Computer system for applying predictive model to determine and indeterminate data

• US 9,378,527 B2
• Filed: 04/08/2008
• Issued: 06/28/2016
• Est. Priority Date: 04/08/2008
• Status: Active Grant

• Current Assignee: Hartford Fire Insurance Company (The Hartford Financial Services Group, Inc.)
• Original Assignee: Hartford Fire Insurance Company (The Hartford Financial Services Group, Inc.)
• Inventors: Adams, Jonathan Kaleb, McGoldrick, John Albert, Ruotolo, James Dominic
• Primary Examiner(s): Wong, Eric T
• Application Number:
  US12/099,388
• Publication Number:
  US 20090254379A1
• Time in Patent Office: 3,003 Days
• Field of Search: 705/4
• US Class Current: 1/1
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 99/00 Subject matter not provided for in other groups of this subclass
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 20/4016 involving fraud or risk level assessment in transaction processing
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

A computer system includes a data storage module. The computer system also includes an indeterminate data capture module. The data storage module receives, stores, and provides access to both determinate data and indeterminate data that was captured by the indeterminate data capture module. The computer system also includes a computer processor and a program memory. The computer processor executes programmed instructions and stores and retrieves the data stored in the data storage module. The data stored in the data storage module is related to historical claim transactions and current claim transactions. A predictive model component is coupled to the data storage module, and determines whether to refer a current claim transaction to an investigation unit. A model training component is coupled to the computer processor and functions to train the predictive model.

Claims

1. A computer system comprising:

a data storage module configured to receive, store, and provide access to indeterminate data and determinate data corresponding to historical claim transactions and to current claim transactions, and data indicative of categories of claims that are highly unlikely to be questionable;

a computer processor configured to;

receive determinate data corresponding to the historical claim transactions and the current claim transactions;

receive raw indeterminate data corresponding to the historical claim transactions and the current claim transactions;

extract, from the raw indeterminate data by an indeterminate data capture module, indeterminate data comprising at least one of claim handler notes, claim handler audio files, witness notes, witness audio files, claimant notes, and claimant audio files, the indeterminate data capture module including a natural language processing application to mine opinions from narrative text notes and a speech recognition application to analyze audio files to detect a speaker'"'"'s state of mind;

train, by a model training component, a predictive model component based on the determinate data and the extracted indeterminate data related to the historical claim transactions, thereby providing a trained predictive model component;

adapt the predictive model component to changing patterns of questionable claim transactions by periodically updating the predictive model component based on new determinate data and new extracted indeterminate data;

compare, by a screening module, a current claim transaction to the data indicative of categories of claims that are highly unlikely to be questionable and categorize, based on the comparing, each of said current claim transactions as one of a current claim suitable for predictive analysis when the current claim transaction does not correspond to the categories of claims that are highly unlikely to be questionable and a current claim not suitable for predictive analysis when the current claim transaction corresponds to at least one of the categories of claims that are highly unlikely to be questionable;

determine, by the predictive model component coupled to the data storage module and based on the indeterminate data and the determinate data, whether to identify a one of said current claim transactions categorized as suitable for predictive analysis for referral to an investigation unit;

an output device, coupled to the computer processor, configured to output an indication of whether said one of said current claim transactions should be referred to the investigation unit, said output generated in response to application of data for said one of said current claim transactions to the trained predictive model component and said screening module; and

a routing module configured to direct workflow corresponding to said one of said current claim transactions based on the output from the output device, to one of an investigative analyst computer and a field investigator computer.

2. The computer system of claim 1, wherein the indeterminate data capture module includes an optical character reader.

3. The computer system of claim 1, wherein the natural language processing module is configured to determine whether the indeterminate data contains one or more statements, and wherein the predictive model component is further configured to determine whether to identify a one of said current claim transactions categorized as suitable for predictive analysis for referral to an investigation unit based on the determined one or more statements contained in the indeterminate data.

4. The computer system of claim 3, wherein the indeterminate data includes claim handlers'"'"' narrative text notes.

5. The computer system of claim 4, wherein the natural language processing module extracts claim handlers'"'"' opinions from the claim handlers'"'"' narrative text notes.

