Residual value forecasting system and method thereof, insurance premium calculation system and method thereof, and computer program product

• US 20020072958A1
• Filed: 10/24/2001
• Published: 06/13/2002
• Est. Priority Date: 10/31/2000
• Status: Active Grant

• Current Assignee: Proto Corp., Tokio Marine NEWA Insurance Company Limited (Tokio Marine & Nichido Fire Insurance Co., Ltd.)
• Original Assignee: Proto Corp., Tokio Marine NEWA Insurance Company Limited (Tokio Marine & Nichido Fire Insurance Co., Ltd.)
• Inventors: Sakai, Takanori, Yuyama, Takuya, Okada, Shumpei, Ino, Kenji
• Granted Patent: US 7,366,679 B2
• Time in Patent Office: Days
• US Class Current: 705/10
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 30/0206 Price or cost determination based on market factors
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

Provided is technology for enabling an accurate forecasting of a residual value of goods based on data showing the similar trend, without making reduction of the number of data caused by ramification, when the subject goods have an attribute that may largely affect the residual value thereof. In order to achieve the above, the system according to the present invention comprises a secondhand price database storing transaction data including the residual value of the object and the elapsed year of the object etc.; means for calculating the standardized forecast residual value in a prescribed future based on the standardized residual value, which is calculated as if the attribute of the object is deemed to be a representative attribute value (standardized residual value), by modifying the residual value of the object included in the transaction data stored in the secondhand price database based on attribute values of a prescribed attribute of the object included in the transaction data and the elapsed time included in the transaction data; and means for seeking a non-standardized forecast residual value by making adjustments corresponding to the attribute values of the object of which residual value will be forecasted, to said standardized forecast residual value.

Claims

1. A residual value forecasting system comprising:

a secondhand price database which stores transaction data including a secondhand circulating price or a ratio of the secondhand circulating price to the original price (the secondhand circulating price and the ratio thereof to the original price are collectively referred to as “

residual value”

) of an object, attribute values of the object, and elapsed time of the object; and

forecast residual value calculation means which calculates a forecast residual value in a prescribed future based on the corresponding relationship between the residual value and elapsed time stored in said secondhand price database.

2. The residual value forecasting system according to claim 1, wherein said object is a car.

3. The residual value forecasting system according to claim 1, wherein said forecast residual value calculation means comprises:

standardized forecast residual value calculation means which adjusts the residual value of the object included in the transaction data stored in said secondhand price database, based on an attribute value of a prescribed attribute of the object included in said transaction data, calculates the residual value as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

), and calculates a standardized forecast residual value in a prescribed future based on such standardized residual value and the elapsed time included in said transaction data; and

non-standardized forecast residual value calculation means which seeks a non-standardized forecast residual value by making adjustment corresponding to the attribute value of the object of which residual value will be forecasted, to said standardized forecast residual value.

4. A residual value forecasting system according to claim 1, wherein said forecast residual value calculation means comprises:

standardized forecast residual value calculation means which calculates a forecast residual value in a prescribed future as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

) (such residual value hereinafter to be referred to as “

standardized forecast residual value”

); and

non-standardized forecast residual value calculation means which randomly selects an attribute value for the attribute of the object based on the component ratio of the attributes, and seeks a non-standardized forecast residual value by making adjustment corresponding to the selected attribute value, to said standardized forecast residual value.

5. A residual value forecasting system according to claim 3 or claim 4, wherein said object is a car, and said attribute is at least either a body color or a grade.

6. A residual value forecasting system according to claim 3, wherein said standardized forecast residual value calculation means comprises:

a function for standardization to calculate standardized residual value in which the residual value of the object included in the transaction data stored in said secondhand price database is adjusted, based on attribute values of a prescribed attribute of the object included in said transaction data;

a function in which a function to calculate the expected value of the standardized forecast residual value (hereinafter referred to as the “

forecast function”

) is sought based on said standardized residual value and the elapsed time included in said transaction data; and

a function in which an expected value of the standardized forecast residual value in a prescribed future is calculated, based on said forecast function.

7. A residual value forecasting system according to claim 4, wherein said standardized forecast residual value calculation means comprises:

a function for standardization to calculate standardized residual value in which the residual value of the object is adjusted based on attribute values of a prescribed attribute of the object;

a function to calculate the expected value of the standardized forecast residual value (hereinafter referred to as the “

forecast function”

) is sought based on said standardized residual value and the elapsed time of said object;

a function in which variance around said forecast function is sought based on said standardized residual value;

a function in which an expected value of the standardized forecast residual value in a prescribed future is calculated, based on said forecast function; and

a function in which a sample value is generated based on a distribution as follows;

(mean value, variance)=(expected value calculated as aforementioned, variance sought as aforementioned) as a standardized forecast residual value.

