JPH0383166A - Transaction limiting fuzzy controller for automatic teller machine - Google Patents

Abstract

PURPOSE:To securely limit illegal transactions by performing fuzzy reasoning control to the best adequate extent of transaction limitation according to a fuzzy rule which contains data on fuzzy variables for obtaining proper extents of transaction limitation corresponding to respective input states. CONSTITUTION:A detecting means detects whether a user is a proper user or illegal user from the input state of an error frequency in input operation, the paid amount of money, etc., and a fuzzy reasoning control means uses the detected element as an antecedent to output an adequate transaction limitation value and performs the fuzzy reasoning control over an automatic transaction limiting means which allows or limit a transaction with said output. Consequently, even if a user who is not familiar with automatic transaction misinputs his or her password number several times, the transaction is not limited and the transaction validity of the password number is maintained. An illegal user inputs the large amount of money to be paid inevitably in addition to password number misinput operation, so the unfamiliar user and illegal user can be discriminated clearly and the reliability of the automatic teller machine is securely improved in addition to the improvement of the security of the automatic teller machine and serviceability to users.

Description

Detailed Description of the Invention (a) Industrial Application Field The present invention relates to automatic transaction machines such as those used in banking, and more specifically, transaction restrictions for automatic transaction machines with enhanced security performance during automatic transactions. Related to fuzzy control device.

(b) Conventional technology In general, this type of automated teller machine has various transaction functions such as deposit, withdrawal, transfer, bankbook entry, balance inquiry, etc. Select and execute automatic trading.

In this case, in order to avoid fraud and provide security to the automated teller machine, when a customer enters a PIN number that identifies the customer, if the PIN number is different three times in a row, then , as there were many input errors, it was determined that the PIN was used fraudulently, and subsequent transactions using this PIN were automatically set to be invalidated.

However, if an inexperienced user makes an incorrect input operation and performs the input operation many times, the number of operation errors immediately reaches the number of invalid transactions, causing an inconvenient problem in which the transaction becomes impossible.

(c) Problems to be solved by the invention This invention focuses on not only the PIN number but also the high withdrawal amount due to fraudulent use, and the transaction restriction of automatic transaction machines that makes it possible to reliably restrict fraudulent transactions. The purpose is to provide a fuzzy control device.

(d) Means for Solving the Problems This invention provides automatic transaction restriction means for restricting automatic transactions depending on the degree of unauthorized input by a customer to an automatic transaction machine, and the number of errors during input operations by the customer and withdrawal transactions. A detection means for detecting the input status of amounts, etc., and fuzzy variable data that is the appropriate degree of transaction restriction corresponding to each input situation are stored for the input status data for the automatic transaction machine detected by the detection means. The present invention is a transaction restriction fuzzy control device for an automatic transaction machine, comprising fuzzy inference control means that performs fuzzy inference control on the automatic transaction restriction means to the most appropriate degree of transaction restriction based on fuzzy rules.

(E) Effect of the Invention According to the present invention, when a customer inputs data into an automatic transaction machine, the detection means determines whether the customer is a valid customer or a fraudulent customer based on the number of errors during input operation and the input status such as the withdrawal amount. The fuzzy inference control means uses this detection element as an antecedent and outputs the most appropriate transaction limit value based on fuzzy rules to find the optimal consequent. The automatic transaction restriction means is controlled by fuzzy inference to limit the transaction.

(f) Effect of the invention Therefore, even if a user who is not familiar with automated transactions enters the PIN number incorrectly several times, the transaction will not be restricted and the validity of subsequent transactions using the PIN number will be maintained. In the case of customers, in addition to inputting incorrect PIN numbers, they inevitably input high withdrawal amounts, so it is possible to clearly distinguish between inexperienced customers and fraudulent customers, which improves the security of automated teller machines. In addition to improving service quality and customer service, the reliability of automatic transaction machines will definitely improve.

(g) Examples An embodiment of the present invention will be described in detail below based on the drawings.