6. The computer system of claim 1, wherein:

the data for said one of said current claim transactions specifies a claimant; and

the indeterminate data includes a recording of the claimant'"'"'s voice.

7. The computer system of claim 6, wherein the indeterminate data capture module includes a component for analyzing the claimant'"'"'s voice.

8. The computer system of claim 1, wherein the indeterminate data includes at least one of:

(a) text captured by optical character reading;

(b) information extracted from text; and

(c) text converted from a voice signal.

9. The computer system of claim 1, wherein the routing module routes said one of the current claim transactions to the investigation unit based on the output from the output device.

10. The computer system of claim 9, wherein the routing module routes the current claim transaction to an investigative analyst in the investigation unit.

11. The computer system of claim 1, wherein the claim transactions are worker'"'"'s compensation claims.

12. The computer system of claim 1, wherein the claim transactions are bodily injury claims.

13. The computer system of claim 1, where the claim transactions are property damage claims.

14. The computer system of claim 1, further comprising:

a processor programmed to;

place said current claim transactions categorized as suitable for predictive analysis in a queue for application of the predictive model component in response to categorizing said current claim transactions as a current claim suitable for analysis by the predictive model component.

15. The computer system of claim 14, wherein a workers'"'"' compensation claim that arises from loss of life comprises a category of claims that is highly unlikely to be questionable.

16. The computer system of claim 1, wherein said computer processor is further configured to:

process the indeterminate data for said one of said current claim transactions to detect at least one of;

(a) an indication that said one of said current claim transactions is disputed;

(b) an indication that a claimant engaged in physical activity; and

(c) an indication that said claimant made inconsistent statements.

17. The computer system of claim 1, wherein the output device being configured to output the indication of whether said one of said current claim transactions should be referred to the investigation unit comprises the output device being configured to generate a score corresponding to the advisability of referring the one of the current claim transactions to the investigation unit;

and

wherein the routing module for directing workflow based on the output from the output device is configured to;

determine whether the score exceeds a threshold value indicative of a claim being advisable for referral to the investigation unit; and

responsive to a determination that the score does exceed the threshold value, transmit data indicative of a case file for the one of the current claim transactions to an investigation unit computer.

18. The computer system of claim 17, wherein the routing module being configured to transmit the data indicative of a case file for the one of the current claim transactions to an investigation unit computer comprises the routing module being configured to:

transmit the data indicative of the case file for the one of the current claim transactions, and data indicative of the score corresponding to the advisability of referring the one of the current claim transactions to an investigation unit, to an investigation unit computer queue which is configured to place data indicative of case files in order according to the scores corresponding to each of the case files.

19. The computer system of claim 17, wherein the routing module is further configured to transmit, responsive to a determination that the score for the one of the current claim transactions does exceed the threshold value, a notification to a claim handler device to suspend processing of the claim transaction.

20. A computer system comprising:

a data storage unit configured to receive, store, and provide access to historical claim data and current claim data, the current claim data representing a plurality of current claim transactions, and said data storage unit also data indicative of categories of claims that are highly unlikely to be questionable;

an indeterminate data capture module configured to extract, from the raw indeterminate data, indeterminate data comprising at least one of claim handler notes, claim handler audio files, witness notes, witness audio files, claimant notes, and claimant audio files, the indeterminate data capture module including a natural language processing application to mine opinions from narrative text notes and a speech recognition application to analyze audio files to detect a speaker'"'"'s state of mind;

a computer processor configured to;

compare, by a screening module, a current claim transaction to the data indicative of categories of claims that are highly unlikely to be questionable and to categorize, based on the comparison, each of the current claim transactions as one of a current claim transaction suitable for predictive analysis when the current claim transaction does not correspond to the categories of claims that are highly unlikely to be questionable and a current claim transaction not suitable for predictive analysis when the current claim transaction corresponds to at least one of the categories of claims that are highly unlikely to be questionable;

train, by a model training component, a predictive model based on the indeterminate data and the determinate data for each current claim transaction categorized as suitable for predictive analysis in the data storage unit and to provide a respective output for each current claim transaction represented by the current claim data, the respective output indicating whether said each current claim transaction should be referred to an investigation unit;

adapt the predictive model component to changing patterns of questionable claim transactions by periodically updating the predictive model component based on new determinate data and new extracted indeterminate data; and

route, by a routing component, responsive to the respective output for each of the current claim transactions, selective ones of the current claim transactions to the investigation unit.

