JP4377443B1 - Credit card payment approval system, credit card used in credit card payment approval system, terminal device and host computer system, and credit card payment approval method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a credit card settlement approval system capable of performing efficient processing while maintaining accuracy in detection of unauthorized use of a credit card equipped with an IC chip.
An IC chip mounted on an IC credit card stores an fraud determination model in which a fraud pattern is modeled, and is used for settlement approval to improve the reliability of offline approval. If it is determined that there is a possibility of unauthorized use in offline approval, the host computer is requested to perform online approval based on a more detailed model, but the update information of the parameters used for the determination is sent together with the approval request, and the host computer When it is confirmed that the latest parameters created by analyzing the transaction trend of the customer are updated, the latest parameters are sent to the IC terminal together with the approval result from the host computer, and the IC chip The parameters set in the fraud determination model are updated to the latest state.
[Selection] Figure 3

Description

The present invention relates to a credit card settlement approval system for efficiently detecting unauthorized use of a credit card equipped with an IC chip, a credit card used in the credit card settlement approval system , a terminal device and a host computer system, and The present invention relates to a credit card settlement approval method.

  When using a credit card, for example, in order to prevent another person who picked up the card from impersonating the person and conducting an illegal transaction, the store accepts the use of the card from the credit card company host computer, Along with checking the credit balance, inquiries about the possibility of unauthorized use are being made. In an unauthorized use inquiry, a scoring model that calculates the possibility of unauthorized use as a score is generally used (see, for example, Patent Document 1).

  On the other hand, if online approval is made by communication with the host computer in order to apply the scoring model every time a credit card is used, the system burden on the credit card company increases and the waiting time for communication with the customer There is a risk of service degradation. In recent years, IC credit cards equipped with IC chips have become widespread in order to deal with such problems, so by setting conditions for some credit judgment on the IC chips, offline approval under certain conditions An invention that enables this is disclosed (for example, see Patent Document 2).

  In addition, an invention is also disclosed that limits the number of online communications by rejecting the use when the conditions written on the IC chip are not met, assuming online approval (for example, patents). Reference 3). As such a condition, it is common that a floor limit that requires online approval is provided for a payment of a certain amount or more.

JP 2007-207011 A JP 2006-313440 A JP 2005-025306 A

  As mentioned above, considering the system burden of the credit card company and the waiting time of customers, it is desirable to reduce the number of online approvals as much as possible. On the other hand, in order to prevent unauthorized use as much as possible, the reliability of detecting unauthorized use is increased. A mechanism to maintain it is also required.

  The invention disclosed in Patent Document 2 approves only a small amount of payment within the range of a certain credit line offline, but if it is within the range of the credit line, unauthorized use may be overlooked. In addition to being peculiar, it is difficult to apply to transactions with large amounts. The invention disclosed in Patent Document 3 determines the possibility of unauthorized use based on the area where the credit card is used or the location where the terminal is installed, and is an element used for determination as compared with a generally used scoring model Are limited, the possibility of detecting unauthorized use offline is limited.

  In other words, an efficient method that can reduce the number of online approvals while maintaining the reliability of detecting fraudulent use that complies with the scoring model even offline can be realized by utilizing the functions of IC credit cards. That is where it is required to do.

The present invention has been made in order to cope with such a problem, and can detect a fraudulent use of a credit card equipped with an IC chip and perform an efficient process while maintaining its reliability. It is an object of the present invention to provide a credit card, a terminal device and a host computer system, and a credit card payment approval method used in the payment approval system .

A first invention for solving the problems of the present application is a host connected via a network to a credit card mounted with an IC chip and a terminal device having a function of reading and writing information stored in the IC chip. A credit card settlement approval system comprising a computer system, wherein the credit card stores a first fraud determination model in which a first fraud parameter in which a pattern of fraudulent use of the credit card is modeled is set. A fraud determination model that reads out a first fraud determination model in which a first fraud parameter is set from the fraud determination model storage means; a determination request storage means that accepts a payment request by the credit card from the terminal device; Read means and the value specified in the settlement request are converted to the first fraudulent parameter. If the specified value matches a predetermined condition, the possibility of fraudulent use of the settlement request corresponds to a criterion that requires online approval. If it is determined that the fraud determination model application unit and the fraud determination model application unit meet the criteria that require online approval, the terminal device reads the payment information generated from the payment request and the IC chip. Further, settlement information transmitting means for transmitting update information specifying the update time of the first fraud parameter to the host computer system, and a first fraud determination model set in the first fraud determination model stored in the fraud determination model storage means. An unauthorized parameter updating means for updating one unauthorized parameter to a new unauthorized parameter, and the host computer The system applies the value included in the payment information received from the terminal device to a second fraud determination model in which a second fraud parameter in which a credit card fraud pattern is modeled is set, and the payment information A payment approval determination means for determining that the possibility of fraudulent use of the payment request corresponds to a criterion that requires online approval, and a determination by the payment approval determination means Received from the terminal apparatus, a determination result transmitting means for transmitting the approval or disapproval determination result to the terminal apparatus, a fraud parameter storage means for storing a third fraud parameter set in the first fraud determination model, and Update that determines whether the updated information matches the update information that specifies the update time of the third incorrect parameter stored in the incorrect parameter storage means An information determining unit; and an unauthorized parameter transmitting unit that reads the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device when the updated information determining unit determines that the update information does not match. Then, the fraud parameter updating unit of the credit card has received the first fraud parameter set in the first fraud determination model stored in the fraud determination model storage unit from the host computer system by the terminal device. It is a settlement approval system characterized by updating to a third fraud parameter.

  In the first invention, a fraud determination model in which fraudulent use patterns are modeled on an IC chip mounted on a credit card is stored, and the fraud determination model in which fraudulent parameters are set is stored in the IC chip. By performing arithmetic processing, it is possible to reduce the number of transactions requiring online approval while improving the reliability of detection of unauthorized use offline. By simplifying the first fraud determination model stored in the IC chip with respect to the second fraud determination model used for determination in the host-side computer system, the terminal-side approval process is simplified and further refined. When analysis is required, it becomes possible to perform a more accurate approval process in the computer system on the host side.

  In addition, when configured as in the first invention, the fraud parameters set in the first fraud determination model used in offline approval are updated to fraud parameters reflecting the latest transaction tendency at the time of online approval. Therefore, it becomes possible to further improve the reliability of detecting unauthorized use offline.

  In the first invention, the first fraud determination model used in the fraud determination model applying means is a model for determining the necessity of online approval by a decision tree, wherein the first fraud parameter is a decision tree. It can also be characterized in that it is set as a reference value for branching at a node.

  In the first invention, the first fraud determination model used in the fraud determination model applying means is a model for determining whether or not online approval is required by a neural network, and the first fraud parameter is a neural network It can also be characterized in that it is set as a coefficient in scoring at an intermediate node.

  Further, according to a first aspect of the present invention, the credit card is stored in a risk parameter storage unit that stores a risk parameter set as a quantitative threshold value that requires online approval for the credit card, and the risk parameter storage unit. A risk parameter reading means for reading the risk parameter, and when the value specified in the settlement request is compared with the risk parameter and the value exceeds a threshold set in the risk parameter, the settlement request is online From the risk parameter application means that determines that the criteria that require approval and the risk parameter application means that meets the criteria that require online approval, the payment information generated from the payment request, and the IC chip When updating the read risk parameters Second settlement information transmitting means for transmitting update information to be specified to the host computer system; and risk parameter updating means for updating the risk parameter stored in the risk parameter storage means to a new risk parameter. The host computer system includes a risk parameter storage unit that stores a risk parameter set as a quantitative threshold that requires online approval for a credit card, and update information received from the terminal device is the risk parameter storage unit. Update information determination means for determining whether or not the update information specifying the update time of the latest risk parameter stored in the update information matches, and when the update information determination means determines that the update information does not match, the latest risk parameter is From the risk parameter storage means A risk parameter transmitting means for transmitting to the terminal device, and the risk parameter updating means for the credit card receives the risk parameter stored in the risk parameter storage means from the host computer system. The terminal device may be updated to the latest risk parameter received.

  When configured in this way, mechanical judgments based on simple criteria such as the amount of usage (online approval required for tens of thousands of yen or more), the number of usage (once online approval required every time, etc.), etc. are offline. Therefore, it is possible to select transactions requiring online approval by multifaceted judgment without increasing the load of offline processing. Moreover, since the risk parameter used in offline approval can be updated to a risk parameter based on a new standard at the time of online approval, it becomes possible to handle the processing standard based on the risk parameter in a unified manner.

  In the first invention, the processing for off-line approval is configured to be executed in the IC chip mounted on the credit card. On the other hand, in the credit card settlement approval system according to the present invention, the second invention It is also possible to configure the terminal device to execute such processing.

