Subject matter expert knowledge mapping using dynamic clustering
The example embodiments are directed to a system and method that applies knowledge developed by a subject matter expert with respect to a physical asset. In one example, the method includes receiving knowledge and issue resolution information developed of subject matter experts in association with historical issues for an asset, generating a plurality of data clusters for the asset based on the knowledge, wherein each historical issue of the asset is mapped to a cluster and includes a plurality of resolutions for the issue, receiving a new issue and new issue information, and processing the new issue by extracting keywords from the new issue information and assigning the new issue to a data cluster from among the plurality of data clusters based on the extracted keywords, and outputting, to a display, a cause of the new issue and potential solutions for the new issue.
1. A computing system, comprising:
a storage configured to store historical work orders associated with an asset; and
a processor configured to
extract keywords from the historical work orders,
cluster the historical work orders into a plurality of different clusters and generate a knowledge mapping algorithm based on execution of a natural language processing algorithm on the extracted keywords and a plurality of causes of failure mapped to the plurality of different clusters, respectively, wherein the knowledge mapping algorithm comprises rules that identify how to map a new work order to a cluster from among the plurality of different clusters, and
store the knowledge mapping algorithm in the storage.
2. The computing system of claim 1 , wherein the processor is configured to extract keywords from opinions and from data used to derive the opinions within the historical work orders.
3. The computing system of claim 1 , wherein the processor is configured to execute the natural language processing algorithm on the historical work orders to extract sequences of words and score the sequences of words.
4. The computing system of claim 1 , wherein the plurality of clusters are mapped to a plurality of different causes of failure of the asset, respectively.
5. The computing system of claim 1 , wherein the method further comprises modifying the plurality of clusters based on user feedback provided via a user interface, and generating a modified knowledge mapping algorithm based on the user feedback.
6. The computing system of claim 1 , wherein the natural language processing algorithm comprises a rapid automatic keyword extraction (RAKE) algorithm.
7. The computing system of claim 1 , wherein the processor is further configured to receive a new work order, extract keywords from the new work order, and execute the generated knowledge mapping algorithm on the extracted keywords to determine a cause of failure associated with the new work order.
8. The computing system of claim 7 , wherein the processor is configured to assign the new work order to a cluster from among the plurality of clusters via execution of the generated knowledge mapping algorithm.
9. A method comprising:
storing historical work orders associated with an asset;
extracting keywords from the historical work orders;
clustering the historical work orders into a plurality of different clusters and generating a knowledge mapping algorithm based on execution of a natural language processing algorithm on the extracted keywords and a plurality of causes of failure mapped to the plurality of different clusters, respectively, wherein the knowledge mapping algorithm comprises rules that identify how to map a new work order to a cluster from among the plurality of different clusters; and
storing the knowledge mapping algorithm in the storage.
10. The method of claim 9 , wherein the extracting comprises extracting keywords from opinions and from data used to derive the opinions within the historical work orders.
11. The method of claim 9 , wherein the extracting comprises executing the natural language processing algorithm on the historical work orders to extract sequences of words and score the sequences of words.
12. The method of claim 9 , wherein the plurality of clusters are mapped to a plurality of different causes of failure of the asset, respectively.
13. The method of claim 9 , wherein the method further comprises modifying the plurality of clusters based on user feedback provided via a user interface, and generating a modified knowledge mapping algorithm based on the user feedback.
14. The method of claim 9 , wherein the natural language processing algorithm comprises a rapid automatic keyword extraction (RAKE) algorithm.
15. The method of claim 9 , wherein the method further comprises receiving a new work order, extracting keywords from the new work order, and executing the generated knowledge mapping algorithm on the extracted keywords to determine a cause of failure associated with the new work order.
16. The method of claim 15 , wherein the clustering comprises assigning the new work order to a cluster from among the plurality of clusters via execution of the generated knowledge mapping algorithm.
17. A non-transitory computer-readable storage medium comprising instructions which when executed by a processor cause a computer to perform a method comprising:
storing historical work orders associated with an asset;
extracting keywords from the historical work orders;
clustering the historical work orders into a plurality of different clusters, and generating a knowledge mapping algorithm based on execution of a natural language processing algorithm on the extracted keywords and a plurality of issues mapped to the plurality of different clusters, respectively, wherein the knowledge mapping algorithm comprises rules that identify how to map a new work order to a cluster from among the plurality of different clusters; and
storing the knowledge mapping algorithm in the storage.
18. The non-transitory computer-readable medium of claim 17 , wherein the extracting comprises extracting keywords from opinions and from data used to derive the opinions within the historical work orders.
19. The non-transitory computer-readable medium of claim 17 , wherein the extracting comprises executing the natural language processing algorithm on the historical work orders to extract sequences of words and score the sequences of words.
20. The non-transitory computer-readable medium of claim 17 , wherein the plurality of clusters are mapped to a plurality of different causes of failure of the asset, respectively.