source node including only the set of nodes that have a weight greater than the threshold;
storing the strong local community as one of a plurality of communities of the network;
removing the set of nodes in the strong local community and edges connected to the set of nodes from the network to generate a reduced network;
selecting a node in the reduced network as a second source node;
generating a second strong local community for the second source node, the second strong local community comprising a second set of nodes;
storing the second strong local community as one of the plurality of communities of the network;
removing the second set of nodes in the second strong local community from the network to generate a second reduced network;
repeating the selecting, generating, storing and removing until a reduced network is generated that comprises only nodes with a degree less than a threshold value; and labeling the set of stored strong local communities as one of a disjoint community structure of the network or an overlapping community structure of the network.
2. The method of claim 1 , wherein the mapping comprises obtaining one or more advertisements from an external advertisement feed.
3. The method of claim 1 , wherein the mapping comprises obtaining one or more advertisements from a manual entry based on user selection of one or more contexts.
4. The method of claim 1 , wherein the mapping further comprises providing an advertisement based on bids received from one or more sources.
5. The method of claim 3 , wherein the mapping comprises obtaining one or more advertisements based on context of a query.
6. A computer implemented method for providing categorized data, the method comprising:
maintaining a representation of a structured web community stored in a database, the structured web community comprising a plurality of contexts, each context comprising a set of content, the set of content in each context determined using a community finding method based on structural closeness and semantic closeness of the content in the context, each context including an array of predefined concepts generated based on terms extracted from the set content in the context and statistical frequency of occurrence of terms within the set of content in the context, each predefined concept characterized by a pattern of terms derived from the set of content in the context, wherein a plurality of the plurality of contexts are associated together to form a hierarchy of overlapping contexts;
mapping advertising content to at least one of the plurality of contexts using the at least one predefined concept; and
storing the mapping of the advertising content to the at least one context in a database;
wherein determining the set of content in each context using a community finding method based on structural closeness and semantic closeness of the content comprises:
selecting a node in a network as a source node; computing a set of local communities for the source node;
identifying a set of nodes in the set of local communities having a weight greater than a threshold;
generating a strong local community for the source node including only the set of nodes that have a weight greater than the threshold;
storing the strong local community as one of a plurality of communities of the network;
removing the set of nodes in the strong local community and edges connected to the set of nodes from the network to generate a reduced network;
selecting a node in the reduced network as a second source node;
generating a second strong local community for the second source node, the second strong local community comprising a second set of nodes;
storing the second strong local community as one of the plurality of communities of the network;
removing the second set of nodes in the second strong local community from the network to generate a second reduced network;
repeating the selecting, generating, storing and removing until a reduced network is generated that comprises only nodes with a degree less than a threshold value; and labeling the set of stored strong local communities as one of a disjoint community structure of the network or an overlapping community structure of the network.
7. The method of claim 6 , wherein the advertising content is obtained from one or more external feeds.
8. The method of claim 6 , wherein the advertising content is obtained from a manual entry based on user selection of one or more contexts.
9. The method of claim 8 , wherein the advertising content is mapped based on context of a query.
10. A computer implemented method for operating context-sensitive databases, the method comprising:
receiving a user selection of one of a plurality of contexts;
determining one or more predefined concepts associated with the one of the plurality of contexts from a representation of a structured web community stored in a database, the structured web community comprising a plurality of contexts, each context comprising a set of content, the set of content included in each context determined using a community finding method based on structural closeness and semantic closeness of the content in the context, each context including an array of predefined concepts generated based on terms extracted from the set content in the context and statistical frequency of occurrence of terms within the set of content in the context, each predefined concept characterized by a pattern of terms derived from the set of content in the context, wherein a plurality of the plurality of contexts are associated together to form a hierarchy of overlapping contexts;
associating an advertisement with the one of the plurality of contexts based on the at least one predefined concept; and
presenting the advertisement in response to the user selection of the one of the plurality of contexts;
wherein determining the set of content in each context using a community finding method based on structural closeness and semantic closeness of the content comprises:
selecting a node in a network as a source node; computing a set of local communities for the source node;
identifying a set of nodes in the set of local communities having a weight greater than a threshold;
generating a strong local community for the source node including only the set of nodes that have a weight greater than the threshold;
storing the strong local community as one of a plurality of communities of the network;
removing the set of nodes in the strong local community and edges connected to the set of nodes from the network to generate a reduced network;
selecting a node in the reduced network as a second source node;
generating a second strong local community for the second source node, the second strong local community comprising a second set of nodes;
storing the second strong local community as one of the plurality of communities of the network;
removing the second set of nodes in the second strong local community from the network to generate a second reduced network;
repeating the selecting, generating, storing and removing until a reduced network is generated that comprises only nodes with a degree less than a threshold value; and labeling the set of stored strong local communities as one of a disjoint community structure of the network or an overlapping community structure of the network.
11. The method of claim 10 , wherein the advertisement is associated based on relevancy of the one of the plurality of contexts to one or more concepts.
12. The method of claim 10 , wherein the presenting comprises:
matching a document retrieved by the user with one of the certain contexts; and
providing the advertisement in a context-sensitive Ad Billboard.
13. The method of claim 12 , wherein matching comprises determining relevance of terms in the document to at least one of the predefined concepts.
14. The method of claim 10 , wherein the associating comprises subscribing to a service implementing the method.
15. The method of claim 1 , wherein the set of content in each context is determined using an information neighborhood analysis.
16. The method of claim 1 , wherein the set of content in each context is determined using a percolation community finding method combined with an agglomerative community fining method.
17. The method of claim 1 , wherein the set of content in each context is determined using a percolation community finding method combined with a second community fining method.
18. The method of claim 1 , wherein each context is further associated with a statistical concept, wherein the statistical concept is determined based on a statistical frequency of occurrence of terms within the context.
19. The method of claim 1 , wherein mapping the advertising content to a context comprises:
receiving user input, wherein the user input comprises a description, a website or an identification of at least one context;
determining relevant contexts based on the user input; and
associating the advertisement with the relevant contexts.
20. The method of claim 1 , wherein determining the set of content in each context using a community finding method based on structural closeness and semantic closeness of the content comprises:
assigning a set of nodes as source nodes;
initiating a plurality of percolation messages from each of source nodes at a plurality of percolation probabilities, wherein the plurality of percolation probabilities are selected from a set of values between 0 and 1;
determining a list of nodes that received the plurality of percolation messages for each of the plurality of source nodes;
aggregating the list of nodes for each of the plurality of source nodes to determine a local neighborhood for each of the plurality of source nodes, wherein each node is weighted based on a number of times each node received the plurality of percolation messages;
plotting the local neighborhood size versus percolation probabilities for each of the local neighborhoods;
locating phase transition points in the plot; and
determining nested local communities for the plurality of source nodes by combining the local neighborhoods at the phase transition points.