  A second invention for solving the problems according to the present application is a credit comprising a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network. A payment approval system for a card, wherein the terminal device includes a payment request receiving means for receiving a payment request by credit card, and a first fraud parameter modeling a fraudulent use pattern of a credit card stored in the IC chip. The fraud determination model reading means for reading the first fraud determination model in which is set and the value specified in the settlement request are applied to the first fraud determination model in which the first fraud parameter is set, and the designation If the calculated value meets the predetermined condition, the possibility of unauthorized use of the settlement request is confirmed online. The fraud determination model applying means for determining that it corresponds to a criterion that requires the payment, and the payment information generated from the payment request when the fraud determination model applying means is determined to be a criterion that requires online approval, and the IC chip Payment information transmitting means for transmitting update information for specifying when to update the first fraud parameter read out from the host computer system, and a first fraud determination model stored in the IC chip. Fraud parameter updating means for updating the fraud parameter to a new fraud parameter, wherein the host computer system uses a credit card fraud pattern as a value included in the payment information received from the terminal device. Fit to the second fraud determination model with the modeled second fraud parameter set A payment approval determination means for determining that the possibility of unauthorized use of the payment request corresponds to a criterion for approval of payment when a value included in the payment information matches a predetermined condition; and a payment approval determination means A determination result transmission unit that transmits the determination result of approval or disapproval determined by the terminal device, an unauthorized parameter storage unit that stores a third unauthorized parameter set in the first unauthorized determination model, and the terminal device Update information determination means for determining whether the update information received from the update information for determining the update time of the third illegal parameter stored in the illegal parameter storage means matches the update information determination means If it is determined that they do not coincide with each other, an unauthorized parameter transmitting unit that reads out the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device; The fraud parameter updating unit of the terminal device includes a third fraud parameter transmitted from the host computer system to the first fraud parameter set in the first fraud determination model stored in the IC chip. It is a settlement approval system characterized by updating to the unauthorized parameter.

  According to a second aspect of the present invention, the first fraud determination model used in the fraud determination model applying means is a model for determining the necessity of online approval using a decision tree, and the first fraud parameter is a decision It can also be characterized in that it is set as a reference value for branching at the nodes of the tree.

  According to a second aspect of the present invention, the first fraud determination model used in the fraud determination model applying means is a model for determining the necessity of online approval by a neural network, wherein the first fraud parameter is a neural network. It can also be characterized by being set as a coefficient in scoring at an intermediate node of the network.

  Further, according to a second aspect of the invention, the terminal device reads a risk parameter reading unit that reads a risk parameter set as a quantitative threshold that requires online approval for the credit card stored in the IC chip, and the settlement request Risk parameter application that determines that the settlement request falls under the criteria that require online approval when the value specified in 1 is compared with the risk parameter and the value exceeds a threshold set in the risk parameter And when the risk parameter applying means determines that it meets the criteria requiring online approval, the payment information generated from the payment request and the update information for specifying the update time of the risk parameter read from the IC chip are , Second settlement information transmission to be transmitted to the host computer system And a risk parameter updating means for updating the risk parameter stored in the IC chip to a new risk parameter, wherein the host computer system requires a quantitative threshold that requires online approval for the credit card. Risk parameter storage means for storing the risk parameter set as: and whether the update information received from the terminal device matches the update information for specifying the update time of the latest risk parameter stored in the risk parameter storage means An update information determination unit for determining, a risk parameter transmission unit for reading the latest risk parameter from the risk parameter storage unit and transmitting to the terminal device when the update information determination unit determines that the update information does not match, The terminal device squirrel Parameter updating means, the risk parameters stored in the IC chip may be characterized by updating to the latest risk parameters transmitted from the host computer system.

  The first invention and the second invention can also be specified as a credit card settlement approval method executed by each configuration of the credit card settlement approval system according to each invention.

  A credit card settlement approval method according to a first aspect of the present invention is connected via a network to a credit card equipped with an IC chip and a terminal device having a function of reading and writing information stored in the IC chip. A credit card settlement approval method executed by a host computer system, wherein the credit card accepts a settlement request by the credit card from the terminal device, and the credit card is the credit card The fraud determination model reading step of reading out the first fraud determination model from the fraud determination model storage means for storing the first fraud determination model in which the first fraud parameter in which the fraud pattern is modeled is set; and the credit card The value specified in the settlement request Applying to the first fraud determination model in which one fraud parameter is set and the specified value meets a predetermined condition, the possibility of fraudulent use of the settlement request requires online approval The fraud determination model applying step for determining that the credit card meets the criteria, and the payment generated from the payment request in the terminal device when the credit card is determined to meet the criteria for requiring online approval in the fraud determination model applying step. A settlement information transmission step for causing the host computer system to transmit information and update information for identifying the update time of the first illegal parameter read from the IC chip; and the settlement received by the host computer system from the terminal device The value included in the information is modeled on the pattern of unauthorized use of credit cards. Applying to the second fraud determination model in which the fraud parameter of 2 is set and the value included in the settlement information matches a predetermined condition, the possibility of unauthorized use of the settlement request approves settlement. A settlement approval determination step for determining that the criterion is satisfied, a determination result transmission step for the host computer system to transmit the determination result of approval or disapproval determined in the settlement approval determination step to the terminal device, and the host computer system The update information received from the terminal device specifies an update time of the third fraud parameter stored in the fraud parameter storage means for storing the third fraud parameter set in the first fraud determination model An update information determination step for determining whether or not the information matches, and the host computer system determines the update information determination If it is determined in the step that the update information does not match, the unauthorized parameter transmission step of reading out the third unauthorized parameter from the unauthorized parameter storage unit and transmitting it to the terminal device; and the credit card in the unauthorized parameter storage unit A fraud parameter updating step of updating the first fraud parameter set in the stored first fraud determination model with the third fraud parameter received by the terminal device from the host computer system. This is a settlement approval method.

  In the credit card settlement approval method according to the first invention, the first fraud determination model used in the fraud determination model applying step is a model for determining whether online approval is necessary or not by a decision tree, The first fraud parameter may be set as a reference value for branching at a node of the decision tree.

  In the credit card settlement approval method according to the first invention, the first fraud determination model used in the fraud determination model applying step is a model for determining whether online approval is necessary by a neural network, The first incorrect parameter may be set as a coefficient in scoring at an intermediate node of the neural network.

  Furthermore, the credit card settlement approval method according to the first aspect of the present invention relates to a risk parameter storage means for storing a risk parameter that is set as a quantitative threshold for the credit card that requires online approval for the credit card. A risk parameter reading step of reading a parameter; and when the credit card compares the value specified in the payment request with the risk parameter and the value exceeds a threshold set in the risk parameter, the payment A risk parameter application step that determines that the request corresponds to a criterion that requires online approval; and a determination that the credit card corresponds to a criterion that requires online approval in the risk parameter application step; Raw from payment request A second settlement information transmission step for causing the host computer system to transmit the settlement information and update information for specifying the update time of the risk parameter read from the IC chip, and the host computer system receives from the terminal device Whether the updated information matches the update information that identifies the update time of the latest risk parameter stored in the risk parameter storage means that stores the risk parameter set as a quantitative threshold that requires online approval for the credit card And when the host computer system determines that the update information does not match in the update information determination step, the latest risk parameter is read from the risk parameter storage means and transmitted to the terminal device Risk parameters And a risk parameter updating step for updating the risk parameter stored in the risk parameter storage means to the latest risk parameter transmitted from the host computer system. It is good.

  A credit card settlement approval method according to a second invention is executed by a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network. And a payment request acceptance step in which the terminal device accepts a payment request by credit card, and the terminal device uses a pattern of fraudulent use of the credit card stored in the IC chip. A fraud determination model reading step of reading the first fraud determination model in which the modeled first fraud parameter is set; and the terminal device sets the value specified in the settlement request as the first fraud parameter. Applying to the first fraud determination model, the specified value meets a predetermined condition The fraud determination model applying step for determining that the possibility of fraudulent use of the payment request corresponds to a criterion requiring online approval, and the terminal device requires online approval in the fraud determination model applying step. If it is determined that it meets the criteria, the settlement information transmission that transmits the settlement information generated from the settlement request and the update information that specifies the update time of the first illegal parameter read from the IC chip to the host computer system A second fraud determination model in which a second fraud parameter in which a value included in the payment information received from the terminal device is modeled by a credit card fraud pattern is set. If the value included in the payment information matches a predetermined condition, A payment approval determination step for determining that the possibility of unauthorized use of the payment request corresponds to a criterion for approval of payment; and a determination result of approval or disapproval determined by the host computer system in the payment approval determination step. A determination result transmission step transmitted to the host computer system, and the update information received by the host computer system from the terminal device is stored in a fraud parameter storage means for storing a third fraud parameter set in the first fraud determination model. An update information determination step for determining whether or not the update information specifying the update time of the third unauthorized parameter that has been updated matches the update information determination step; The terminal device is read out from the unauthorized parameter storage means with a third unauthorized parameter A fraudulent parameter transmission step of transmitting to the host computer system the first fraud parameter set in the first fraud determination model stored in the IC chip; And a fraud parameter updating step for updating to a parameter.

  Further, in the credit card settlement approval method according to the second invention, the first fraud determination model used in the fraud determination model applying step is a model for determining the necessity of online approval by a decision tree, The first fraud parameter may be set as a reference value for branching at a node of the decision tree.

  In the credit card settlement approval method according to the second invention, the first fraud determination model used in the fraud determination model applying step is a model for determining whether online approval is necessary or not by a neural network, The first incorrect parameter may be set as a coefficient in scoring at an intermediate node of the neural network.