The drawing shows a transaction restriction fuzzy control device for an automatic transaction machine,
This automatic transaction machine 11 has a CRT 1 on the front panel that can be used for both evening and sochino channels to display and guide customers regarding transaction operations.
2, a speaker 13 for voice guidance, a bankbook insertion slot 14, a bill slot 15, a coin slot 16, a card slot 17, a handling display 18, a call button 19 for customers to call an attendant, etc. It has various transaction functions such as deposits, withdrawals, transfers, passbook entries, and balance inquiries.

FIG. 2 shows the fuzzy inference control circuit at the time of outputting the transaction limit, and the CPU 21 operates according to the detection signal input from the automatic transaction machine 11 and the program stored in the ROM 22.
A fuzzy inference engine (hereinafter referred to as FIE) is installed in the transaction restriction device 23.
) 24, and the necessary control data at this time is stored in the RAM 25.

The above-mentioned F' I E 24 converts the current detection element determined by the detection signal from the automatic transaction machine 11 into the antecedent part Xl.
,
(Conclusion part) A fuzzy rule is set to output the most appropriate trading limit value Z1 based on the fuzzy rule from the above-mentioned detection elements, and this output Zl manages the suitability of each transaction status of the automatic transaction machine 11. Fuzzy inference control is performed to ensure high security against fraudulent transactions.

The above-mentioned fuzzy rule is set according to a predetermined fuzzy rule table 26, and is set in advance by setting an appropriate value corresponding to the customer's error frequency Xl and the customer's withdrawal amount X2. Then, compare this appropriate value with the detected facts and draw the most appropriate conclusion based on these facts in Figures 3 (a), (b) (antecedent part) and Figure 4 (consequent part). ) is set using a membership function using fuzzy variables.

In this membership function, the fuzzy set degree (
Each label (fuzzy value) indicating the grade) is divided into the direction display group of (positive direction P/standard 2) and the direction display group of (large L/
The membership function that is set to display in combination with the degree display group of medium M and small S) and shows the degree of error number XI of the customer in Figure 3 (a) is Z: 0 times (no error count) PS : A small number of errors (2 times) PM: A certain number of errors (4 times) PL: A considerable number of errors (6 or more).

The membership function showing the customer's withdrawal amount X2 in Figure 3 (b) is: Z: No withdrawal, 0 yen) PS: Relatively small withdrawal amount (10% yen) PM: Some withdrawal amount (20% yen) PL: Indicates that the withdrawal amount is considerable (30% yen or more).

The membership functions showing the customer's degree of fraud Yl in Figure 4 are as follows: Z: Outputs no fraud (0%) PS + Outputs minor fraud degree (30%) PM: Outputs medium fraud degree (60%) ) PL: Outputs a high degree of fraud (90%).

Furthermore, as an example of the setting of the fuzzy rule table 26 mentioned above, as shown in FIG. If the number of errors XI is not 1 (XI = Z), even if the customer's withdrawal amount X2 is to some extent (X2 = PM), the withdrawal will be made through proper input operations, so there will be no fraud in the withdrawal transaction. Judging as (
Yl =Z), at that time, the FIE 24 sets the output to the transaction restriction device 123 to allow normal withdrawal transactions.

Eye XI =Z X2 =PMthen
YI=Z (Rule 2) If a user inputs information into the automatic transaction machine 11 for the purpose of withdrawal, if the user's error count degree XI is a little (XI = PS), the customer's withdrawal amount degree xl If there is even a small amount (X2 = PS), it is determined that the withdrawal transaction has a very small possibility of fraud (Y1 = PS), then FIE24
While displaying on the staff monitoring panel etc. that there is a slight possibility of fraud, output settings are made to the transaction restriction device 23 to permit withdrawal transactions.

if X1=PS X2=PSthen
Yl=PS (Rule 3) If a customer inputs information into the automatic transaction machine 11 for the purpose of withdrawal, if the user's error count degree Xl is a little (Xl = PS), the customer's withdrawal amount degree xl If there are many (X2 = PL), it is determined that the withdrawal transaction has a certain possibility of fraud (Y1 = PM), then FIE2
Step 4 displays on the staff monitoring panel that there is a certain possibility of fraud, and outputs settings to the transaction restriction device 23 to permit withdrawal transactions.