8. A residual value forecasting method, comprising a forecast residual value calculation step in which a forecast residual value in a prescribed future is calculated, based on the corresponding relationship between the residual value and elapsed time, by referring to a secondhand price database that stores transaction data including a secondhand circulating price or a ratio of the secondhand circulating price to the original price (the secondhand circulating price and the ratio thereof to the original price are collectively referred to as “

residual value”

) of an object, attribute of the object, and elapsed time of the object.

9. A residual value forecasting method according to claim 8, wherein said object is a car.

10. A residual value forecasting method according to claim 8, wherein said forecast residual value calculation step comprises:

standardized forecast residual value calculation step in which the residual value of the object included in the transaction data stored in said secondhand price database is adjusted, based on attribute values of a prescribed attribute of the object included in said transaction data, as if the attribute of said object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

), and in which a standardized forecast residual value in a prescribed future is calculated based on such standardized forecast residual value and the elapsed time included in said transaction data; and

non-standardized forecast residual value calculation step in which a non-standardized forecast residual value is sought by making adjustment corresponding to the attribute of the object of which residual value will be forecasted, to said standardized forecast residual value.

11. A residual value forecasting method according to claim 8, wherein said forecast residual value calculation step comprises:

standardized forecast residual value calculation step in which a forecast residual value in a prescribed future is calculated as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

) (such residual value hereinafter to be referred to as “

standardized forecast residual value”

); and

non-standardized forecast residual value calculation step in which an attribute value of the attribute of the object is randomly selected, based on the component ratio of the attributes, and in which a non-standardized forecast residual value is sought by making adjustment corresponding to said selected attribute value, to said standardized forecast residual value.

12. A residual value forecasting method according to claim 10 or claim 11, wherein said object is a car, and said attribute is at least either a body color or grade.

13. A residual value forecasting method according to claim 10, wherein said standardized forecast residual value calculation step comprises:

a step for standardization to calculate standardized residual value in which the residual value of the object included in the transaction data stored in said secondhand price database is adjusted, based on attribute values of a prescribed attribute of the object included in said transaction data;

a step in which a function to calculate the expected value of the standardized forecast residual value (hereinafter referred to as the “

forecast function”

) is sought based on said standardized residual value and the elapsed time included in said transaction data; and

a step in which an expected value of the standardized forecast residual value in a prescribed future is calculated, based on said forecast function.

14. A residual value forecasting method according to claim 11, wherein said standardized forecast residual value calculation step comprises:

a step for standardization to calculate standardized residual value in which the residual value of the object is adjusted based on attribute values of a prescribed attribute of the object;

a step in which a function to calculate the expected value of the standardized forecast residual value (hereinafter referred to as the “

forecast function”

) is sought based on said standardized residual value and the elapsed time of said object;

a step in which variance around said forecast function is sought based on said standardized residual value;

a step in which an expected value of the standardized forecast residual value in a prescribed future is calculated, based on said forecast function; and

a step in which a sample value is generated based on a distribution as follows;

(mean value, variance)=(expected value calculated as aforementioned, variance sought as aforementioned) as a standardized forecast residual value.

15. A computer program product comprising:

a computer usable medium having computer readable program code means embodied in said medium for a residual value forecasting method according to claim 8.

18. An insurance premium calculation system according to claim 17, further comprising:

a function in which the standardized residual value of the object, for each data, is calculated based on the residual value database of the object;

a function in which a function to calculate the expected value of the standardized forecast residual value is sought based on said standardized residual value and the elapsed time of said object and is stored in said function storing means; and

a function in which variance around said function is sought based on said standardized residual value and is stored in said variance storage means.

16. An insurance premium calculation system for calculating the premium for residual value insurance, comprising:

standardized forecast residual value calculation means which calculates a residual value in a prescribed future as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

) (such residual value hereinafter to be referred to as “

standardized forecast residual value”

); and

non-standardized forecast residual value calculation means which randomly selects an attribute value of the attribute of the object based on the component ratio of the attributes, and seeks a non-standardized forecast residual value (hereinafter referred to as “

non-standardized forecast residual value) by making adjustment corresponding to the selected attribute value, to said standardized forecast residual value; and

insurance premium calculation means, which calculates an insurance premium based on a plurality of said non-standardized forecast residual values.

19. An insurance premium calculation system according to any one of claims 16 to 18, wherein said object is a car, and said attribute is at least either a body color or grade.