  Furthermore, in the credit card settlement approval method according to the second invention, the risk that the terminal device reads a risk parameter set as a quantitative threshold that requires online approval for the credit card stored in the IC chip. When the parameter reading step and the terminal device compare the value specified in the settlement request with the risk parameter and the value exceeds a threshold set in the risk parameter, the settlement request is approved online. A risk parameter application step that determines that the request falls under a criterion, and the terminal device determines that it falls under a criterion that requires online approval in the risk parameter application step, payment information generated from the payment request, Update of the risk parameter read from the IC chip A second settlement information transmission step for transmitting update information for identifying the credit card to the host computer system, and quantitative information that the host computer system requires online approval for the credit card for the update information received from the terminal device. An update information determination step for determining whether or not the update information for identifying the update time of the latest risk parameter stored in the risk parameter storage means for storing the risk parameter set as a threshold value is matched, and the host computer system includes: If it is determined in the update information determination step that the update information does not match, the risk parameter transmission step of reading the latest risk parameter from the risk parameter storage means and transmitting it to the terminal device; and the terminal device in the IC chip Remembered risk Meter, and risk parameter updating step of updating the latest risk parameters transmitted from the host computer system may be characterized by having a.

  According to the present invention, taking advantage of the function of an IC credit card, the fraud determination model in which fraud parameters stored in the IC chip are set improves the reliability of offline approval and reduces the number of transactions requiring online approval. Therefore, it is possible to contribute to the improvement of customer service by reducing the burden of system processing in a host computer system such as a credit card company and shortening the waiting time until approval.

  In addition, fraud parameters set in the first fraud determination model used for offline approval can be updated to fraud parameters reflecting the latest transaction trends at the time of online approval. It becomes possible to further increase the reliability of.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings. The following description is an example of an embodiment of the present invention, and the present invention is not limited to such an embodiment. For example, offline approval is performed using a part of parameters and history information exemplified below. For example, some parameters may be common to card members, or calculation processing related to offline approval may be performed by a terminal device installed in a credit card member store or the like.

  1 and 2 are first and second block diagrams, respectively, showing the configuration of a conventional credit card settlement approval system. 3 and 4 are block diagrams showing configurations of the credit card settlement approval system according to the present invention in the first and second embodiments, respectively. FIG. 5 and FIG. 6 are diagrams showing first and second examples of history information used in the credit card settlement approval system according to the present invention, respectively. FIG. 7 is a diagram showing an example of authorization data used in the credit card settlement approval system according to the present invention. FIG. 8 is a flowchart showing a flow of approval processing in the IC card in the first embodiment of the credit card settlement approval method according to the present invention. FIG. 9 is a flowchart showing a flow of transmission processing on the terminal side in the first embodiment of the credit card settlement approval method according to the present invention. FIG. 10 is a flowchart showing the processing flow of the computer system on the host side in the credit card settlement approval method according to the present invention. FIG. 11 is a flowchart showing a flow of update processing of parameters and the like on the terminal side of the credit card settlement approval method according to the present invention. FIG. 12 is a flowchart showing the flow of terminal-side approval processing in the second embodiment of the credit card settlement approval method according to the present invention. FIGS. 13 and 14 are first and second diagrams, respectively, showing that the approval process can be made more efficient while maintaining the reliability by the credit card settlement approval system according to the present invention. FIG. 15 is a diagram showing an example of a fraud determination model based on a decision tree used in the credit card settlement approval system according to the present invention. FIGS. 16 and 17 are first and second diagrams, respectively, showing examples of fraud determination models using neural networks used in the credit card settlement approval system according to the present invention.

  1 and 2 show a general configuration of a payment approval system using a current IC credit card. When payment is settled using the IC credit card 10, the information stored in the IC credit card 10 is read by the IC terminal 20 installed in the member store of the credit card, and the IC terminal 20 and the IC credit card 10 are read. Whether or not settlement by the IC credit card 10 is possible is determined by offline approval using the credit card or online approval by the host computer 30 such as a credit card company.

  In FIG. 1, storage areas corresponding to the card information storage unit 11 and the risk parameter storage unit 12 are allocated to the IC chip mounted on the IC credit card 10, and a CPU for performing arithmetic processing is also provided. An authentication processing unit 15 and an approval processing unit 16 are provided, both of which are functionally specified. In order to realize each function, an IC chip of the IC credit card 10 is provided. The arithmetic processing is executed by the CPU provided in the IC chip of the IC credit card 10 based on the application program stored in the above.

  2, functions corresponding to the authentication processing unit 15 and the approval processing unit 16 of the IC credit card 10 may be provided in the IC terminal 20. In this case, the IC terminal 20 is a computer that can read information from the IC chip of the IC credit card 10 and perform a predetermined calculation process, and includes a settlement request reception unit 21, an authentication processing unit 22, and an approval processing unit 23. Is provided. The settlement request accepting unit 21, the authentication processing unit 22, and the approval processing unit 23 are all functionally specified. In order to realize each function, each of the memory areas provided in the IC terminal 20 is provided. An application program corresponding to the above function is read out, and arithmetic processing is executed by the CPU of the IC terminal 20.

  The card information storage unit 11 stores a member number for identifying the IC credit card 10, a personal identification number registered by the user of the IC credit card 10, a key for determining a forged card, and the like. The authentication processing unit 15 reads out the information from the card information storage unit 11 to determine the forged card using the key, or whether the password entered in the IC terminal 20 matches the password. The authentication process of the IC credit card 10 is performed by performing personal authentication from the beginning.

  The authentication process may be performed by the authentication processing unit 22 of the IC terminal 20 by reading information necessary for the authentication process from the card information storage unit 11 to the IC terminal 20. If an error occurs in the authentication process, payment using the IC credit card 10 is not accepted.

  The risk parameter storage unit 12 sets quantitative criteria such as a floor limit that requires online approval for payments over a certain amount, and a frequency that requires online approval, or generates random numbers. A risk parameter in which a rule requiring online approval is set, such as random online approval, is stored. When the payment request receiving unit 21 of the IC terminal 20 receives a payment request in which the usage amount is specified, the information is transferred to the approval processing unit 16 of the IC credit card 10 and the approval processing unit 16 of the IC credit card 10 Then, the risk parameter is read from the risk parameter storage unit 12, and the usage amount specified in the settlement request is applied to the risk parameter to determine whether the rule that requires online approval is met.

  Such approval processing may be performed in the approval processing unit 23 of the IC terminal 20 by reading the risk parameters from the risk parameter storage unit 12 into the IC terminal 20. Further, the authentication process described above may be performed by the IC terminal 20 and then the approval process may be performed by the IC credit card 10.

  If it is determined that online approval is required, authorization data including a membership number, a member store code, a usage amount and the like necessary for approval of payment is transmitted to the host computer 30. If it is determined that online approval is not required, the payment is approved offline and the payment by the IC credit card 10 is accepted.

  The host computer 30 includes an approval processing unit 31, an authorization data storage unit 32, and a fraud determination model storage unit 33. The approval processing unit 31 is functionally specified, and an approval program is read out by reading an application program for the approval processing into a memory area provided in the host computer 30 and executing arithmetic processing by the CPU. Is done. An authorization data storage unit 32 that stores authorization data related to past transactions for each member, and an fraud determination model storage unit 33 that stores fraud determination models in which fraud parameters that model fraudulent use patterns of credit cards are set. May be allocated to a predetermined storage area of the HDD of the host computer 30 or provided in a database server or the like connected to the host computer 30.

  The host computer 30 is not limited to a single computer, and may be a plurality of computers. Some of the functions are externally connected to a card company computer via a network. The computer may be configured.

  In the host computer 30 that has received the authorization data, the authorization processing unit 31 is activated, and the received authorization data is applied to the fraud determination model stored in the fraud determination model storage unit 33 to approve settlement. To determine whether it meets the criteria for When the unauthorized use is determined by scoring the possibility of fraud, the past transaction history of the member stored in the authorization data storage unit 32 may be used. The approval / disapproval is transmitted to the IC terminal 20 and indicates whether or not the IC credit card 10 can accept payment.

  In this way, according to the current general method, in the offline approval, the fraud determination model in which the fraud parameter that models the fraud pattern is set is not used, but based on quantitative criteria and random numbers. The only decision is whether to approve online. Considering the burden on the communication line and the host computer system, it is desirable to increase the offline approval ratio as much as possible. However, according to such a method, as shown in FIG. In proportion to this, the number of fraudulent transactions in offline transactions increases, resulting in a problem that reliability is lowered. Similarly, in FIG. 14, the number of fraudulent transactions that occur in offline approval increases as the risk parameter is narrowed down to increase the offline approval ratio and reduce the burden on the communication line and host computer system. Will increase.

  In order to improve the reliability in offline approval for such problems, it is desirable to perform determination using an fraud determination model in which fraud parameters are set by statistical analysis even in offline approval. By introducing statistical analysis in offline approval, as shown in FIG. 13, even if the ratio of offline transactions is increased, it is possible to suppress the decrease in reliability associated therewith, as shown in FIG. As described above, even if the narrowing down by the risk parameter is eased, the increase in the number of fraudulent transactions that occur in offline approval can also be stopped moderately.

  Based on the above concept, what is configured to be able to make a determination using fraudulent parameters by statistical analysis even in offline approval is the example of the credit card payment approval system according to the present invention illustrated in FIGS. It is a configuration. In the first embodiment of FIG. 3, the approval process or the like is performed by the IC chip of the IC credit card 10, and in the second embodiment of FIG. 4, the approval process or the like is performed by the IC terminal 20.

