Using a dynamically-generated content-level newsworthiness rating to provide content recommendations

• US 8,386,457 B2
• Filed: 06/22/2011
• Issued: 02/26/2013
• Est. Priority Date: 06/22/2011
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Trovati, Marcello, Zimmet, Carol Sue, McCloskey, Daniel John
• Primary Examiner(s): To, Baoquoc N
• Application Number:
  US13/166,286
• Publication Number:
  US 20120330936A1
• Time in Patent Office: 615 Days
• Field of Search: 707/748, 707/749, 707/750
• US Class Current: 707/706
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A method for providing content-level data artifact recommendations can begin with the creation of a semantic library from the textual content of data artifacts by a newsworthy content recommendation engine. A base newsworthiness rating can be calculated using global newsworthiness parameters and behavioral functions that model newsworthy influences for each relationship contained in the semantic library. A user-specific search network can be generated that represents user-entered criteria and/or user task-related criteria. Within the semantic library, potential newsworthy semantic networks can be identified. Newsworthy content from each identified potential newsworthy semantic network can be dynamically determined based upon the base newsworthiness rating and a predefined newsworthiness threshold. The newsworthy content from the identified potential newsworthy semantic network can be related to the user-specific search network at the common node, creating a newsworthy content recommendation graph. The newsworthy content recommendation graph can be presented within a user interface.

Claims

1. A system comprising:

one or more processors;

a plurality of data sources, stored in at least one memory, housing a plurality of data artifacts having textual content;

a user context that represents at least one of at least one user-entered criterion and at least one user task-related criterion, wherein the at least one user task-related criterion is captured from a user session;

a newsworthy content recommendation engine, comprising program instructions stored in a memory, said program instructions being executable by at least one of the one or more processors, configured to generate a newsworthy content recommendation graph for the user context, said newsworthy content recommendation engine comprising;

a semantic library representing the textual content of the plurality of data artifacts as a plurality of semantic networks, wherein each of the plurality of semantic networks comprises a plurality of semantic units that express a relationship between a subject node and an object node, wherein separate one of the plurality of semantic networks are inter-related at common nodes;

at least one semantic ontology defining at least one domain for describing the plurality of semantic networks;

a semantic network generator configured to create the semantic library from the textual content of the plurality of data artifacts using the at least one semantic ontology;

a semantic recommendation handler configured to determine one or more of the plurality of semantic networks within the semantic library related to the user context;

at least one behavioral function configured to model an influence of the subject node upon the object node;

a set of global newsworthiness parameters defining default values for variables utilized in the at least one behavioral function;

a predefined newsworthiness threshold defining a minimum value for a semantic unit to be determined as newsworthy with respect to the user context; and

a newsworthiness calculator configured to calculate a newsworthiness rating for the semantic units contained in the semantic networks identified by the semantic recommendation handler with respect to the user context using the at least one behavioral function, the set of global newsworthiness parameters, and the predefined newsworthiness threshold, wherein semantic units determined to be newsworthy are included in the newsworthy content recommendation graph.

2. The system of claim 1, further comprising:

a plurality of user-specific newsworthiness parameters representing user-configured values for at least one of the set of global newsworthiness parameters, wherein said user-configured values are used by the newsworthy content recommendation engine in place of values for the corresponding set of global newsworthiness parameters.

3. The system of claim 1, wherein the newsworthy content recommendation engine is a component of a content analysis system.

4. The system of claim 1, wherein the newsworthy content recommendation engine further comprises:

a user interface configured to serve as an interaction mechanism for accessing functionality provided by the newsworthy content recommendation engine, wherein said functionality comprises at least an ability to semantically search the semantic library, an ability to navigate through nodes of the newsworthy content recommendation graph, and an ability to explore an impact of a change to at least one of the set of global newsworthiness parameters upon generation of the newsworthy content recommendation graph.

5. A computer program product comprising a non-transitory computer readable storage medium having computer usable program code embodied therewith, the non-transitory computer usable program code comprising:

computer usable program code configured to create a semantic library from textual content of data artifacts by a newsworthy content recommendation engine;

computer usable program code configured to calculate a base newsworthiness rating using at least one global newsworthiness parameter and at least one behavioral function that model at least one newsworthy influence for each relationship contained in the semantic library;

computer usable program code configured to generate a user-specific search network that represents user-entered criteria and/or user task-related criteria, wherein within the semantic library, at least one potential newsworthy semantic network is identified;

computer usable program code configured to dynamically determine newsworthy content from each identified potential newsworthy semantic network based upon the base newsworthiness rating and a predefined newsworthiness threshold, wherein the newsworthy content from the identified potential newsworthy semantic network is related to the user-specific search network at a common node; and

computer usable program code configured to create a newsworthy content recommendation graph using the determined newsworthy content.

6. The computer program product of claim 5, further comprising:

computer usable program code configured to detect a change to the textual content of one of the data artifacts represented in the semantic library;

computer usable program code configured to modify the semantic network corresponding to the one data artifact to reflect the detected change;

computer usable program code configured to update relationships of the modified semantic network to other semantic networks in the semantic library, wherein said updating results in at least one of a creation of new relationship to a node of another semantic network that is now common and a removal of an existing relationship to a node of another semantic network that is no longer common; and

computer usable program code configured to adjust the base newsworthiness rating for the updated relationships.

7. The computer program product of claim 5, wherein determining the existence of newsworthy content further comprises:

computer usable program code configured to identify an existence of at least one user-configured newsworthiness parameter for the at least one global newsworthiness parameter; and

computer usable program code configured to, when at least one user-configured newsworthiness parameter exists, dynamically adjust the calculation of the base newsworthiness rating for the newsworthy content of the identified semantic network with respect to the at least one user-configured newsworthiness parameter, wherein a value for a user-configured newsworthiness parameter takes precedence over a corresponding global newsworthiness parameter, wherein said adjusted newsworthiness rating is subsequently utilized in place of the base newsworthiness rating.

8. The computer program product of claim 5, wherein determining the existence of newsworthy content further comprises:

computer usable program code configured to compare the base newsworthiness rating of the newsworthy content to the predefined newsworthiness threshold; and

computer usable program code configured to, when the base newsworthiness rating of the newsworthy content satisfies the predefined newsworthiness threshold, continue with the relating of the newsworthy content to the user-specific search network.

9. The computer program product of claim 5, wherein identifying of the at least one semantic network further comprises:

computer usable program code configured to, for each identified potential newsworthy semantic network, compare the base newsworthiness rating of each semantic unit to the predefined newsworthiness threshold; and

computer usable program code configured to, when a difference between the base newsworthiness rating of the semantic unit and the predefined newsworthiness threshold meets a predetermined value, remove the semantic unit from the identified potential newsworthy semantic network, wherein semantic units subordinate to the removed semantic unit are also removed.

10. The computer program product of claim 5, further comprising:

computer usable program code configured to, in response to a user-selected command within the user interface, present at least one current value for the at least one global newsworthiness parameter used when determining the base newsworthiness rating for a selected relationship in the newsworthy content recommendation graph and the predefined newsworthiness threshold;

computer usable program code configured to receive via the user interface a modification to a current value for at least one global newsworthiness parameter or a value of the predefined newsworthiness threshold;

computer usable program code configured to create a second newsworthy content recommendation graph using the user-specific search network, the received modification, and unmodified values for the plurality of global newsworthiness parameters; and

computer usable program code configured to present the second newsworthy content recommendation graph within the user interface such that both the newsworthy content recommendation graph and the second newsworthy content recommendation graph are capable of being visually compared with each other.