Automated method for identification of reinsurance claims

• US 6,049,773 A
• Filed: 10/14/1997
• Issued: 04/11/2000
• Est. Priority Date: 10/14/1997
• Status: Expired due to Fees

• Current Assignee: Reclaim Technology and Services Limited
• Original Assignee: Reclaim Technology and Services Limited
• Inventors: McCormack, Douglas R., McCormack, James S.
• Primary Examiner(s): Voeltz, Emanuel Todd
• Assistant Examiner(s): Morgan, George D.
• Application Number:
  US08/943,298
• Time in Patent Office: 910 Days
• Field of Search: 705/4, 705/7, 705/10, 705/35, 705/36, 705/2, 706/925, 706/52, 706/21, 706/60, 706/47, 707/104
• US Class Current: 705/4
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

An automated method for identifying those claims of a raw claims insurance claims database for which reinsurance is applicable. The method develops a database of uniquely clustered catastrophic events such as storm reports. This "CatNodes" database may be developed manually or automatically through the use of a neural network approach. A fuzzy degree of belonging is employed to quantify the likelihood that a given insurance claim is properly associated with a given catastrophic event or storm. The assignment of a degree of belonging is derived by approaches which consider four factors: the date of the loss; the location of the loss; the type of the loss; and the presence of special keywords in the claim description.

Claims

1. A method for identifying insurance claims having a relationship with reinsurance related catastrophic events, comprising the steps of:

providing a catastrophic database having identified cluster nodes of catastrophic events derived by reported parameters, each said cluster node having an earth position parameter, and a date of occurrence parameter;

providing an insurance claims database of claim reports, each such claim report thereof having, an earth location datum and a date datum;

deriving a place degree of belonging for each said earth location datum of each said claim report with respect to each said cluster node earth position parameter;

deriving a date degree of belonging for each said date datum of each said claims report with respect to each said date of occurrence parameter;

selecting a said identified cluster node for each said claim report as that best cluster node optimally corresponding with the derived place degree of belonging and date degree of belonging; and

generating a tabular claims file compilation incorporating data from each said claim report, each said claim report being associated with said selected identified best cluster node, with said best place degree of belonging and date degree of belonging.

2. A method for producing, from a memory retained group of insurance claim reports, a compilation of insurance claims having a relationship with catastrophes for which reinsurance coverage is applicable, comprising the steps of:

providing a city database wherein a plurality of cities name entries are retained in memory, each being retained in combination with geographic coordinates identifying their position on earth;

opening a claims database having fields including a loss city field, a loss state field, a claim identification field, and a loss date field;

importing claim data available from each said claim report of said memory retained group of insurance claims reports into corresponding said fields;

for each said claim report, scanning said city database to identify a city name match and retaining said geographic coordinates as claim geographic coordinates in memory in conjunction with said claims report when a city name match occurs;

providing a catastrophic database having identified clusters of catastrophic events derived by reported parameters, each said cluster having an earth position parameter and a date of occurrence parameter;

deriving a degree of belonging corresponding with each said claim geographic coordinates with respect to each said cluster earth position parameter, and corresponding with each said loss date of each said claim report with respect to each said cluster date of occurrence parameter;

selecting the best cluster representing a best said degree of belonging for each said claim report; and

generating a tabular claims file compilation incorporating data from each said claim report, each said claims report tabulated therein being combined with said selected best cluster of catastrophic events.

3. The method of claim 2 including the steps of:

providing a zip code database wherein a plurality of postal zip codes are retained in memory in combination with geographic coordinates identifying their position on earth; and

for each said claim report, scanning said zip code database to identify a zip code match in the absence of a city name match and retaining said geographic coordinates as claim geographic coordinates in memory in conjunction with said claim report when a zip code match occurs.

4. The method of claim 3 including the steps of:

providing a telephone number database wherein a plurality of components of telephone numbers are retained in memory in combination with geographic coordinates identifying their position on earth; and

for each said claim report, scanning said telephone number database to identify a telephone number match in the absence of a said zip code match and retaining said geographic coordinates as claims geographic coordinates in memory in conjunction with said claims report when a telephone match occurs.