  3 and 4, when payment is made using the IC credit card 10, the information stored in the IC credit card 10 is read by the IC terminal 20 installed in a credit card member store or the like. Whether or not settlement by the IC credit card 10 is possible is determined by offline approval using the IC terminal 20 and the IC credit card 10 or online approval by the host computer 30 such as a credit card company.

  In FIG. 3, storage areas corresponding to the card information storage unit 11, the risk parameter storage unit 12, the unauthorized parameter storage unit 13, and the history information storage unit 14 are allocated to the IC chip mounted on the IC credit card 10. A CPU for performing arithmetic processing is also provided. Further, an authentication processing unit 15, an approval processing unit 16, and an update processing unit 17 are provided, all of which are functionally specified, and in order to realize each function, an IC credit is provided. Based on the application program stored in the IC chip of the card 10, the arithmetic processing is executed by the CPU provided in the IC chip of the IC credit card 10.

  The authentication processing unit 15, the approval processing unit 16, and the update processing unit 17 do not perform the necessary arithmetic processing only in the IC chip, but the corresponding authentication processing unit 22, approval processing unit 23, It is good also as what cooperates with the update process part 24. FIG. For example, an authentication processing unit for general-purpose authentication processing, such as verification of a personal identification number, which passes information set for each card company to authenticate whether or not it is a legitimate card from the authentication processing unit 22 to the authentication processing unit 15. A method of executing at 22 is conceivable.

  Furthermore, as shown in FIG. 4, functions corresponding to the authentication processing unit 15, the approval processing unit 16, and the update processing unit 17 of the IC credit card 10 may be provided in the IC terminal 20. In this case, the IC terminal 20 is a computer capable of reading information from the IC chip of the IC credit card 10 and performing a predetermined calculation process. The payment request reception unit 21, the authentication processing unit 22, the approval processing unit 23, An update processing unit 24 is provided. The settlement request receiving unit 21, the authentication processing unit 22, the approval processing unit 23, and the update processing unit 24 are all functionally specified, and are provided in the IC terminal 20 in order to realize each function. The application program corresponding to each function is read into the memory area, and the arithmetic processing is executed by the CPU of the IC terminal 20.

  The card information storage unit 11 stores a member number for identifying the IC credit card 10, a personal identification number registered by the user of the IC credit card 10, a key for determining a forged card, and the like. The authentication processing unit 15 reads out the information from the card information storage unit 11 to determine the forged card using the key, or whether the password entered in the IC terminal 20 matches the password. The authentication process of the IC credit card 10 is performed by performing personal authentication from the beginning. The authentication process of the IC credit card 10 is performed.

  The authentication process may be performed by the authentication processing unit 22 of the IC terminal 20 by reading information necessary for the authentication process from the card information storage unit 11 to the IC terminal 20. If an error occurs in the authentication process, payment using the IC credit card 10 is not accepted.

  The risk parameter storage unit 12 sets quantitative criteria such as a floor limit that requires online approval for payments over a certain amount, and a frequency that requires online approval, or generates random numbers. A risk parameter in which a rule requiring online approval is set, such as random online approval, is stored. When the payment request receiving unit 21 of the IC terminal 20 receives a payment request in which the usage amount is specified, the information is transferred to the approval processing unit 16 of the IC credit card 10 and the approval processing unit 16 of the IC credit card 10 Then, the IC terminal 20 reads out the risk parameters from the risk parameter storage unit 12, applies the usage amount specified in the settlement request to the risk parameters, and determines whether the rule requires online approval. The CPU is executed.

  The approval result in the approval processing unit 16 of the IC credit card 10 is delivered to the approval processing unit 23 of the IC terminal 20. The risk parameter may be a common parameter among members, but a different parameter may be set for each member in order to make a more reliable determination reflecting the transaction trend of each member. preferable.

  Such approval processing may be performed in the approval processing unit 23 of the IC terminal 20 by reading the risk parameters from the risk parameter storage unit 12 into the IC terminal 20. Further, the authentication process described above may be performed by the IC terminal 20 and then the approval process may be performed by the IC credit card 10.

  The fraud determination model storage unit 13 statistically processes the tendency of fraudulent use that has occurred in the past (the use of late-night or short-term continuous use has a high probability of fraudulent transactions, etc.) and the tendency of normal transactions for each member. The fraud determination model in which the fraud parameters set by modeling the fraudulent use pattern of the credit card analyzed by the above is stored. In response to the payment request in which the usage amount received by the payment request reception unit 21 of the IC terminal 20 is specified, the approval processing unit 23 reads the fraud determination model in which the fraud parameter is set from the fraud determination model storage unit 13. Data such as the usage amount specified in the payment request and the reception date and time when the payment request is received are applied to a fraud determination model in which fraud parameters are set, and it is determined whether the rule requires online approval.

  The type of fraud determination model used here is not particularly limited, but by applying data such as the usage amount specified in the payment request and the date and time of receipt of the payment request to the decision tree (decision tree) Apply the decision tree model for determining whether there is a possibility of fraud, the usage amount specified in the payment request, the date and time of acceptance of the payment request, etc. to a predetermined calculation formula, and the calculated score is a predetermined threshold It is possible to apply a scoring model that determines whether there is a possibility of fraud depending on whether or not.

  FIG. 15 shows an example of a fraud determination model using a decision tree. The age of the customer, the product to be purchased, the usage amount, and the like become the nodes of the decision tree, and values that serve as criteria for branching at the nodes ( In the example of FIG. 15, “30”, “50”, “10,000”, etc.) are set as illegal parameters.

  When the scoring model is adopted, in the predetermined calculation formula for calculating the score, the usage amount specified in the settlement request, the reception date / time when the settlement request is accepted, etc. are variables, but are set as coefficients. Numeric value is an invalid parameter.

  As one form of the scoring model format, a model using a neural network as shown in FIGS. 16 and 17 may be employed. In this example, the customer's age and usage amount are input nodes, and the intermediate node assigns the value of the input node to the variable part of a predetermined arithmetic expression that is a variable, and the score exceeds a predetermined threshold value. Whether or not the value to be substituted into the variable portion is derived at the next intermediate node, and finally it is determined whether or not a predetermined threshold value is exceeded and output to the output node.

  For example, as shown in FIG. 17, when an input of age “34” and money amount “3000” is received, the first intermediate node sums values obtained by multiplying these by coefficients of “0.004” and “0.00012”, respectively. Then, a score of “0.496” is calculated. Since this score does not exceed the threshold value of “0.5”, a numerical value of “0” when the threshold value is not exceeded is delivered to the variable of the next intermediate node.

  As described above, in a model using a neural network, it is set as a coefficient in scoring at an intermediate node of the neural network (“0.004”, “0.00012”, etc. in the example of FIG. 17). It is an invalid parameter.

  Note that the calculation processing for approval using the unauthorized parameter as described above performed in the approval processing unit 23 of FIG. 3 is also performed in the approval processing unit 23 of the IC terminal 20 shown in FIG. Also good. In this case, the fraud determination model in which the fraud parameters read from the fraud determination model storage unit 13 of the IC credit card 10 are set is delivered to the IC terminal 20, and the arithmetic processing is performed by the CPU provided in the IC terminal 20.

  Here, as the fraud parameters set in the fraud determination model, parameters common to members may be used. However, in order to make a more reliable determination reflecting the transaction tendency of each member, It is preferable to set different parameters for each.

  The history information storage unit 14 stores history information related to recent credit card usage histories of IC credit card 10 members. In the determination based on the fraud determination model, it is preferable to use the latest usage history in order to determine the possibility of fraudulent transaction from the normal transaction tendency for each member. For example, a member who normally uses a credit card only several times a month repeats the use of the card many times in a short time, or a member who usually only pays about tens of thousands of yen pays a large amount of more than 100,000 yen In order to determine that there is a high possibility of fraudulent transactions, etc., the latest history information is required.

  Therefore, when the fraud determination model in which the fraud parameter is set is read from the fraud determination model storage unit 13 and it is determined whether it corresponds to a rule that requires online approval, the usage amount specified in the settlement request or the settlement request The history information is read from the history information storage unit 14 instead of the reception date / time when the information is received, and these numerical values are compared with the incorrect parameters, or the incorrect parameters are multiplied by the incorrect parameters. It is preferable to be configured to apply. In this case, the unauthorized parameter stored in the unauthorized parameter storage unit 13 includes a model using history information.

  Note that the format of the history information used here is not particularly limited. As shown in the example of FIG. 5, a part of the authorization data at the most recent use is stored for each record. Alternatively, as shown in the example of FIG. 6, the previous usage information, the average value of several previous times, or the like may be stored.

  Incidentally, the fraud determination model that models the pattern of fraudulent use of credit cards is also used for online approval processing. In the present invention, the fraudulent parameters are also used for the off-line approval process of the approval processing unit 16 (in the case of FIG. 3) or the approval processing unit 23 (in the case of FIG. 4). In consideration of the storage capacity of the IC chip to be processed and the arithmetic processing capability of the IC chip or the IC terminal 20, it is difficult to use the same fraud determination model and history information as the host side as they are. Therefore, in the present invention, it is determined that there is a possibility of unauthorized use in offline processing by adopting a fraud determination model used on the host side or a relatively simple model in which history information is simplified for offline processing. In such a case, authorization data is sent to the host computer 30 for online approval.

