Enhanced concern indicator failure prediction system

• US 20030158702A1
• Filed: 02/15/2002
• Published: 08/21/2003
• Est. Priority Date: 02/15/2002
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: Busche, Frederick D., Calkins, William D., Kish, Jeffrey D., Swain, Scott L.
• Granted Patent: US 6,714,893 B2
• Time in Patent Office: Days
• US Class Current: 702/181
• CPC Class Codes: G06F 16/284 Relational databases

Abstract

An enhanced concern indicator failure prediction system is provided to predict possible product failures with automatic notification of people as well as systems. The system integrates data mining, spatial analysis, linear programming, narrowcasting, data warehousing, visualization, and text mining. As a result, failure conditions, attributes, complaints, locations, consequences, and sequence of events are analyzed using data mining technologies. This data is fed into an optimization module that assesses the efficiency of the failure process such that failures can be assessed as to their priority. These priorities are then used to feed a triggering engine that triggers notification of systems and individuals using narrowcasting technology. This system is one that allows early warning of potential problems to occur and integrates data from call centers, legacy systems, retailers, manufacturers, vendor supplied parts, and transportation of parts.

Claims

1. A method for predicting product failures, comprising:

defining predictive indicators;

performing data mining on product data using the predictive indicators;

determining whether triggering data is discovered by the data mining; and

distributing triggering data to subscribers if triggering data is discovered.

2. The method of claim 1, further comprising performing data visualization on the product data.

3. The method of claim 2, wherein the step of defining predictive indicators comprises defining predictive indicators based on results of the data visualization.

4. The method of claim 2, wherein the step of performing data mining comprises performing data mining based on results of the data visualization.

5. The method of claim 1, wherein the data mining comprises at least one of segmentation, classification, statistical verification prediction, associations, outlier analysis, profile modeling, and model scoring.

6. The method of claim 1, further comprising using a weighting engine to optimize multiple inputs from the predictive indicators and multiple outputs to product data storage.

7. The method of claim 6, wherein the weighting engine optimizes the multiple inputs and multiple outputs using scoring from the data mining.

8. The method of claim 1, further comprising performing text mining on the product data to collect and manipulate information that may be used to predict product failures.

9. The method of claim 1, wherein the step of determining whether triggering data is discovered by the data mining comprises:

discovering potential product failures;

assessing priorities of the potential product failures; and

sending the assessed priorities to a triggering engine.

10. The method of claim 9, wherein the triggering engine triggers notification of at least one system.

11. The method of claim 10, wherein the at least one system comprises a pager.

12. The method of claim 10, wherein the at least one system comprises an e-mail account.

13. The method of claim 9, wherein the triggering engine triggers notification of an individual at a user dashboard.

14. The method of claim 1, wherein the step of distributing triggering data to subscribers comprises distributing the triggering data using narrowcasting.

15. An apparatus for predicting product failures, comprising:

definition means for defining predictive indicators;

data mining means for performing data mining on product data using the predictive indicators;

determination means for determining whether triggering data is discovered by the data mining; and

distribution means for distributing triggering data to subscribers if triggering data is discovered.

16. The apparatus of claim 15, further comprising data visualization means for performing data visualization on the product data.

17. The apparatus of claim 16, wherein the definition means comprises means for defining predictive indicators based on results of the data visualization means.

18. The apparatus of claim 16, wherein the data mining means comprises means for performing data mining based on results of the data visualization means.

19. The apparatus of claim 15, wherein the data mining means comprises tools for performing at least one of segmentation, classification, statistical verification prediction, associations, outlier analysis, profile modeling, and model scoring.

20. The apparatus of claim 15, further comprising a weighting means for optimizing multiple inputs from the predictive indicators and multiple outputs to product data storage.

21. The apparatus of claim 20, wherein the weighting means optimizes the multiple inputs and multiple outputs using scoring from the data mining means.

22. The apparatus of claim 15, further comprising text mining means for performing text mining on the product data to collect and manipulate information that may be used to predict product failures.

23. The apparatus of claim 15, wherein the determination means comprises:

means for discovering potential product failures;

means for assessing priorities of the potential product failures; and

means for sending the assessed priorities to a triggering engine.

24. The apparatus of claim 23, wherein the triggering engine triggers notification of at least one system.

25. The apparatus of claim 24, wherein the at least one system comprises a pager.

26. The apparatus of claim 24, wherein the at least one system comprises an e-mail account.

27. The apparatus of claim 23, wherein the triggering engine triggers notification of an individual at a user dashboard.

28. A computer program product, in a computer readable medium, for predicting product failures, comprising:

instructions for defining predictive indicators;

instructions for performing data mining on product data using the predictive indicators;

instructions for determining whether triggering data is discovered by the data mining; and

instructions for distributing triggering data to subscribers if triggering data is discovered.