5. The method of claim 4 in which:

said city database is provided having a state name entry in conjunction with each said city name and is provided having a state average entry in conjunction with a said state name entry, each said state average entry being retained in memory in combination with geographic coordinates identifying a predetermined state contained position on earth; and

including the step;

for each said claim report, scanning said city database to identify a state name match in the absence of a said telephone match and further scanning said city database to identify that state average entry corresponding with said state name match, and retaining said geographic coordinates identifying a predetermined state contained position on earth as claim geographic coordinates in memory when said state average entry is identified.

6. The method of claim 3 including the steps of:

providing a telephone number database wherein a plurality of components of telephone numbers are retained in memory in combination with geographic coordinates identifying their position on earth; and

for each said claim report, scanning said telephone number database to identify a telephone number match in the absence of a said city name match and retaining said geographic coordinates as claims geographic coordinates in memory in conjunction with said claims report when a telephone match occurs.

7. The method of claim 3 in which:

said city database is provided having a state name entry in conjunction with each said city name and is provided having a state average entry in conjunction with a said state name entry, each said state average entry being retained in memory in combination with geographic coordinates identifying a predetermined state contained position on earth; and

including the step;

for each said claim report, scanning said city database to identify a state name match in the absence of a said zip code match and further scanning said city database to identify that state average entry corresponding with said state name match, and retaining said geographic coordinates identifying a predetermined state contained position on earth as claim geographic coordinates in memory when said state average entry is identified.

8. The method of claim 2 in which:

said city database is provided having a state name entry in conjunction with each said city name; and

including the step;

for each said claim report, scanning said city database to identify a state name match in the presence of a said city name match and carrying out a retention of said geographic coordinates as clam geographic coordinates in memory only in the presence of both said city name match and said state name match.

9. The method of claim 2 in which:

said city database is provided having a state name entry in conjunction with each said city name and is provided having a state average entry in conjunction with a said state name entry, each said state average entry being retained in memory in combination with geographic coordinates identifying a predetermined state contained position on earth; and

including the step;

for each said claim report, scanning said city database to identify a state name match in the absence of a said city name match and further scanning said city database to identify that state average entry corresponding with said state name match, and retaining said geographic coordinates identifying a predetermined state contained position on earth as claim geographic coordinates in memory when said state average entry is identified.

10. The method of claim 2 including the steps of:

providing a loss types database wherein a plurality of predetermined types of insurable loss entries are retained in memory;

said importing claim data step includes the step of importing loss data from said claim reports into a corresponding loss type field to provide a loss type entry in association with each said claim report;

providing a loss-type to degree of belonging database wherein each of said predetermined types of insurable loss entries are assigned a loss-type degree of belonging;

said step of selecting said best cluster is carried out in conjunction with said loss-type degree of belonging when the value thereof is less than the value of said best degree of belonging.

11. A method for identifying insurance claim having a relationship with reinsurance related catastrophic events, comprising the steps of:

(a1) providing a catastrophic database as a memory retained compilation of catastrophic event records, each such record having an event identification, a date of event occurrence, and a catastrophic geographic coordinate identifying its position on earth;

(a2) selecting a first place vigilance value;

(a3) selecting a first date vigilance value;

(a4) opening a catnodes database;

(a5) selecting an initial node as one said event record with a node coordinate and a node date corresponding respectively with said catastrophic geographic coordinates and said date of event occurrence;

(a6) scanning said catastrophic database and for each record thereof, deriving a place degree of belonging in correspondence with said node coordinate and said catastrophic geographic coordinate, and further deriving a date degree of belonging in correspondence with said node date and said date of event occurrence;

(a7) comparing the value of said place degree of belonging with the said first place vigilance value and diminishing said place degree of belonging when the value thereof is less than said first place vigilance value;

(a8) comparing the value of said date degree of belonging with said first date vigilance value and diminishing said date degree of belonging when the value thereof is less than said first date vigilance value;

(a9) determining an overall degree of belonging in correspondence with said place degree of belonging and said date degree of belonging;