  In the determination using the fraud determination model in which the fraud parameter is set in the approval processing section 16 (in the case of FIG. 3) or the approval processing section 23 (in the case of FIG. 4), it is determined whether it corresponds to the rule requiring online approval. The method of doing is not particularly limited, and for example, a decision tree model or a scoring model as described above is used.

  Regarding the history information stored in the history information storage unit 14, after a new transaction is performed using the IC credit card 10, it is necessary to update the history information to reflect the transaction. Therefore, when the offline approval is performed by the approval processing unit 16 (in the case of FIG. 3) or the approval processing unit 23 (in the case of FIG. 4), or the notification that the online approval is performed from the approval processing unit 31 of the host computer 30. If accepted, the authorization data required for the update is delivered from the update processing unit 24 of the IC terminal 20 to the update processing unit 17 of the IC credit card 10, and the update processing unit 17 is activated to update the IC credit card 10. The history information stored in the history information storage unit 14 is updated to reflect the transaction acceptance time, usage amount, etc. approved.

  Note that the arithmetic processing for updating the history information as described above performed in the update processing unit 24 of FIG. 3 may also be performed in the update processing unit 24 of the IC terminal 20 shown in FIG. In this case, the history information stored in the history information storage unit 14 of the IC credit card 10 is updated by the processing of the update processing unit 24 of the IC terminal 20.

  In the approval processing unit 16 (in the case of FIG. 3) or the approval processing unit 23 (in the case of FIG. 4), as a result of applying the risk parameter or the fraud determination model in which the fraud parameter is set, a rule requiring online approval If it is determined to be applicable, a payment request such as authorization data including the usage amount received by the payment request receiving unit 21 is transmitted from the IC terminal 20 to the host computer 30. FIG. 7 shows an example of the authorization data to be transmitted, which includes information such as a membership number (card number) that can identify the user of the IC credit card 10, reception time, member store code, usage amount, etc. It is.

  Here, when sending authorization data, it also sends update information that identifies when the fraud parameters and risk parameters set in the fraud determination model read for use in offline approval processing are updated. It is preferable to configure as described above. The format of the update information is not particularly limited. For example, the last update date and time, a code indicating the file version, and the like may be used.

  The host computer 30 includes an approval processing unit 31, an authorization data storage unit 32, a fraud determination model storage unit 33, and a risk parameter storage unit 34. The approval processing unit 31 is functionally specified, and an approval program is read out by reading an application program for the approval processing into a memory area provided in the host computer 30 and executing arithmetic processing by the CPU. Is done. An authorization data storage unit 32 that stores authorization data related to past transactions for each member, and an fraud determination model storage unit 33 that stores fraud determination models in which fraud parameters that model fraudulent use patterns of credit cards are set. May be allocated to a predetermined storage area of the HDD of the host computer 30 or provided in a database server or the like connected to the host computer 30.

  In the host computer 30 that has received the authorization data, the authorization processing unit 31 is activated, and the received authorization data is applied to the fraud determination model in which the fraud parameters stored in the fraud determination model storage unit 33 are set. Determine if the possibility falls within the criteria for approving payment.

  Here, the fraud parameters used in the fraud determination model may be parameters common among members, but in order to make a more reliable determination reflecting the transaction trend of each member, It is preferable to set different parameters. In this case, the fraud determination model storage unit 33 stores, for each member, a fraud parameter to which a member number for identifying the member is attached. Therefore, the fraud parameter used for the determination by using the member number included in the authorization data as a key. Should be specified.

  Compared with the fraud determination model stored in the fraud determination model storage unit 13 of the IC credit card 10 used for offline approval, the fraud determination model stored in the fraud determination model storage unit 33 includes the storage capacity of the computer and As described above, it is possible to use a more detailed model in consideration of the difference in processing capability.

  Furthermore, when scoring the possibility of fraudulent use using fraud parameters, as in the case of offline approval, historical information such as past transaction histories of members stored in the authorization data storage unit 32 should be used. It is good. Compared with the history information stored in the history information storage unit 14 of the IC credit card 10 used for offline approval, more information can be stored in the authorization data storage unit 32, enabling more detailed analysis. This is as described above.

  The approval / disapproval of the payment by the IC credit card 10 determined by the approval processing unit 31 is transmitted to the IC terminal 20. In the approval process in the host computer 30, the payment is not approved even if the usage amount specified in the authorization data exceeds the usage limit amount with reference to the usage limit amount set for each member. It is normal. In the IC terminal 20, whether or not the acceptance of the settlement by the IC credit card 10 is indicated, and the update processing unit 17 (in the case of FIG. 3) or the update processing unit 24 (in the case of FIG. 4) is activated, and the approved transaction is performed. Reflecting this, the history information stored in the history information storage unit 14 of the IC credit card 10 is updated.

  In conjunction with the above approval process, the approval processing unit 31 uses the fraud parameters set in the fraud determination model stored in the fraud determination model storage unit 33 to update the fraud parameters and risk parameter update information transmitted from the IC terminal 20. Or, the risk parameters stored in the risk parameter storage unit 34 are compared with the update information that identifies the update time when the latest information is updated. The parameters stored in each of the fraud determination model storage unit 33 and the risk parameter storage unit 34 are created as models reflecting new trends as transaction information including fraudulent use is accumulated. The new parameters have been updated.

  If any update information does not match, the latest fraud parameter or risk parameter is updated from the fraud determination model storage unit 33 or the risk parameter storage unit 34 in order to update the fraud parameter or risk parameter stored in the IC credit card 10. Is transmitted from the host computer 30 to the IC terminal 20 together with the approval / disapproval of payment by the IC credit card 10. In the IC terminal 20, the update processing unit (in the case of FIG. 3) 17 or the update processing unit 24 (in the case of FIG. 4) is activated, and the risk parameter storage unit 12 or the fraud determination model storage unit 13 of the IC credit card 10 is The latest fraud parameter or risk parameter received from the host computer 30 is stored, and the fraud parameter or risk parameter is updated.

  As described above, the fraud determination model storage unit 33 of the host computer 30 has a more detailed fraud determination model than the fraud determination model stored in the fraud determination model storage unit 13 of the IC credit card 10. Therefore, when the update information is compared, not all the fraud parameters of the fraud determination model stored in the fraud determination model storage unit 33 of the host computer 30 but the fraud determination model storage unit of the IC credit card 10 Some fraud parameters that are also set in the fraud determination model stored in FIG. Alternatively, when there are fraud parameters used only in the fraud determination model stored in the fraud determination model storage unit 13 of the IC credit card 10, the fraud determination model stored in the fraud determination model storage unit 33 of the host computer 30 is used. Even if the fraud parameters are not used, the fraud determination model storage unit 33 is configured to store the latest information.

  In the host computer 30, the parameters stored in the fraud determination model storage unit 33 and the risk parameter storage unit 34 are appropriately updated. Therefore, when the update information does not match, the host computer 30 is normal. The update information on the side should be new. As described above, the update process to the latest parameter is also performed at the time of the online approval process, so that the IC credit card 10 also has the latest update without performing a specific operation for parameter update. Parameters are easily reflected, and it becomes possible to improve the reliability of offline approval.

  A processing flow in the first embodiment of the credit card settlement approval method according to the present invention will be described with reference to the flowcharts of FIGS. 8 shows the flow of approval processing in the IC card that has accepted a payment request via the IC terminal, FIG. 9 shows the flow of transmission processing in the IC terminal, and FIG. 10 shows the processing flow of the computer system on the host side that performs payment approval. FIG. 11 shows the processing flow of the IC terminal that has accepted the approval result.

  When payment is to be made with an IC credit card at a merchant of the credit card, etc., the IC credit card is read by the IC terminal, and the usage amount, which is the price of the product to be purchased, is input to the IC terminal. A settlement request is made. In the IC terminal, the information such as the usage amount received for the input is delivered to the IC chip mounted on the IC credit card, and the information is received by the IC chip (S01).

  In the IC chip mounted on the IC credit card, the card information stored in the IC chip is read, and from the read information, for example, a key for identifying a counterfeit card, a personal identification number for personal authentication, etc. The card information authentication process is executed by the arithmetic circuit of the IC chip (S02). Information used for these authentications may be received from an IC terminal. If it is determined that the card used is not a legitimate card or is not used by the card holder (S03), error processing is performed and the IC credit card settlement is stopped. It is also possible to store the usage limit amount per time in the card information and perform error processing even when the usage amount of the payment request exceeds the usage limit amount.

  When it is determined that the card holder is authorized use (S03), the risk parameter in which a rule requiring online approval is set is read (S04), and the accepted settlement request requires online approval as the risk parameter. Whether it does not exceed the set threshold (S05), for example, the usage amount does not exceed the amount set as the amount that requires online approval, or the number of times of use by offline approval until the previous time exceeds the predetermined number It is confirmed whether it is a thing. When such a threshold value is exceeded, it is determined that online approval is necessary without applying to the fraud determination model.

