Method and system for determining the importance of individual variables in a statistical model

• US 8,200,511 B2
• Filed: 11/28/2001
• Issued: 06/12/2012
• Est. Priority Date: 11/28/2001
• Status: Expired due to Fees

• Current Assignee: Deloitte Development LLC
• Original Assignee: Deloitte Development LLC
• Inventors: Zizzamia, Frank M., Wu, Cheng-Sheng Peter, Stukel, Raymond E., Adamopoulos, Hrisanthi, Lucker, John R.
• Primary Examiner(s): Rines, R. David
• Application Number:
  US09/996,065
• Publication Number:
  US 20030101080A1
• Time in Patent Office: 3,849 Days
• Field of Search: 705/4
• US Class Current: 705/4
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

A method and system for determining the importance of each of the variables that contribute to the overall score of a model for predicting the profitability of an insurance policy. For each variable in the model, an importance is calculated based on the calculated slope and deviance of the predictive variable. Since the score is developed using complex mathematical calculations combining large numbers of parameters with predictive variables, it is often difficult to interpret from the mathematical formula for example, why some policyholders receive low scores while other receive high scores. Such clear communication and interpretation of insurance profitability scores is critical if they are used by the various interested insurance parties including policyholders, agents, underwriters, and regulators.

Claims

1. A system for calculating the relative importance of one or more individual input variables to an overall insurance policy profitability score, comprising:

a database for storing values for various input variables;

a display; and

at least one data processor configured to;

receive an insurance policy profitability scoring formula, said insurance policy profitability scoring formula comprising a sum of a plurality of predictive input variables each having a weighting co-efficient, values for some or all said variables being stored in the database;

calculate a score using said scoring formula and a set of input variable values corresponding to a particular proposed or actual insured insurance policy;

calculate a partial derivative of the scoring formula with respect to each of the input variables in said set;

calculate a deviance value for each of the input variables in said set;

calculate the relative importance of each of the input variables in said set to a score calculated using the scoring formula by multiplying the partial derivative and deviance values for that variable;

create a rank for each of the input variables in said set based on its relative importance to the score; and

display the score, and one or more of said input variables in said set, its value and its rank to a user.

2. The system of claim 1 wherein the data processor further executes a software module that takes the first derivative of the scoring formula with respect to each a selected predictive input variable.

3. The system of claim 1 wherein the input variables in said set further describe characteristics of at least one of an existing policyholder and a potential policyholder, and wherein the scoring formula is used to generate a score reflective of the expected loss/premium ratio for an insurance policy.

4. The system of claim 3 wherein a premium for the insurance policy is based on the score.

5. The system of claim 1 wherein the data processor further executes a software module that receives inputs for a mean value and a standard deviation value, and the deviance value is calculated using the formula:

Deviance of x_i=(x_i−

μ

_i)/σ

_i

where μ

_i is the mean for x_i and σ

_i is the standard deviation for predictive input variable x_i.

6. The method of claim 1 wherein a score generated by the model determines the price for the insurance policy and the contribution is used to identify which of the predictive input variables had the greatest effect on the price.

7. The method of claim 1, further comprising populating the database with a mean value and a standard deviation value for each of the plurality of predictive variables.

8. The system of claim 1, wherein each of said input variables, its value and its rank is displayed to a user.

9. The system of claim 1, wherein said data processor is further configured to generate one or more insurance policy profitability scoring formulae, and to use one of them as the received scoring formula.

10. A computerized method of evaluating the relative individual contribution of one or more predictive input variables to an insurance policy profitability score, comprising:

receiving a scoring formula based on a multivariate statistical model to predict insurance policy profitability, said scoring formula comprising a sum of predictive input variables each having a weighting co-efficient;

receiving values for a set comprising a plurality of said predictive input variables wherein the received value correspond to a particular proposed or insured insurance policy;

calculating a score using the scoring formula and the received values;

calculating a partial derivative of the scoring formula with respect to each of the plurality of predictive input variables;

populating a database with a mean value and standard deviation value for each of the plurality of predictive input variables;

calculating a deviance value based on the mean value and the standard deviation value for each of the plurality of predictive input variables, wherein said calculation is performed by one of a data processor and a digital computer;

multiplying the deviance value and partial derivative value for each of the plurality of predictive input variables to determine the contribution of each of the plurality of predictive input variables to a given score;

creating a rank for each of the plurality of input variables based on its relative importance to the score;

and

displaying said score, one or more of the plurality of input variables, its received value and its rank to a user.

11. The method of claim 10, further comprising defining at least one assumption for the mean value associated with at least one of the plurality of predictive input variables.

12. The method of claim 10 wherein said calculating the slope further comprises calculating the first derivative of the scoring formula with respect to the predictive variable of the plurality of predictive variables that is being analyzed.

13. The method of claim 10 wherein the deviance value is calculated as follows:

Deviance of x_i=(x_i−

μ

_i)/σ

_i

where μ

_i is the mean for x_i and σ

_i is the standard deviation for predictive variable x_i.

14. The method of claim 10, wherein each of said input variables, its value and its rank is displayed to a user.

15. The method of claim 10, further comprising generating one or more insurance policy profitability scoring formulae, and using one of them as the received scoring formula.