System to determine a credibility weighting for electronic records

• US 10,445,354 B2
• Filed: 10/05/2016
• Issued: 10/15/2019
• Est. Priority Date: 10/05/2016
• 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: Houser, Reuben Emory, Wang, Qiao, Drennan, III, Arthur Paul, Sims, Nicole Rose, Steger, Tracey Ellen
• Primary Examiner(s): Reyes, Mariela
• Assistant Examiner(s): Harmon, Courtney
• Application Number:
  US15/285,790
• Publication Number:
  US 20180096055A1
• Time in Patent Office: 1,105 Days
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Text input data may be aggregated and mapped to create composite text input data for electronic records. A semantic event may be automatically detected, triggered by a semantic rule and associated semantic tag. The detected semantic event may be flagged, and a text mining result database may be updated by adding an entry to the database. An indication associated with the event may be transmitted, and a back-end application computer server may establish a hierarchy for multiple elements of the electronic records. The computer server may determine a weight variable and response variable for each element in the hierarchy and apply a hierarchical credibility weighting methodology to the elements from level 1, representing the least granular level in the hierarchy, to level n, representing the most granular level in the hierarchy, calculated recursively from level n, to determine a final estimated credibility for the electronic records.

Claims

1. A system for determining a credibility weighting associated with electronic records, comprising:

a text mining platform, including;

a text mining communication device to receive text input data associated with the electronic records;

a text mining processor coupled to the text mining communication device; and

a text mining storage device in communication with said text mining processor and storing instructions adapted to be executed by said text mining processor to;

(i) aggregate and map the received text input data to create composite text input data,

(ii) automatically detect a semantic event in the composite text input data triggered by a semantic rule and associated semantic tag,

(iii) flag the detected semantic event as meeting a pre-determined condition,

(iv) update a text mining result database, responsive to the flag, by adding an entry to the database identifying the detected semantic event and the triggering semantic rule, and

(v) transmit an indication associated with the event based on the associated data in the text mining result database; and

a back-end application computer server coupled to the text mining platform, including;

a back-end communication device to receive the indication for the electronic record transmitted by the text mining platform;

a back-end processor coupled to the back end communication device; and

a back-end storage device in communication with said back-end processor and storing instructions adapted to be executed by said back-end processor to;

(i) establish a hierarchy for multiple elements of the electronic records,

(ii) determine a weight variable and response variable for each element in the established hierarchy,

(iii) apply a hierarchical credibility weighting methodology to the elements of the electronic records from level 1, representing the least granular level in the hierarchy, to level n, representing the most granular level in the hierarchy, calculated recursively from level n, to determine a final estimated credibility for the electronic records based on the weight variable and response variable, wherein ω

_i represents a weight for level i, X_i represents an observed experience for level i, K represents a credibility factor, and the credibility for level i is calculated as;

2. The system of claim 1, wherein if there is not enough data associated with a granular level in the hierarchy to be credible, then the level of granularity is decreased until enough data is obtained for the level to be credible.

3. The system of claim 1, wherein the semantic event is associated with at least one of:

(i) a word, (ii) a phrase, (iii) a shorthand term, (iv) a course of action, and (v) an enterprise name.

4. The system of claim 1, wherein the triggering semantic rule is associated with at least one of:

(i) a noun, (ii) a verb, (iii) a definition, (iv) a semantic tree, (v) a named entity recognition rule, (vi) a root, (vii) a noun phrase, (viii) a prepositional phrase, and (ix) a verb phrase.

5. The system of claim 4, wherein the triggering semantic rule was defined by an administrator using a graphical user interface.

6. The system of claim 1, wherein the observed experience for level i, excluding lower levels in the hierarchy, is calculated as:

7. The system of claim 6, wherein the credibility for level i, excluding lower levels in the hierarchy, is calculated as:

Z_i′

=Z_i−

Z_i+1.

8. The system of claim 7, wherein the credibility weighted estimate for level n is calculated as:

{circumflex over (X)}_n=Σ

_i=1ⁿX_i′

*Z_i′

.

9. The system of claim 1, wherein the electronic records are associated with insurance claims, the weight variable comprises a text flag total claim count, and the response variable comprises a text flag ratio.

10. The system of claim 9, wherein the insurance claims comprise workers'"'"' compensation insurance claims.

11. The system of claim 10, wherein the final estimated credibility is associated with at least one of claim litigiousness and adverse claim severity summarized at a geographic level.