  If the threshold value is not exceeded (S05), the fraud determination model (1) in which fraud parameters are set modeling the fraudulent use pattern of the credit card (the fraud determination model (2) used in the computer system on the host side) Simplified) and history information that is the latest transaction history of the credit card is read (S06) and applied to the fraud determination model (1) (S07), and the value included in the settlement request is illegal as a variable. If it is applied to the decision model (1), it does not correspond to the pattern of unauthorized use (S08). For example, the usage time, usage amount, usage location specified from the member store code, etc. A scoring model in which target products are also identified as variables by being identified as codes) Fitting, the scoring of the possibility that a fraud is carried out. When the calculated score exceeds a predetermined threshold and it is determined that there is a possibility of unauthorized use, it is determined that online approval is necessary.

  The order of the process for applying the risk parameter described in S04 to S05 and the process for applying the fraud determination model (1) described in S06 to S08 is not particularly limited, and the order is reversed. It may be.

  When it is determined that the possibility of fraud is low, for example, the calculated score does not exceed a predetermined threshold value, the IC terminal is instructed that offline approval may be performed (S09). Approval is displayed. Further, the history information stored in the IC chip mounted on the IC credit card is updated to reflect the approved transaction (S10). Thereafter, in the IC terminal, a process of transmitting authorized data to be stored in the host computer system is performed.

  When it is determined in S05 or S08 that online approval is necessary, the risk parameter used in the approval process or update information regarding the update of the fraud parameter set in the fraud determination model (1) is delivered to the IC terminal, An online approval process is requested to the IC terminal (S11).

  The processing of the IC terminal that has received the information is as shown in the flowchart of FIG. When the update information of the risk parameter or fraud parameter and the request for online approval processing are received (S12), the received settlement request is received in the form of authorization data, etc. together with the update information of the risk parameter or fraud parameter delivered from the IC chip, The data is transmitted to the host computer (S13).

  The processing of the host computer that has received such information is as shown in the flowchart of FIG. The host computer receives the authorization data and the update information of the risk parameter or illegal parameter transmitted from the IC terminal (S14). Update information related to the latest risk parameter or illegal parameter update stored in the computer system on the host side is referred to (S15) and compared with the received update information related to the updated risk parameter or illegal parameter ( S16) When the update time of the host computer system is updated and is newer than the update time of the IC chip, the latest risk parameter or illegal parameter is read (S17) and temporarily stored.

  Subsequently, in order to perform a more detailed analysis process than offline, a fraud determination model (2) in which fraud parameters that model a fraudulent use pattern of a credit card are set (a fraud determination model stored in an IC credit card ( (1) More detailed) is read out (S18) and applied to the fraud determination model (2), and the received authorization data does not correspond to the fraudulent use pattern in the fraud determination model (2) ( S19) The value included in the authorization data is applied as a variable to the fraud determination model (2) to determine the possibility of unauthorized use. Here, reading information about past transaction history from a database or the like for storing authorization data and using it for scoring is the same as using history information offline.

  When it is determined that the possibility of fraud is low, for example, the calculated score does not exceed a predetermined threshold, the settlement is approved and the latest risk parameter or fraud parameter temporarily stored Is transmitted to the IC terminal (S20). In addition, the approval process in the host computer system also refers to the usage limit amount set for each member, and even if the usage amount specified in the authorization data exceeds the usage limit amount, the payment is not approved. It is usual to be taken. If the parameter has not been updated, only the approval result is transmitted.

  When it is determined that the calculated score exceeds a predetermined threshold value and it is determined to correspond to unauthorized use, payment by IC credit card is disapproved, and the latest risk parameter or illegal parameter temporarily stored is the approval result. Is transmitted to the IC terminal (S21). When the parameter is not updated, only the approval result is transmitted as in the case of the approval. The authorization data to be determined is stored in a database for saving the authorization data together with the approval result in both cases of approval and non-approval (S22).

  The processing of the IC terminal that has received the approval result is as shown in the flowchart of FIG. In the IC terminal that has received the settlement approval result and the latest risk parameter or fraud parameter (S23), the settlement approval result is displayed on the IC terminal (S24), and the received latest risk parameter or fraud parameter is displayed as an IC credit. This information is updated by writing on the IC chip mounted on the card (S25).

  A processing flow in the second embodiment of the credit card settlement approval method according to the present invention will be described with reference to the flowchart of FIG. The flowchart of FIG. 12 shows a processing flow in the IC terminal that has accepted the settlement request.

  When payment is to be made with an IC credit card at a merchant of the credit card, etc., the IC credit card is read by the IC terminal, and the usage amount, which is the price of the product to be purchased, is input to the IC terminal. A settlement request is made. The IC terminal reads card information such as a membership number from an IC chip mounted on the IC credit card, and accepts input such as a usage amount (S31).

  The IC terminal performs card information authentication processing using, for example, a key for identifying a counterfeit card, a personal identification number for personal authentication, etc. from the card information read from the IC chip (S32). If it is determined that the card used is not a legitimate card or is not used by the cardholder (S33), error processing is performed and the IC credit card settlement is stopped. It is also possible to store the usage limit amount per time in the card information and perform error processing even when the usage amount of the payment request exceeds the usage limit amount.

  If it is determined that the card holder is properly used (S33), a risk parameter in which a rule requiring online approval is set is read from an IC chip mounted on the IC credit card (S34), and a settlement request is made. Does not exceed the threshold set to require online approval in the risk parameter (S35), for example, whether the usage amount exceeds the amount set as the amount that requires online approval, or offline approval up to the previous time Confirmation is made whether the number of times of use exceeds a predetermined number. If such a threshold value is exceeded, it is determined that online approval is necessary, without applying the fraudulent parameter.

  If the threshold value is not exceeded (S35), the fraud determination model (1) (host side computer) in which fraud parameters are set, which models the fraudulent use pattern of the credit card from the IC chip mounted on the IC credit card. The fraud determination model (2) used in the system is simplified) and history information that is the latest transaction history of the credit card is read (S36), and is applied to the fraud determination model (1) (S37). When the value included in the settlement request is applied to the fraud determination model (1) as a variable, it does not correspond to the fraud pattern (S38), for example, specified from the usage time, the usage amount, and the member store code Variables (such as gender and other attributes, and target products, etc.) Be understood as a variable by applying the scoring model the same) to a first embodiment, the scoring of potentially fraudulent use is made. When the calculated score exceeds a predetermined threshold and it is determined that there is a possibility of unauthorized use, it is determined that online approval is necessary.

  The order of the process for applying the risk parameter described in S34 to S35 and the process for applying the fraud determination model (1) described in S36 to S38 is not particularly limited, and the order is reversed. It may be.

  When it is determined that the possibility of fraud is low, for example, the calculated score does not exceed a predetermined threshold, offline approval is performed, and payment approval is displayed on the IC terminal (S39). Further, the history information stored in the IC chip mounted on the IC credit card is updated to reflect the approved transaction (S40), and the approved authorization data is stored in the host computer system for storage in the host computer system. It transmits to the computer (S41).

  When it is determined in S05 or S08 that online approval is required, the payment request accepted by the IC terminal is set in the risk parameter or fraud determination model (1) used in the approval process in the form of authorization data or the like. The update information related to the update of the invalid parameter is transmitted to the host computer (S12).

  The processing after the host computer receives these pieces of information and the subsequent processing at the IC terminal are the same as those in the first embodiment shown in the flowcharts of FIGS.

It is a 1st block diagram which shows the structure of the settlement approval system of the conventional credit card. It is a 2nd block diagram which shows the structure of the settlement approval system of the conventional credit card. It is a block diagram which shows the structure in 1st Embodiment of the payment approval system of the credit card concerning this invention. It is a block diagram which shows the structure in 2nd Embodiment of the payment approval system of the credit card concerning this invention. It is a figure which shows the 1st example of the historical information used with the payment approval system of the credit card concerning this invention. It is a figure which shows the 2nd example of the historical information used with the payment approval system of the credit card concerning this invention. It is a figure which shows an example of the authorization data used with the payment approval system of the credit card concerning this invention. It is a flowchart which shows the flow of the approval process in IC card in 1st Embodiment of the payment approval method of the credit card concerning this invention. It is a flowchart which shows the flow of the transmission process by the side of the terminal in 1st Embodiment of the payment approval method of the credit card concerning this invention. It is a flowchart which shows the processing flow of the computer system of the host side of the payment approval method of the credit card concerning this invention. It is a flowchart which shows the flow of the parameter etc. update process by the side of the terminal of the payment approval method of the credit card concerning this invention. It is a flowchart which shows the flow of the approval process by the side of the terminal in 2nd Embodiment of the payment approval method of the credit card concerning this invention. It is a 1st figure which shows that the approval process can be made efficient, maintaining reliability by the payment approval system of the credit card concerning this invention. It is a 2nd figure which shows that the approval process can be made efficient, maintaining reliability by the payment approval system of the credit card concerning this invention. It is a figure which shows an example of the fraud determination model by the decision tree used with the payment approval system of the credit card concerning this invention. It is a 1st figure which shows an example of the fraud determination model by the neural network used with the payment approval system of the credit card concerning this invention. It is a 2nd figure which shows an example of the fraud determination model by the neural network used with the payment approval system of the credit card concerning this invention.