(a10) determining a best degree of belonging from all overall degree of belonging values derived from said scanning step (a6) and assigning the place degree of belonging and the date degree of belonging corresponding with said best degree of belonging respectively as best place degree of belonging and best date degree of belonging;

(a11) creating a new node corresponding with that event record deriving said best degree of belonging when the value of said best place degree of belonging is less than said first place vigilance value or when the value of said best date degree of belonging is less than said first date vigilance value;

(a12) entering said new node into said catnodes database;

(a13) adjusting said node coordinates and said node date in correspondence with said best place degree of belonging and best date degree of belonging when the value of said best place degree of belonging is greater than said first place vigilance value and said best date degree of belonging is greater than said first date vigilance value;

(a14) reiterating steps (a6) through (a13) for each new node created to form a catnodes database;

(b) providing an insurance claims database of claim reports, each such claim report thereof having claim place geographic coordinates identifying a position on earth and a claim loss date;

(c) deriving a claim place degree of belonging corresponding with each of said claim place geographic coordinates with respect to each said node coordinate of said catnodes database;

(d) deriving a claims date degree of belonging corresponding with each said claim loss date with respect to each said node date of said catnodes database;

(e) deriving a claims overall degree of belonging corresponding with each said claims place degree of belonging and said claim loss date degree of belonging;

(f) selecting the best node of said catnodes database representing a best overall degree of belonging for each said claim report; and

generating a tabular ClaimsFile compilation incorporating data from each said claim report, each said claim report tabulated therein being combined with said selected best node.

12. The method of claim 11 including the step:

(a15) amalgamating said created nodes into cluster nodes in accordance with a cluster place vigilance value less than said first place vigilance value and in accordance with a cluster date vigilance value less than said first date vigilance value.

13. The method of claim 11 in which said step (a6) includes the steps of:

(a6i) determining the distance between said node coordinates and said catastrophic, geographic coordinates; and

(a6ii) deriving said place degree of belonging as the value of one minus the value of said distance divided by the value of a predetermined terrestrial domain representing a predetermined distance extending from the location of said node coordinates.

14. The method of claim 13 in which said predetermined distance is an integer fraction of the radius of the earth.

15. The method of claim 13 in which said step (a7) diminishes said place degree of belonging to zero when the value thereof is less than said first place vigilance value.

16. The method of claim 13 in which said step (a6) includes the steps of:

(a6iii) determining the days apart interval between said node date and said date of event occurrence; and

(a6iv) deriving said date degree of belonging as the value of one minus the value of said days apart interval divided by a predetermined time domain.

17. The method of claim 16 in which said predetermined time domain is a predetermined number of days.

18. The method of claim 17 in which said predetermined number of days is 365.

19. The method of claim 16 in which said step (a8) diminishes said date degree of belonging to zero when the value thereof is less than said first date vigilance value.

20. The method of claim 11 in which said step (a9) overall degree of belonging is the average of said place degree of belonging and said date degree of belonging.

21. The method of claim 11 in which said step (a11) includes the step of:

(a11i) assigning said new node said corresponding event record date of occurrence and catastrophic geographic coordinates.

22. The method of claim 11 in which said step (a13) includes the steps of:

(a13i) determining the number, K, of adjustments which will have been made to said node to be adjusted;

(a13ii) adjusting said node coordinates toward the catastrophic geographic coordinates of that record corresponding with said best place degree of belonging in correspondence with the ratio 1/K; and

(a13iii) adjusting said node date toward the record date of event occurrence corresponding with said best date degree of belonging.

