Intelligent self-enabled solution discovery

• US 8,291,319 B2
• Filed: 08/28/2009
• Issued: 10/16/2012
• Est. Priority Date: 08/28/2009
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Li, Huajing, Pingali, Gopal Sarma, Vukovic, Maja
• Primary Examiner(s): Nunez, Jordany
• Application Number:
  US12/549,644
• Publication Number:
  US 20110055699A1
• Time in Patent Office: 1,145 Days
• Field of Search: 715/709, 704/9, 706/45, 706/52, 707/999.003, 707/999.004, 707/999.005
• US Class Current: 715/709
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06Q 10/06316 Sequencing of tasks or work

Abstract

Solutions for solving a problem experienced by a user are retrieved. In response to receiving a query from the user describing the problem, relevant candidate solutions to the problem are sent to the user. In response to receiving a selection of one relevant candidate solution from the relevant candidate solutions, instructions steps within the one relevant candidate solution selected by the user are analyzed. An instruction step similarity is calculated between the instruction steps within the one relevant candidate solution selected and other instructions steps within other solutions stored in a storage device. Then, similar solutions are sent to the user containing similar instruction steps to the instruction steps contained within the one relevant candidate solution selected based on the calculated instruction step similarity.

Claims

1. A computer implemented method for retrieving solutions that solve a problem experienced by a user, the computer implemented method comprising:

generating, by a computer, a candidate solution document set for solving the problem, wherein a customized solution procedure for solving the problem is generated by the computer from a plurality of stored solution documents, and wherein a modified solution procedure with another set of instruction steps is generated by the computer for solving the problem based on the computer receiving an input rejecting one or more instruction steps included in the customized solution procedure;

generating, by the computer, a document object model tree for the generated candidate solution document set;

simplifying, by the computer, the generated document object model tree for the generated candidate solution document set by filtering out nodes in the generated document object model tree that do not have structural effects;

generating, by the computer, a template based on the simplified document object model tree;

calculating, by the computer, a structural similarity score for solution documents by comparing document object model trees of the solution documents with the generated template;

determining, by the computer, whether the structural similarity score for the solution documents is greater than a predetermined threshold;

responsive to the computer determining that the structural similarity score is greater than the predetermined threshold, storing, by the computer, the solution documents with structural similarity scores greater than the predetermined threshold;

responsive to the computer receiving a query describing the problem, sending, by the computer, relevant candidate solutions to the problem, wherein the relevant candidate solutions include unstructured hypertext markup language solution documents found on a world wide web, and wherein the unstructured hypertext markup language solution documents include solution data found in web logs, instant messaging chat sessions, and online message boards;

responsive to the computer receiving a selection of one relevant candidate solution from the relevant candidate solutions, analyzing, by the computer, instructions steps within the one relevant candidate solution selected;

calculating, by the computer, an instruction step similarity between the instruction steps within the one relevant candidate solution selected and other instructions steps within the stored solution documents; and

sending, by the computer, similar solutions containing similar instruction steps to the instruction steps contained within the one relevant candidate solution selected based on the calculated instruction step similarity.

2. The computer implemented method of claim 1, further comprising:

responsive to the computer receiving the query describing the problem, retrieving, by the computer, a query history associated with the user;

determining, by the computer, whether a problem topic taxonomy is available;

responsive to the computer determining that a problem topic taxonomy is available, determining, by the computer, a taxonomy-based context for the problem based on a description of the problem within the query and the retrieved query history associated with the user;

generating, by the computer, a context for the problem;

translating, by the computer, the query into an internal form that includes original terms in the query and the generated context for the problem; and

accessing, by the computer, full-text indices using the translated query to determine a textual relevance of candidate solutions to the problem.

3. The computer implemented method of claim 2, wherein the query history associated with the user includes a history of problem queries submitted by the user, user search and browse history, and user profile information.

4. The computer implemented method of claim 1, further comprising:

segmenting, by the computer, a solution into a series of individual instruction steps;

assigning, by the computer, a label to each of the individual instruction steps in the series that represents an individual instruction step'"'"'s function within the solution;

generating, by the computer, a solution path by assembling labels assigned to each of the individual instruction steps within the solution; and

storing, by the computer, the solution path.

5. The computer implemented method of claim 1, wherein the instruction steps include click-through instruction steps and core instruction steps, and wherein the click-through instruction steps and core instruction steps are labeled for identification.

6. The computer implemented method of claim 5, wherein the core instruction steps are highlighted, and wherein the core instruction steps are action steps that the user follows to solve the problem, and wherein the click-through instruction steps guide the user to the core instruction steps.

7. The computer implemented method of claim 5, wherein in response to determining a current solution the user is browsing, a textual relevance of the core instruction steps of this currently browsed solution is determined in order to recommend other relevant solutions.

8. The computer implemented method of claim 1, wherein metadata is used to determine similarity between solutions.

9. The computer implemented method of claim 1, wherein the customized solution procedure for solving the problem is generated by combining one or more instruction steps selected from a first solution document with one or more other instruction steps selected from a second solution document.

10. The computer implemented method of claim 1, wherein a new solution procedure is generated from the plurality of stored solution documents based on an interactive process with the user during a recommendation process.

11. The computer implemented method of claim 1, wherein the solution data found on the world wide web is analyzed to determine which problems these available solution data solve in order to store and reuse the solution data in a context of solving problems submitted by other users.