21. The computer system of claim 20, wherein the computer processor, under control by the routing component, routes ones of the current claim transactions via electronic mail messages to the investigation unit.

22. The computer system of claim 21, wherein investigative analysts in the investigation unit are recipients of the electronic mail messages.

23. The computer system of claim 20, wherein the indeterminate data capture module includes an optical character reader for converting paper documents to text files.

24. The computer system of claim 23, wherein the indeterminate data capture module includes:

a data extraction component for controlling the computer processor to extract information from the text files.

25. The computer system of claim 20, wherein:

the indeterminate data includes claim handlers'"'"' narrative text notes; and

the indeterminate data capture module includes;

a data extraction component for controlling the computer processor to extract information from the claim handlers'"'"' narrative text notes.

26. The computer system of claim 25, wherein the extracted information includes information indicative of the claim handlers'"'"' opinions.

27. The computer system of claim 20, wherein:

the indeterminate data includes audio information files, the audio information files including recordings of claimants'"'"' voices.

28. The computer system of claim 20, wherein said computer processor is further configured to:

process the indeterminate data for said each current claim transaction to detect at least one of;

(a) an indication that said each current claim transaction is disputed;

(b) an indication that a claimant engaged in physical activity; and

(c) an indication that said claimant made inconsistent statements.

29. A method of operating a computer system, the method comprising:

receiving determinate data corresponding to the historical claim transactions and the current claim transactions;

receiving raw indeterminate data corresponding to the historical claim transactions and the current claim transactions;

extracting, from the raw indeterminate data by an indeterminate data capture module, indeterminate data comprising at least one of claim handler notes, claim handler audio files, witness notes, witness audio files, claimant notes, and claimant audio files, the indeterminate data capture module including a natural language processing application to mine opinions from narrative text notes and a speech recognition application to analyze audio files to detect a speaker'"'"'s state of mind;

storing, in a computer storage unit, the determinate data and the information extracted from the indeterminate data as historical claim data;

using the historical claim data to train and verify a predictive model, the predictive model for evaluating claim transactions to determine whether to refer the claim transactions to an investigation unit;

adapting the predictive model component to changing patterns of questionable claim transactions by periodically updating the predictive model component based on new determinate data and new extracted indeterminate data;

using the plurality of input devices to provide determinate data and indeterminate data for current claim transactions;

processing the indeterminate data for the current claim transactions to extract information from the indeterminate data for the current claim transactions;

storing, in the computer storage unit, the determinate data for the current claim transactions and the information extracted from the indeterminate data for the current claim transactions;

identifying at least a first current claim transaction as one suitable for analysis by the trained and verified predictive model, wherein the first current claim transaction comprises a claim corresponding to at least one category of claims that are highly unlikely to be questionable;

applying the trained and verified predictive model to the determinate data for the at least first current claim transaction and to the information extracted from the indeterminate data for the at least first current claim transaction, to generate an output for the at least first current claim transaction; and

selectively routing the at least first current claim transaction to an investigative unit based on the outputs generated by the predictive model.

30. The method of claim 29, wherein the processing the indeterminate data includes applying optical character recognition.

31. The method of claim 29, wherein the processing the indeterminate data includes parsing text files.

32. The method of claim 29, wherein the processing the indeterminate data includes analyzing recordings of claimants'"'"' voices.

33. The method of claim 29, wherein the processing steps include:

processing the indeterminate data for said current claim transactions to detect at least one of;

(a) an indication that one of said current claim transactions is disputed;

(b) an indication that a claimant engaged in physical activity; and

(c) an indication that said claimant made inconsistent statements.

34. The method of claim 29, wherein the extracted information includes claim handlers'"'"' opinions.