Explanation of symbols

10 IC credit card 11 card information storage unit 12 risk parameter storage unit 13 fraud determination model storage unit 14 history information storage unit 15 authentication processing unit 16 approval processing unit 17 update processing unit 20 IC terminal 21 settlement request reception unit 22 authentication processing unit 23 Approval processing unit 24 Update processing unit 30 Host computer 31 Approval processing unit 32 Authorization data storage unit 33 Fraud determination model storage unit 34 Risk parameter storage unit

Claims (10)

A credit card settlement approval system comprising a credit card equipped with an IC chip, and a host computer system connected via a network to a terminal device having a function of reading and writing information stored in the IC chip. ,
The credit card is
A decision tree model that determines whether or not online approval of a credit card is necessary using a decision tree, and stores a decision tree model in which a first fraud parameter is set as a reference value for branching at a decision tree contact point Storage means;
Payment request receiving means for receiving a payment request by the credit card from the terminal device;
A decision tree model reading means before decision tree model storage unit reads the first decision tree model incorrect parameters are set,
When the values specified in the settlement request are applied in the order of the decision tree model to the contact points of the decision tree in which the first fraud parameter is set and the predetermined conditions are met, the fraudulent use of the settlement request is performed. A decision tree model application means that determines that the probability falls under a criterion that requires online approval;
When it is determined that the decision tree model applying means corresponds to a criterion requiring online approval, the terminal device is updated with the payment information generated from the payment request and the first fraud parameter read from the IC chip. Payment information transmitting means for causing the host computer system to transmit update information for identifying
And illegal parameter updating means for updating the first fraud parameters set to the stored decision tree model in the decision tree model storage means to a new invalid parameters,
With
The host computer system is
A value included in the payment information received from the terminal device is applied to an fraud determination model in which a second fraud parameter that models a fraudulent usage pattern of a credit card is set, and the value included in the payment information is predetermined. A payment approval determination means for determining that the possibility of fraudulent use of the payment request corresponds to a criterion that requires online approval,
A determination result transmitting means for transmitting a determination result of approval or disapproval determined by the settlement approval determination means to the terminal device;
Illegal parameter storage means for storing a third illegal parameter set in the decision tree model;
Update information determination means for determining whether the update information received from the terminal device matches the update information specifying the update time of the third unauthorized parameter stored in the unauthorized parameter storage means;
When the update information determination unit determines that the update information does not match, the unauthorized parameter transmission unit reads out the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device;
With
The illegal parameter updating means of a credit card, the first incorrect parameter set to the stored decision tree model in the decision tree model storage unit, a third illegal parameter received is the terminal apparatus from the host computer system A payment approval system characterized by being updated. The credit card is
Risk parameter storage means for storing a risk parameter set as a quantitative threshold that requires online approval for the credit card;
Risk parameter reading means for reading a risk parameter stored in the risk parameter storage means;
The value specified in the settlement request is compared with the risk parameter, and when the value exceeds a threshold set in the risk parameter, it is determined that the settlement request meets a criterion that requires online approval. Risk parameter application means;
When it is determined that the risk parameter application means falls under a criterion that requires online approval, payment information generated from the payment request and update information that specifies when the risk parameter read from the IC chip is updated are updated to the host Second settlement information transmitting means for transmitting to the computer system;
Risk parameter update means for updating the risk parameter stored in the risk parameter storage means to a new risk parameter;
With
The host computer system is
A risk parameter storage means for storing a risk parameter set as a quantitative threshold value that requires online approval for a credit card;
Second update information determination means for determining whether the update information received from the terminal device matches the update information for specifying the update time of the latest risk parameter stored in the risk parameter storage means;
When the second update information determination means determines that the update information does not match, the risk parameter transmission means for reading the latest risk parameter from the risk parameter storage means and transmitting to the terminal device;
With
The risk parameter updating unit credit card, the risk parameters risk parameters stored in the storage unit, according to claim 1, wherein the update to the latest risk parameters the terminal device receives from the host computer system Payment approval system. A credit card settlement approval system comprising a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network,
The terminal device
A payment request receiving means for receiving a payment request by credit card;
In a decision tree model determines the necessity of the decision tree of online credit card authorization stored in the IC chip, the first decision tree model illegal parameter is set as a value serving as a reference of a branch at the contact point of the decision tree A decision tree model reading means for reading
When the values specified in the settlement request are applied in the order of the decision tree model to the contact points of the decision tree in which the first fraud parameter is set and the predetermined conditions are met, the fraudulent use of the settlement request is performed. A decision tree model application means that determines that the probability falls under a criterion that requires online approval;
When it is determined that the decision tree model application means meets the criteria requiring online approval, the payment information generated from the payment request and the update information for specifying the update time of the first illegal parameter read from the IC chip Payment information transmitting means for transmitting to the host computer system;
A fraud parameter updating means for updating the first fraud parameter set in the decision tree model stored in the IC chip to a new fraud parameter;
With
The host computer system is
A value included in the payment information received from the terminal device is applied to an fraud determination model in which a second fraud parameter that models a fraudulent usage pattern of a credit card is set, and the value included in the payment information is predetermined. A payment approval determination means for determining that the possibility of unauthorized use of the payment request corresponds to a criterion for approval of payment,
A determination result transmitting means for transmitting a determination result of approval or disapproval determined by the settlement approval determination means to the terminal device;
Illegal parameter storage means for storing a third illegal parameter set in the decision tree model;
Update information determination means for determining whether the update information received from the terminal device matches the update information specifying the update time of the third unauthorized parameter stored in the unauthorized parameter storage means;
When the update information determination unit determines that the update information does not match, the unauthorized parameter transmission unit reads the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device;
With
The unauthorized parameter updating means of the terminal device updates the first unauthorized parameter set in the decision tree model stored in the IC chip to the third unauthorized parameter transmitted from the host computer system. Settlement approval system. The terminal device
Risk parameter reading means for reading a risk parameter set as a quantitative threshold value that requires online approval for the credit card stored in the IC chip;
The value specified in the settlement request is compared with the risk parameter, and when the value exceeds a threshold set in the risk parameter, it is determined that the settlement request meets a criterion that requires online approval. Risk parameter application means;
When it is determined that the risk parameter application means falls under a criterion that requires online approval, payment information generated from the payment request and update information that specifies when the risk parameter read from the IC chip is updated are updated to the host Second settlement information transmitting means for transmitting to the computer system;
Risk parameter update means for updating the risk parameter stored in the IC chip to a new risk parameter;
With
The host computer system is
A risk parameter storage means for storing a risk parameter set as a quantitative threshold value that requires online approval for a credit card;
Second update information determination means for determining whether the update information received from the terminal device matches the update information for specifying the update time of the latest risk parameter stored in the risk parameter storage means;
When the second update information determination means determines that the update information does not match, the risk parameter transmission means for reading the latest risk parameter from the risk parameter storage means and transmitting to the terminal device;
With
4. The settlement approval system according to claim 3 , wherein the risk parameter updating means of the terminal device updates the risk parameter stored in the IC chip to the latest risk parameter transmitted from the host computer system. Used in a credit card settlement approval system comprising a credit card equipped with an IC chip and a host computer system connected via a network to a terminal device having a function of reading and writing information stored in the IC chip. A credit card,
A decision tree model that determines whether or not online approval of a credit card is necessary using a decision tree, and stores a decision tree model in which a first fraud parameter is set as a reference value for branching at a decision tree contact point Storage means;
Payment request receiving means for receiving a payment request by the credit card from the terminal device;
A decision tree model reading means for reading a decision tree model in which a first incorrect parameter is set from the decision tree model storage means;
When the values specified in the settlement request are applied in the order of the decision tree model to the contact points of the decision tree in which the first fraud parameter is set and the predetermined conditions are met, the fraudulent use of the settlement request is performed. A decision tree model application means that determines that the probability falls under a criterion that requires online approval;
When it is determined that the decision tree model applying means corresponds to a criterion requiring online approval, the terminal device is updated with the payment information generated from the payment request and the first fraud parameter read from the IC chip. Payment information transmitting means for causing the host computer system to transmit update information for identifying
And illegal parameter updating means for updating the first fraud parameters set to the stored decision tree model in the decision tree model storage means to a new invalid parameters,
With
The illegal parameter updating means, a first fraud parameters set to the stored decision tree model in the decision tree model storage unit,
A value included in the payment information received from the terminal device is applied to an fraud determination model in which a second fraud parameter that models a fraudulent usage pattern of a credit card is set, and the value included in the payment information is predetermined. The payment approval determination means for determining that the possibility of fraudulent use of the payment request corresponds to a criterion requiring online approval, and the approval or disapproval determination determined by the payment approval determination means Determination result transmitting means for transmitting a result to the terminal device, illegal parameter storage means for storing a third illegal parameter set in the decision tree model, and update information received from the terminal device is the illegal parameter storage means Update information determination means for determining whether or not the update information for specifying the update time of the third illegal parameter stored in the update information matches, and the update If the information determination means determines that the update information does not match, the terminal computer system comprising: an unauthorized parameter transmission means that reads out the third unauthorized parameter from the unauthorized parameter storage means and transmits it to the terminal device; A credit card, which is updated to a third fraud parameter received by the device. A terminal device used in a credit card settlement approval system comprising a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network. ,
A payment request receiving means for receiving a payment request by credit card;
In a decision tree model determines the necessity of the decision tree of online credit card authorization stored in the IC chip, the first decision tree model illegal parameter is set as a value serving as a reference of a branch at the contact point of the decision tree A decision tree model reading means for reading