23. The method of claim 11 including the steps of:

(a15) selecting a second place vigilance value less than said first place vigilance value;

(a16) selecting a second date vigilance value less than said first date vigilance value;

(a17) selecting an initial cluster node from said catnodes database and assigning it a cluster identifier;

(a18) scanning said catnodes database and for each scanned node thereof deriving a node place degree of belonging with respect to the identifier assigned node node coordinate and the node coordinates of said scanned node, and further deriving a node date degree of belonging with respect to the identifier assigned node node date and the node date of said scanned node;

(a19) comparing the value of said node place degree of belonging with said second place vigilance value and diminishing said node place degree of belonging when the value thereof is less than said second place vigilance value;

(a20) comparing the value of said node date degree of belonging with said second date vigilance value and diminishing said node date degree of belonging when the value thereof is less than said second date vigilance value;

(a21) determining a node overall degree of belonging in correspondence with said node place degree of belonging and said node date degree of belonging;

(a22) determining a node best degree of belonging from all node overall degree of belonging values derived from said scanning step (a18) and assigning the node place degree of belonging and the node date degree of belonging corresponding with said node best degree of belonging respectively as node best degree of belonging and node best date degree of belonging;

(a23) creating a new cluster node and assigning a next cluster identifier thereto with that scanned node deriving said node best degree of belonging when the value of said node best place degree of belonging is less than said second place vigilance value or when the value of said node best date degree of belonging is less than said second date vigilance value;

(a24) entering said new cluster node into said catnodes database;

(a25) adjusting said cluster node node coordinates and node date in correspondence with said node best place degree of belonging and node best date degree of belonging when the value of said node best place degree of belonging is greater than said second place vigilance value and said node best date degree of belonging is greater than said second date vigilance value; and

(a26) reiterating steps (a18) through (a25) for each new cluster node created.

24. The method of claim 23 in which said step (a18) includes the steps of:

(a18i) determining the distance between the node coordinates of said cluster node and the node coordinates of the scanned node; and

(a18ii) deriving said node place degree of belonging as the value of one minus the value of said distance divided by the value of a predetermined terrestrial domain representing a predetermined distance extending from the location of the cluster node coordinates.

25. The method of claim 24 in which said predetermined distance is an integer fraction of the radius of the earth.

26. The method of claim 24 in which said step (a18) diminishes said node place degree of belonging to zero when the value thereof is less than said second place vigilance value.

27. The method of claim 24 in which said step (a20) includes the steps of:

(a20iii) determining the days apart interval between the node date of the cluster node and the node date of the scanned node; and

(a20iv) deriving said node date degree of belonging as the value of one minus the value of said days apart interval divided by a predetermined time domain.

28. The method of claim 27 in which said predetermined time domain is a predetermined number of days.

29. The method of claim 28 in which said predetermined number of days is 365.

30. The method of claim 27 in which said step (a20) diminishes said node date degree of belonging to zero when the value thereof is less than said second date vigilance value.

31. The method of claim 23 in which said step (a21) node overall degree of belonging is the average of said node place degree of belonging and said node date degree of belonging.

32. The method of claim 23 in which said step (a25) includes the steps of:

(a25i) determining the number, K, of adjustments which will have been made to said cluster node to be adjusted;

(a25ii) adjusting said cluster node coordinates toward the node coordinates of the scanned node corresponding with said node best degree of belonging in accordance with the ratio 1/K; and

(a25iii) adjusting the cluster node date toward the scanned node date corresponding with the node best date degree of belonging.

33. A method for identifying insurance claims having a relationship with reinsurance related catastrophic events, comprising the steps of:

providing a catastrophic database having identified cluster nodes of catastrophic events derived by reported parameters, each said cluster node having an earth position parameter, and a date of occurrence parameter;

providing an insurance claims database of claim reports, each such claim report thereof having an earth location datum and a date datum;

deriving a place degree of belonging for each said earth location datum of each said claim report with respect to each said cluster node earth position parameter;

deriving a date degree of belonging for each said date datum of each said claims report with respect to each said date of occurrence parameter;

selecting a said identified cluster node for each said claim report as that best cluster node optimally corresponding with the derived place degree of belonging and date degree of belonging by;

selectively combining the values of said place degree of belonging and said date degree of belonging to derive an overall degree of belonging, and

selecting said optimally corresponding best cluster node as that cluster node corresponding with the best said overall degree of belonging; and

generating a tabular claims file compilation incorporating data from each said claim report, each said claim report being associated with said selected identified best cluster node, with said best place degree of belonging and date degree of belonging.