12. The system of claim 11, wherein the established hierarchy includes at least four of:

(i) effective year, (ii) a risk state, (iii) a Combined Statistical Area (“

CSA”

), (iv) a Core Based Statistical Area (“

CBSA”

), (v) a Federal Information Processing Standard (“

FIPS”

) code, (vi) a county, and (vii) a five-digital ZIP code.

13. The system of claim 1, wherein higher levels of the established hierarchy exclude the experience and credibility from the lower levels, credibility weighted values of the full hierarchy are set relative to the credibility weighted values of the hierarchy to normalize the data, and the final estimated credibility is used in connection with at least one of:

(i) an insurance pricing process, and (ii) a risk score model.

14. The system of claim 1, wherein the text input data is associated with at least one of:

(i) an insurance claim file, (ii) an insurance claim note, (iii) a medical report, (iv) a police report, (v) social network data, (vi) big data information, (vii) a loss description, (viii) an injury description, (ix) a first notice of loss statement, (x) telephone call transcript, (xi) optical character recognition data, (xii) third-party data, and (xiii) a governmental agency.

15. A computer-implemented method for determining a credibility weighting associated with electronic records, comprising:

aggregating and mapping, by a text mining platform processor, received text input data to create composite text input data for the electronic records;

automatically detecting, by the text mining platform processor, a semantic event in the composite text input data triggered by a semantic rule and associated semantic tag;

flagging, by the text mining platform processor, the detected semantic event as meeting a pre-determined condition;

updating, by the text mining platform processor, a text mining result database, responsive to the flag, by adding an entry to the database identifying the detected semantic event and the triggering semantic rule;

transmitting, by the text mining platform processor, an indication associated with the event based on the associated data in the text mining result database;

establishing, by a back-end application computer processor, a hierarchy for multiple elements of the electronic records;

determining, by the back-end application computer processor, a weight variable and response variable for each element in the established hierarchy;

applying, by the back-end application computer processor, a hierarchical credibility weighting methodology to the elements of the electronic records from level 1, representing the least granular level in the hierarchy, to level n, representing the most granular level in the hierarchy, calculated recursively from level n, to determine a final estimated credibility for the electronic records based on the weight variable and response variable, wherein ω

_i represents a weight for level i, X_i represents an observed experience for level i, K represents a credibility factor, and the credibility for level i is calculated as;

16. The method of claim 15, wherein the observed experience for level i, excluding lower levels in the hierarchy, is calculated as:

17. The method of claim 16, wherein the credibility for level i, excluding lower levels in the hierarchy, is calculated as:

Z_i′

=Z_i−

Z_i+1, and

the credibility weighted estimate for level n is calculated as;

{circumflex over (X)}_n=Σ

_i=1ⁿX_i′

*Z_i′

.

18. The method of claim 15, wherein the electronic records are associated with insurance claims, the weight variable comprises a text flag total claim count, and the response variable comprises a text flag ratio.

19. A system for determining a credibility weighting associated with electronic records, comprising:

a text mining platform, including;

a text mining communication device to receive text input data associated with the electronic records;

a text mining processor coupled to the text mining communication device; and

a text mining storage device in communication with said text mining processor and storing instructions adapted to be executed by said text mining processor to;

(i) aggregate and map the received text input data to create composite text input data,

(ii) automatically detect a semantic event in the composite text input data triggered by a semantic rule and associated semantic tag, and

(iii) transmit an indication associated with the event based on the associated data in the text mining result database; and

a back-end application computer server coupled to the text mining platform, including;

a back-end communication device to receive the indication for the electronic record transmitted by the text mining platform;

a back-end processor coupled to the back end communication device; and

a back-end storage device in communication with said back-end processor and storing instructions adapted to be executed by said back-end processor to;

(i) establish a hierarchy for multiple elements of the electronic records,

(ii) determine a weight variable and response variable for each element in the established hierarchy,

(iii) apply a hierarchical credibility weighting methodology to the elements of the electronic records from level 1, representing the least granular level in the hierarchy, to level n, representing the most granular level in the hierarchy, calculated recursively from level n, to determine a final estimated credibility for the electronic records based on the weight variable and response variable, wherein ω

_i represents a weight for level i, X_i represents an observed experience for level i, K represents a credibility factor, and the credibility for level i is calculated as;

20. The system of claim 19, wherein the semantic rule and associated semantic tag were graphically defined by a user.