When the values specified in the settlement request are applied in the order of the decision tree model to the contact points of the decision tree in which the first fraud parameter is set and the predetermined conditions are met, the fraudulent use of the settlement request is performed. A decision tree model application means that determines that the probability falls under a criterion that requires online approval;
When it is determined that the decision tree model application means meets the criteria requiring online approval, the payment information generated from the payment request and the update information for specifying the update time of the first illegal parameter read from the IC chip Payment information transmitting means for transmitting to the host computer system;
A fraud parameter updating means for updating the first fraud parameter set in the decision tree model stored in the IC chip to a new fraud parameter;
With
The unauthorized parameter updating means is configured to replace the first unauthorized parameter set in the decision tree model stored in the IC chip,
A value included in the payment information received from the terminal device is applied to an fraud determination model in which a second fraud parameter that models a fraudulent usage pattern of a credit card is set, and the value included in the payment information is predetermined. The payment approval determination means for determining that the possibility of fraudulent use of the payment request corresponds to a criterion for approval of payment, and the approval / disapproval determination result determined by the payment approval determination means. Determination result transmitting means for transmitting to the terminal device, incorrect parameter storage means for storing a third incorrect parameter set in the decision tree model, and update information received from the terminal device is stored in the incorrect parameter storage means Update information determination means for determining whether or not the update information specifying the updated time of the third illegal parameter is identical, and the update information determination means If it is determined that the update information does not match, the third fraud parameter transmitted from the host computer system, the third fraud parameter is read from the fraudulent parameter storage unit and transmitted to the terminal device. A terminal device that is updated to a parameter. Used in a credit card settlement approval system comprising a credit card equipped with an IC chip and a host computer system connected via a network to a terminal device having a function of reading and writing information stored in the IC chip. A host computer system,
A decision tree model that determines whether or not online approval of a credit card is necessary using a decision tree, and stores a decision tree model in which a first fraud parameter is set as a reference value for branching at a decision tree contact point storage means, the payment request accepting means and the first incorrect parameter from said decision tree model storage means reads decision tree model set decision tree model reading means from said terminal apparatus receives the payment request by the credit card, the When the values specified in the settlement request are applied in the order of the decision tree model to the contact point of the decision tree in which the first fraud parameter is set, and the predetermined condition is met, the settlement request can be illegally used. a decision tree model application unit determines that gender satisfy the criteria that require online authorization, said decision tree model application means If the terminal device is determined to satisfy the criteria that require online approval, the terminal device is provided with the payment information generated from the payment request and the update information that specifies the update time of the first fraud parameter read from the IC chip. A payment information transmission means to be transmitted to a host computer system, and a value included in the payment information received from the terminal device that has transmitted the payment information and the update information by a credit card as a model of an illegal use pattern of the credit card Applying to the fraud determination model in which the second fraud parameter is set, and the value included in the settlement information matches a predetermined condition, the possibility of fraudulent use of the settlement request requires online approval A payment approval determination means for determining that it meets the criteria
A determination result transmitting means for transmitting a determination result of approval or disapproval determined by the settlement approval determination means to the terminal device;
Illegal parameter storage means for storing a third illegal parameter set in the decision tree model;
Update information determination means for determining whether the update information received from the terminal device matches the update information specifying the update time of the third unauthorized parameter stored in the unauthorized parameter storage means;
When the update information determination unit determines that the update information does not match, the unauthorized parameter transmission unit reads out the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device;
With
In the credit card, first fraud parameters set to the stored decision tree model in the decision tree model storage means is updated from the illegal parameter transmission means to the third incorrect parameter received said terminal device A host computer system characterized by the above. A host computer system used in a credit card settlement approval system comprising a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network. A payment request receiving means for receiving a payment request by credit card, and a decision tree model for determining whether or not online approval of a credit card stored in the IC chip is necessary by a decision tree, determining first and decision tree model reading means for incorrect parameter reads a decision tree model, which is set as a value, the first contact of the decision tree invalid parameter has been set, the value specified in the payment request comprising by applying to the order of the tree model, if it meets the predetermined condition A decision tree model application unit determines that corresponds to the reference potential for unauthorized use of the payment request requires an online authorization, if the decision tree model application unit determines that satisfy the criteria that require online authorization, From a terminal device comprising payment information generated from the payment request, and payment information transmission means for transmitting to the host computer system update information that identifies when the first illegal parameter read from the IC chip is updated A value included in the received payment information is applied to a second fraud determination model in which a second fraud parameter in which a credit card fraud pattern is modeled is set. If the conditions are met, a payment approval is determined that the possibility of unauthorized use of the payment request falls within the criteria for approval of payment. And the judgment means,
A determination result transmitting means for transmitting a determination result of approval or disapproval determined by the settlement approval determination means to the terminal device;
Illegal parameter storage means for storing a third illegal parameter set in the decision tree model;
Update information determination means for determining whether the update information received from the terminal device matches the update information specifying the update time of the third unauthorized parameter stored in the unauthorized parameter storage means;
When the update information determination unit determines that the update information does not match, the unauthorized parameter transmission unit reads out the third unauthorized parameter from the unauthorized parameter storage unit and transmits it to the terminal device;
With
In the terminal device, the first unauthorized parameter set in the decision tree model stored in the IC chip is updated to the third unauthorized parameter transmitted from the unauthorized parameter transmitting means. Computer system. A credit card settlement approval method executed by a host computer system connected via a network to a credit card equipped with an IC chip and a terminal device having a function of reading and writing information stored in the IC chip. There,
A payment request receiving step for receiving a payment request by the credit card from the terminal device;
In the decision tree model in which the credit card determines whether or not online approval of the credit card is necessary by using a decision tree, a decision tree model in which a first fraud parameter is set as a reference value of a branch at a decision tree contact point A decision tree model reading step of reading the first decision tree model from the stored decision tree model storage means;
When the credit card applies the values specified in the settlement request in the order of the decision tree model to the contact points of the decision tree in which the first fraud parameter is set , A decision tree model application step for determining that the possibility of unauthorized use of the request corresponds to a criterion requiring online approval;
If it is determined that the credit card corresponds to a criterion that requires online approval in the decision tree model application step, the terminal device receives the payment information generated from the payment request and the first read from the IC chip. A settlement information transmission step for causing the host computer system to transmit update information for specifying an update time of an illegal parameter;
The host computer system applies a value included in the payment information received from the terminal device to an fraud determination model in which a second fraud parameter in which a credit card fraud pattern is modeled is set, and the payment information A payment approval determination step for determining that the possibility of fraudulent use of the payment request corresponds to a criterion for approval of payment when the value included in the value matches a predetermined condition;
A determination result transmission step in which the host computer system transmits an approval or disapproval determination result determined in the payment approval determination step to the terminal device;
The update information received from the terminal device by the host computer system specifies the time of updating the third illegal parameter stored in the illegal parameter storage means for storing the third illegal parameter set in the decision tree model An update information determination step for determining whether the update information matches the update information to be updated;
When the host computer system determines that the update information does not match in the update information determination step, an unauthorized parameter transmission step of reading the third unauthorized parameter from the unauthorized parameter storage unit and transmitting it to the terminal device;
Incorrect said credit card is first incorrect parameter set to the stored decision tree model in the decision tree model storage unit, updates from the host computer system to the third incorrect parameter received said terminal device A parameter update step;
A settlement approval method characterized by comprising: A credit card settlement approval method executed by a terminal device having a function of reading and writing an IC chip mounted on a credit card, and a host computer system connected to the terminal via a network,
A payment request receiving step in which the terminal device receives a payment request by credit card;
In the decision tree model in which the terminal device determines whether or not online approval of a credit card stored in the IC chip is necessary by using a decision tree, a first fraud parameter is set as a reference value for branching at a decision tree contact point and the decision tree model reading step of reading the decision tree model that has been,
When the terminal device applies the values specified in the settlement request in the order of the decision tree model to the contact points of the decision tree in which the first illegal parameter is set, and the predetermined condition is met, the settlement A decision tree model application step for determining that the possibility of unauthorized use of the request corresponds to a criterion requiring online approval;
When the terminal device determines that it meets the criteria that require online approval in the decision tree model applying step, the payment information generated from the payment request and the update time of the first illegal parameter read from the IC chip A payment information transmission step of transmitting update information for identifying the information to the host computer system;
The host computer system applies a value included in the payment information received from the terminal device to an fraud determination model in which a second fraud parameter that models a fraudulent use pattern of a credit card is set, and the payment A payment approval determination step for determining that the possibility of unauthorized use of the payment request corresponds to a criterion for approval of payment when a value included in the information matches a predetermined condition;
A determination result transmission step in which the host computer system transmits an approval or disapproval determination result determined in the payment approval determination step to the terminal device;
The update information received from the terminal device by the host computer system specifies the time of updating the third illegal parameter stored in the illegal parameter storage means for storing the third illegal parameter set in the decision tree model An update information determination step for determining whether the update information matches the update information to be
When the host computer system determines that the update information does not match in the update information determination step, an unauthorized parameter transmission step of reading the third unauthorized parameter from the unauthorized parameter storage unit and transmitting it to the terminal device;
An unauthorized parameter update step in which the terminal device updates the first unauthorized parameter set in the decision tree model stored in the IC chip to the third unauthorized parameter transmitted from the host computer system;
A settlement approval method characterized by comprising: