Transforming a knowledge base into a machine readable format for an automated system

1. A method for transforming a knowledge base into a machine readable format, comprising:

clustering, by a processor of a knowledge base processing server, two or more documents of a plurality of documents in the knowledge base that are similar based upon a domain specific shingling function;

identifying, by the processor, a pattern from each document of the plurality of documents that are clustered, wherein the pattern comprises a sequence of steps;

aggregating, by the processor, the pattern of the each document of the plurality of documents that are clustered into a trie data-structure that is machine readable;

receiving, by the processor, a request; and

applying, by the processor, the trie data-structure to provide a solution to the request.

2. The method of claim 1 , further comprising:

pre-processing, by the processor, the plurality of documents to extract metadata.

3. The method of claim 2 , wherein the domain specific shingling function comprises identifying a plurality of different domains in each one of the plurality of documents based on the metadata.

4. The method of claim 3 , wherein the clustering comprises:

calculating, by the processor, a similarity score for each one of the plurality of different domains of a document to a corresponding domain of each remaining document of the plurality of documents;

calculating, by the processor, an overall similarity score based on the similarity score for the each one of the plurality of different domains for the document to the each remaining document of the plurality of documents; and

clustering, by the processor, the document and one or more of the each remaining document of the plurality of documents when the overall similarity score is above a threshold.

5. The method of claim 1 , wherein the identifying the pattern, comprises:

building, by the processor, a suffix tree for each one of the plurality of documents that are clustered;

extracting, by the processor, each pattern of a plurality of patterns from the suffix tree for the each one of the plurality of documents that are clustered; and

identifying, by the processor, the pattern from the plurality of patterns that occurs above a predefined number of documents of the plurality of documents that are clustered.

6. The method of claim 5 , wherein the extracting each pattern of the plurality of patterns comprises using a suffix link when traversing the suffix tree.

7. The method of claim 1 , wherein the aggregating comprises:

combining, by the processor, a plurality of patterns into a single pattern represented by the trie data-structure.

8. The method of claim 1 , wherein the clustering, the identifying and the aggregating are repeated periodically to account for new documents added to the knowledge base.

9. The method of claim 1 , wherein the applying the trie data-structure to provide the solution to the request comprises navigating the trie data-structure node by node until the solution is provided.

10. An apparatus, comprising:

a processor; and

a computer readable medium in communication with the processor, the computer readable medium comprising:

a clustering module to cluster two or more documents of a plurality of documents in a knowledge base that are similar based upon a domain specific shingling function;

a pattern identification module to identify a pattern from each document of the plurality of documents that are clustered, wherein the pattern comprises a sequence of steps;

an aggregation module to aggregate the pattern of the each document of the plurality of documents that are clustered into a trie data-structure that is machine readable; and

a customer interaction module to receive a request and apply the trie data-structure to provide a solution to the request.

11. The apparatus of claim 10 , wherein the processor is in communication with the knowledge base to obtain the plurality of documents.

12. The apparatus of claim 10 , wherein the domain specific shingling function comprises identifying a plurality of different domains in each one of the plurality of documents based on metadata extracted from the each one of the plurality of documents.

13. The apparatus of claim 12 , wherein the clustering module is further configured to:

calculate a similarity score for each one of the plurality of different domains of a document to a corresponding domain of each remaining document of the plurality of documents;

calculate an overall similarity score based on the similarity score for the each one of the plurality of different domains for the document to the each remaining document of the plurality of documents; and

cluster the document and one or more of the each remaining document of the plurality of documents when the overall similarity score is above a threshold.

14. The apparatus of claim 10 , wherein the pattern identification module is further configured to:

build a suffix tree for each one of the plurality of documents that are clustered;

extract each pattern of a plurality of patterns from the suffix tree for the each one of the plurality of documents that are clustered; and

identify the pattern from the plurality of patterns that occurs above a predefined number of documents of the plurality of documents that are clustered.

15. The apparatus of claim 14 , wherein the pattern identification module extracts each pattern of the plurality of patterns using a suffix link when traversing the suffix tree.

16. The apparatus of claim 10 , wherein the aggregation module is further configured to:

combine a plurality of patterns into a single pattern represented by the trie data-structure.

17. The apparatus of claim 10 , the cluster module, the pattern identification module and the aggregation module may be periodically activated to account for new documents added to the knowledge base.

18. The apparatus of claim 10 , wherein the customer interaction module applies the trie data-structure to provide the solution to the request by navigating the trie data-structure node by node until the solution is provided.

19. A method for transforming a knowledge base into a machine readable format, comprising:

extracting, by a processor of a knowledge base processing server, HTML tags for each document of a plurality of documents in the knowledge base, wherein the HTML tags define a plurality of domains for each one of the plurality of documents;

calculating, by the processor, an overall similarity score for each pair of documents of the plurality of documents based on a similarity score of each one of the plurality of domains of the each pair of documents;

clustering, by the processor, two or more documents of the plurality of documents into a cluster of documents that have the overall similarity score above a threshold value;

identifying, by the processor, a plurality of patterns comprising a pattern from each document of the cluster of documents using suffix links of a suffix tree associated with the each document of the cluster of documents, wherein the pattern comprises a sequence of steps;

aggregating, by the processor, the plurality of patterns into a trie data-structure that is machine readable;

receiving, by the processor, a customer care request; and

applying, by the processor, the trie data-structure to provide a solution to the customer care request.

20. The method of claim 19 , wherein the extracting, the calculating, the clustering, the identifying and the aggregating are repeated periodically to account for new documents added to the knowledge base.