17. An insurance premium calculation system, comprising:

function storage means which stores a function to calculate an expected value of a forecast residual value as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

) (such residual value hereinafter to be referred to as “

standardized forecast residual value”

);

variance storage means which stores variance around said function;

component ratio storage means which stores the component ratio of the attributes regarding the object; and

non-standardized factor storage means which stores non-standardized factor for each attribute value regarding the attributes of the object;

wherein said insurance premium calculation system further comprises;

a function in which the expected value of the standardized forecast residual value in a prescribed future is calculated based on the function stored in said function storage means, in which the variance from said variance storage means is read out, and in which the standardized forecast residual value is calculated in accordance with a distribution of (mean value, variance)=(expected value calculated as aforementioned, variance read out as aforementioned);

a function in which the component ratio is read out from said component ratio storage means, in which attribute values are randomly selected setting the probability of the selected respective attribute values as the probability corresponding to said component ratio read out, in which non-standardized factor corresponding to said selected attribute values is read out from said non-standardized factor storage means, and in which non-standardized forecast residual value is calculated by making adjustments, to said standardized forecast residual value, based on said non-standardized factor read out; and

a function in which difference values are calculated by subtracting said non-standardized forecast residual values from respective prescribed reference residual values, in which the mean value of said difference values are calculated as the average amount of compensation, and in which the insurance premium is calculated based on said average amount of compensation.

20. An insurance premium calculation method for residual value insurance, comprising:

standardized forecast residual value calculation step in which a residual value in a prescribed future as if the attribute of the object is deemed to be a representative attribute value (to deem to be a representative attribute value will hereinafter be referred to as “

to standardize”

) (such residual value hereinafter to be referred to as “

standardized forecast residual value”

);

non-standardized forecast residual value calculation step in which an attribute value of the attribute of the object is randomly selected, based on the component ratio of the attributes, and in which a non-standardized forecast residual value (hereinafter referred to as “

non-standardized forecast residual value) is sought by making adjustment corresponding to the selected attribute value, to said standardized forecast residual value; and

insurance premium calculation step in which insurance premium is calculated based on a plurality of said non-standardized forecast residual values.

21. An insurance premium calculation method according to claim 20, wherein said standardized forecast residual value calculation step comprises:

a step in which a function to calculate the expected value of the standardized forecast residual value of the object and variance around said function are read out;

a step in which the expected value of the standardized forecast residual value in a prescribed future is calculated based on said function read out; and

a step in which a sample value is generated based on a distribution of (mean value, variance)=(expected value calculated as aforementioned, variance read out as aforementioned) as a standardized forecast residual value.

22. An insurance premium calculation method according to claim 20, wherein said standardized forecast residual value calculation step comprises:

a step in which a function to calculate the expected value of the standardized forecast residual value of the object and variance around said function are read out;

a step in which the expected value of the standardized forecast residual value in a prescribed future is calculated based on said function read out;

a step in which auto-correlation factor is read out;

a step in which random numbers are generated based on a distribution of (mean value, variance=0,

     1);

a step in which random number used to calculate the standardized forecast residual value are calculated based on said auto-correlation factor read out, by referring said generated random numbers and random numbers used to calculate standardized forecast residual value utilized heretofore; and

a step in which a standardized forecast residual value is calculated based on the said calculated expected value of standardized forecast residual value, said variance read out, and said calculated random numbers used to calculate standardized forecast residual value.

23. An insurance premium calculation method according to claim 21 or claims 22, wherein said standardized forecast residual value calculation step further comprises:

a standardization step in which the standardized residual value of the object for each data is calculated based on the residual value database for the object;

a step in which a function to calculate the expected value of the standardized forecast residual value is sought based on said standardized residual value and the elapsed time of said object; and

a step in which variance around said function is sought based on said standardized residual value.

24. An insurance premium calculation method according to claim 20, wherein said non-standardized forecast residual value calculation step comprises:

a step in which the component ratio of the attributes is read out regarding a prescribed attribute of the object;

a step in which attribute value are randomly selected setting the probability of the selected respective attribute values as the probability corresponding to said component ratio read out;

a step in which non-standardized factor corresponding to said selected attribute value is read out; and

a step in which non-standardized forecast residual value is calculated by making adjustment, to said standardized forecast residual value, based on said non-standardized factor read out.

25. An insurance premium calculation method according to claim 20, wherein said insurance premium calculation step comprises:

a step in which a plurality of said non-standardized forecast residual values is read out;

a step in which difference value is calculated by subtracting each of said non-standardized forecast residual values read out from respective prescribed reference value;

a step in which the mean value of said difference values is calculated as the average amount of compensation; and

a step in which the insurance premium is calculated based on said average amount of compensation.

26. An insurance premium calculation method according to claim 25, wherein said insurance premium calculation step seeks an average amount of compensation for each set, considering a prescribed number of samples for the non-standardized forecast residual values as one set, and calculates the insurance premium based on the mean value of the average amount of compensation per set.

27. An insurance premium calculation method according to claim 20, wherein said object is a car, and said attribute is at least either a body color or grade

28. A computer program product comprising:

a computer usable medium having computer readable program code means embodied in said medium for a insurance premium calculation method according to claim 20.