Probabilistic model generation method, apparatus, and program

• US 7,899,767 B2
• Filed: 08/01/2006
• Issued: 03/01/2011
• Est. Priority Date: 08/12/2005
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Nishikawa, Takeichiro, Nakatsugawa, Minoru
• Primary Examiner(s): Sparks; Donald
• Assistant Examiner(s): OLUDE AFOLABI, OLATOYOSI
• Application Number:
  US11/496,718
• Publication Number:
  US 20070038558A1
• Time in Patent Office: 1,673 Days
• Field of Search: 706/15, 706/45, 706/52
• US Class Current: 706/45
• 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 40/03 Credit; Loans; Processing thereof

Abstract

There is provided with a method, including: optimizing a first objective function defined by using an explanatory variable belonging an attribute in each sample, a target variable in each sample, and a first conversion parameter to find a value of the first conversion parameter; generating by using the first conversion parameter corresponding to the attribute a conversion function for converting an explanatory variable belonging to the attribute to an intermediate variable with certain range; optimizing a second objective function defined by using a plurality of the intermediate variables corresponding to the plurality of variables in each sample, the target variable in each sample, and a second conversion parameter to find a value of the second conversion parameter; and generating by using the second conversion parameter a probabilistic model for calculating from a plurality of intermediate variables a probability that a predetermined event occurs or does not occur.

Claims

1. A bankruptcy probabilistic model generation apparatus including a computer readable storage medium containing a set of instructions that cause a computer processor to perform a data analyzing process for generating a bankruptcy probabilistic model calculating a probability that a bankruptcy or an accident occurs or does not occur, comprising:

a database storing financial data as a set of samples each of which includes a plurality of explanatory variables belonging to respectively different financial attributes and a target variable representing occurrence or nonoccurrence of a bankruptcy or an accident;

an upper and lower limit setter for each financial attribute, which optimizes a first objective function defined by using the variable belonging to the financial attribute in each sample, the target variable in each sample, and a first conversion parameter to find a value of the first conversion parameter, and calculates a first conversion function defined by the first conversion parameter and the variable belonging to the financial attribute to convert the variable to an intermediate variable with a certain range of value for each sample of the financial data, wherein the certain range of value has an upper limit, a lower limit or both an upper and lower limit; and

a bankruptcy probability calculator, which optimizes a second objective function defined by using a plurality of intermediate variables corresponding to the plurality of variables in each sample of the financial data, the target variable in each sample of the bankruptcy data, and a second conversion parameter to find a value of the second conversion parameter, and calculates a second conversion function defined by the second conversion parameter and a plurality of intermediate variables belonging to the financial attributes to obtain a bankruptcy probability that the bankruptcy or accident occurs or does not occur for each sample of the financial data,

wherein the first and second objective functions are logarithmic likelihood functions.

2. The apparatus according to claim 1, wherein:

the upper and lower limit setter or the bankruptcy probability calculator performs optimizing by maximizing the corresponding one of the logarithmic likelihood functions.

3. The apparatus according to claim 1, wherein the first conversion function is a logistic function.

4. The apparatus according to claim 3, wherein the logistic function has an nth-order term of a variable (where n is an integer of at least 2), or a logarithmic term of a variable.

5. The apparatus according to claim 1, wherein:

the first conversion function is a broken line function,

when a variable to be input into the broken line function is a first threshold or more, the broken line function outputs a first constant value as an intermediate variable, and

when a variable to be input into the broken line function is a second threshold or less, the broken line function outputs a second constant value as an intermediate variable.

6. The apparatus according to claim 1, wherein

the second conversion function is a binary logit model which has a linear combination of a plurality of intermediate variables.

7. A bankruptcy probabilistic model generation apparatus including a computer readable storage medium containing a set of instructions that cause a computer processor to perform a data analyzing process for generating a bankruptcy probabilistic model calculating a probability that a bankruptcy or an accident occurs or does not occur, comprising:

a database storing financial data as a set of samples each of which includes a plurality of explanatory variables belonging to respectively different financial attributes and a target variable representing occurrence or nonoccurrence of a bankruptcy or an accident;

an upper and lower limit setting and bankruptcy probability calculator which optimizes an objective function defined by using the plurality of variables in each sample, the target variable in each sample, a first conversion parameter provided for every financial attribute and a second conversion parameter, to find values of the first conversion parameters and a value of the second conversion parameter, calculates a first conversion function defined by the first conversion parameter and the variable belonging to the financial attribute to convert the variable to an intermediate variable with a certain range of value for each financial attribute in each sample of the financial data, wherein the certain range of value has an upper limit, a lower limit or both an upper limit and a lower limit, and calculates a second conversion function defined by the second conversion parameter and a plurality of intermediate variables belonging to the financial attributes to obtain a bankruptcy probability that the bankruptcy or accident occurs or does not occur for each sample of the financial data,

wherein the objective function is a logarithmic likelihood function.

8. The apparatus according to claim 7, wherein

the upper and lower limit setting and bankruptcy probability calculator performs optimizing by maximizing the logarithmic likelihood function.

9. The apparatus according to claim 7, wherein the first conversion function is a logistic function.

10. The apparatus according to claim 9, wherein the logistic function has an nth-order term of a variable (where n is an integer of at least 2), or a logarithmic term of a variable.

11. The apparatus according to claim 7, wherein:

the first conversion function is a broken line function,

when a variable to be input into the broken line function is a first threshold or more, the broken line function outputs a first constant value as an intermediate variable, and

when a variable to be input into the broken line function is a second threshold or less, the broken line function outputs a second constant value as an intermediate variable.

12. The apparatus according to claim 7, wherein the second conversion function is a binary logit model which has a linear combination of a plurality of intermediate variables.

13. A computer readable storage medium storing a computer program including instructions for causing a computer processor generating a bankruptcy probabilistic model calculating a probability that a bankruptcy or an accident occurs or does not occur, to perform steps comprising:

accessing a database storing financial data as a set of samples each of which includes a plurality of explanatory variables belonging to respectively different financial attributes and a target variable representing occurrence or nonoccurrence of a bankruptcy or an accident;

for every financial attribute, optimizing a first objective function defined by using the variable belonging to the financial attribute in each sample, the target variable in each sample, and a first conversion parameter to find a value of the first conversion parameter, and calculating a first conversion function defined by the first conversion parameter and the variable belonging to the financial attribute to convert the variable to an intermediate variable with certain range of value for each sample of the financial data, wherein the certain range of value has an upper limit, a lower limit or both an upper limit and a lower limit; and

optimizing a second objective function defined by using a plurality of intermediate variables corresponding to the plurality of variables in each sample of the financial data, the target variable in each sample of the bankruptcy data, and a second conversion parameter to find a value of the second conversion parameter, and calculating a second conversion function defined by the second conversion parameter and a plurality of intermediate variables belonging to the financial attributes to obtain a bankruptcy probability that the bankruptcy or accident occurs or does not occur for each sample of the financial data,

wherein the first and second objective functions are logarithmic likelihood functions.