34. The method of claim 33 in which:

said insurance claims database is provided including a loss-type datum for each said claim report;

each said loss-type datum being provided from a group of designated types of loss;

each said loss-type datum of said group of designated types of loss having a predetermined type of loss degree of belonging assigned thereto; and

said step of selecting said optimally corresponding best cluster node is carried out in conjunction with said predetermined type of loss degree of belonging when the value thereof is less than the value of said best overall degree of belonging.

35. The method of claim 34 in which:

said insurance claims database is provided including a memo datum corresponding with a said claim report;

including the steps of;

providing a keywords database formed of a plurality of predetermined keywords,

determining the presence of a said predetermined keyword within said memo datum; and

disregarding said type of loss degree of belonging when a said predetermined keyword is determined to be present within said memo datum.

36. A method for identifying insurance claims having a relationship with reinsurance related catastrophic events, comprising the steps of:

providing a catastrophic database having identified cluster nodes of catastrophic events derived by reported parameters, each said cluster node having an earth position parameter, and a date of occurrence parameter;

providing an insurance claims database of claim reports, each such claim report thereof having an earth location datum and a date datum;

deriving a place degree of belonging for each said earth location datum of each said claim report with respect to each said cluster node earth position parameter;

deriving a date degree of belonging for each said date datum of each said claims report with respect to each said date of occurrence parameter;

selecting a said identified cluster node for each said claim report as that best cluster node optimally corresponding with the derived place degree of belonging and date degree of belonging;

generating a tabular claims file compilation incorporating data from each said claim report, each said claim report being associated with said selected identified best cluster node, with said best place degree of belonging and date degree of belonging;

selecting a target identified cluster node; and

compiling a tabulation of all said claim records wherein said target identified cluster node is said best cluster node.

37. A method for identifying insurance claims having a relationship with reinsurance related catastrophic events, comprising the steps of:

providing a catastrophic database having identified cluster nodes of catastrophic events derived by reported parameters, each said cluster node having an earth position parameter, and a date of occurrence parameter;

providing an insurance claims database of claim reports, each such claim report thereof having an earth location datum and a date datum;

deriving a place degree of belonging for each said earth location datum of each said claim report with respect to each said cluster node earth position parameter;

deriving a date degree of belonging for each said date datum of each said claims report with respect to each said date of occurrence parameter;

selecting a said identified cluster node for each said claim report as that best cluster node optimally corresponding with the derived place degree of belonging and date degree of belonging;

generating a tabular claims file compilation incorporating data from each said claim report, each said claim report being associated with said selected identified best cluster node, with said best place degree of belonging and date degree of belonging;

and wherein;

said catastrophic database earth position parameter includes a fuzzy set with principal geographical coordinates and a territorial domain extending a predetermined distance therefrom; and

said place degree of belonging is derived in correspondence with the distance between said principal geographical coordinates and said earth location datum, divided by a fuzzy set territorial domain extending a predetermined distance from said geographical coordinates.

38. A method for identifying insurance claims having a relationship with reinsurance related catastrophic events, comprising the steps of:

providing a catastrophic database having identified cluster nodes of catastrophic events derived by reported parameters, each said cluster node having an earth position parameter, and a date of occurrence parameter;

providing an insurance claims database of claim reports, each such claim report thereof having an earth location datum and a date datum;

deriving a place degree of belonging for each said earth location datum of each said claim report with respect to each said cluster node earth position parameter;

deriving a date degree of belonging for each said date datum of each said claims report with respect to each said date of occurrence parameter, said date degree of belonging being derived in correspondence with the time difference between said claim report date datum and said date of occurrence parameter divided by a predetermined fuzzy set time domain value;

selecting a said identified cluster node for each said claim report as that best cluster node optimally corresponding with the derived place degree of belonging and date degree of belonging; and

generating a tabular claims file compilation incorporating data from each said claim report, each said claim report being associated with said selected identified best cluster node, with said best place degree of belonging and date degree of belonging.

39. The method of claim 38 in which said time difference is enlarged in value in correspondence with a predetermined tilt factor when said claim report date datum occurs prior in time to said date of occurrence parameter.