12. A computer system for retrieving solutions that solve a problem experienced by a user, the computer system comprising:

a bus system;

a storage device connected to the bus system, wherein the storage device stores a set of instructions; and

a processing unit connected to the bus system, wherein the processing unit executes the set of instructions to generating a candidate solution document set for solving the problem, wherein a customized solution procedure for solving the problem is generated from a plurality of stored solution documents, and wherein a modified solution procedure with another set of instruction steps is generated for solving the problem based on the computer system receiving an input rejecting one or more instruction steps included in the customized solution procedure;

generate a document object model tree for the generated candidate solution document set;

simplify the generated document object model tree for the generated candidate solution document set by filtering out nodes in the generated document object model tree that do not have structural effects;

generate a template based on the simplified document object model tree;

calculate a structural similarity score for solution documents by comparing document object model trees of the solution documents with the generated template;

determine whether the structural similarity score for the solution documents is greater than a predetermined threshold;

store the solution documents with structural similarity scores greater than the predetermined threshold in response to determining that the structural similarity score is greater than the predetermined threshold;

send relevant candidate solutions to the problem in response to receiving a query describing the problem, wherein the relevant candidate solutions include unstructured hypertext markup language solution documents found on a world wide web, and wherein the unstructured hypertext markup language solution documents include solution data found in web logs, instant messaging chat sessions, and online message boards;

analyze instructions steps within one relevant candidate solution selected in response to receiving a selection of the one relevant candidate solution from the relevant candidate solutions;

calculate an instruction step similarity between the instruction steps within the one relevant candidate solution selected and other instructions steps within the stored solution documents; and

send similar solutions containing similar instruction steps to the instruction steps contained within the one relevant candidate solution selected based on the calculated instruction step similarity.

13. A computer program product stored on a computer readable storage device having computer usable program code embodied thereon that is executable by a computer for retrieving solutions that solve a problem experienced by a user, the computer program product comprising:

computer usable program code for generating a candidate solution document set for solving the problem, wherein a customized solution procedure for solving the problem is generated from a plurality of stored solution documents, and wherein a modified solution procedure with another set of instruction steps is generated for solving the problem based on the computer receiving an input rejecting one or more instruction steps included in the customized solution procedure;

computer usable program code for generating a document object model tree for the generated candidate solution document set;

computer usable program code for simplifying the generated document object model tree for the generated candidate solution document set by filtering out nodes in the generated document object model tree that do not have structural effects;

computer usable program code for generating a template based on the simplified document object model tree;

computer usable program code for calculating a structural similarity score for solution documents by comparing document object model trees of the solution documents with the generated template;

computer usable program code for determining whether the structural similarity score for the solution documents is greater than a predetermined threshold;

computer usable program code for storing the solution documents with structural similarity scores greater than the predetermined threshold in response to the computer determining that the structural similarity score is greater than the predetermined threshold;

computer usable program code for sending relevant candidate solutions to the problem in response to receiving a query describing the problem, wherein the relevant candidate solutions include unstructured hypertext markup language solution documents found on a world wide web, and wherein the unstructured hypertext markup language solution documents include solution data found in web logs, instant messaging chat sessions, and online message boards;

computer usable program code for analyzing instructions steps within one relevant candidate solution selected in response to receiving a selection of the one relevant candidate solution from the relevant candidate solutions;

computer usable program code for calculating an instruction step similarity between the instruction steps within the one relevant candidate solution selected and other instructions steps within the stored solution documents; and

computer usable program code for sending similar solutions containing similar instruction steps to the instruction steps contained within the one relevant candidate solution selected based on the calculated instruction step similarity.

14. The computer program product of claim 13, further comprising:

computer usable program code for retrieving a query history associated with the user in response to receiving the query describing the problem;

computer usable program code for determining whether a problem topic taxonomy is available;

computer usable program code for determining a taxonomy-based context for the problem based on a description of the problem within the query and the retrieved query history associated with the user in response to determining that a problem topic taxonomy is available;

computer usable program code for generating a context for the problem;

computer usable program code for translating the query into an internal form that includes original terms in the query and the generated context for the problem; and

computer usable program code for accessing full-text indices using the translated query to determine a textual relevance of candidate solutions to the problem.

15. The computer program product of claim 13, further comprising:

computer usable program code for segmenting a solution into a series of individual instruction steps;

computer usable program code for assigning a label to each of the individual instruction steps in the series that represents an individual instruction step'"'"'s function within the solution;

computer usable program code for generating a solution path by assembling labels assigned to each of the individual instruction steps within the solution; and

computer usable program code for storing the solution path.

16. The computer program product of claim 13, wherein the instruction steps include click-through instruction steps and core instruction steps, and wherein the click-through instruction steps and core instruction steps are labeled for identification.

17. The computer program product of claim 16, wherein the core instruction steps are highlighted, and wherein the core instruction steps are action steps that the user follows to solve the problem, and wherein the click-through instruction steps guide the user to the core instruction steps.

18. The computer program product of claim 16, wherein in response to determining a current solution the user is browsing, a textual relevance of the core instruction steps of this currently browsed solution is determined in order to recommend other relevant solutions.

19. The computer program product of claim 13, wherein metadata is used to determine similarity between solutions.

20. The computer program product of claim 13, wherein the customized solution procedure for solving the problem is generated by combining one or more instruction steps selected from a first solution document with one or more other instruction steps selected from a second solution document.