if X1=PS X2=PLthen
YI=PM (Rule 4) If a customer inputs information into the automatic transaction machine 11 for the purpose of withdrawal, if the user's error frequency xl is considerable (XI = PL), the customer's withdrawal amount x2 If there are many (X2 = PL), it is determined that the withdrawal transaction is highly likely to be fraudulent (Yl = PL), then FIE2
4 outputs a predetermined transaction limit value Zl to the transaction restriction device 23 to restrict the transaction, displays a message indicating fraudulent activity on the staff monitoring panel, and displays the message "This card cannot be used" on the CRT 12 of the automatic transaction machine 11. '' and output settings to disable this card.

if XI=PL X2=PLthen
Yl = PL In this way, as shown in (rules 1 to 4)..., if
A rule is set by making the consequent part of then... correspond to the input of the antecedent part of .

The user's withdrawal amount level X2 described above can be set in accordance with the customer's balance level, and as for the input operation procedure, when the withdrawal transaction mode is specified, the indication of fraudulent activity is first set. By first inputting the withdrawal amount and then inputting and specifying the PIN number, fraud detection becomes more accurate and security is further enhanced.

The processing operation of the transaction restriction fuzzy control device of the automated teller machine 11 configured as described above will be described next.

Currently, even if a customer who is unfamiliar with automated transactions, such as entering a PIN incorrectly during a withdrawal transaction, makes a withdrawal transaction, if the withdrawal amount is small, this is inconsistent with fraudulent activity, so withdrawal transactions are allowed. However, the effectiveness of the customer's subsequent transactions is maintained, and transactions are not inadvertently restricted.

On the other hand, if in addition to inputting a high withdrawal amount, the incorrect PIN number is input several times, this is consistent with fraudulent activity, so it can be clearly determined to be fraudulent use, and in this case, withdrawal transactions will be stopped immediately. It is possible to clearly determine whether or not a transaction is appropriate based on the degree of fraud, such as by checking the degree of fraud, which not only improves the security of automatic transaction machines and improves customer service, but also improves the reliability of automatic transaction machines.

In the correspondence between this invention and the configuration of the above-described embodiment, the automatic transaction restriction means of this invention corresponds to the transaction restriction device 23 of the embodiment, and the detection means corresponds to the CPU 21, and the fuzzy inference The control means is a fuzzy inference engine (FI
E) Corresponding to 24, the present invention is not limited to the configuration of the above-described embodiment.

[Brief explanation of drawings]

The drawings show an embodiment of the present invention. Fig. 1 is an external perspective view of an automated teller machine, Fig. 2 is a fuzzy control block diagram when the automated teller machine outputs transaction limits, and Fig. 3 (a) is a customer FIG. 3(B) is a membership function diagram showing the degree of customer withdrawal amount; FIG. 4 is a membership function diagram showing the degree of fraud by the customer. 11... Automatic transaction machine 21... CPU23...
・Transaction restriction device 24...Fuzzy inference engine 26...Fuzzy rule table Fig. 2 Automatic sawing force 6n^51 叡mも7〃醜0 Fuzzy city Kanshin block diagram Fig. 3 (a), fJ Metsupan Soa ~ Banko map (b) showing the number of errors in usage and harm (B), fj user O out and slope () and Menno ``h set 1) Member Schiff Narishi Bunzu consequent 540 fuzzy fu 1z --

Claims (1)

[Claims] (1) Automatic transaction restriction means that restricts automatic transactions depending on the degree of fraudulent input by the customer to the automatic transaction machine; and detection means that detects the number of errors made by the customer during input operations and the status of input such as withdrawal transaction amount; The above-mentioned automatic transaction restriction means is applied to the input status data for the automatic transaction machine detected by the above-mentioned detection means, based on fuzzy rules containing data of fuzzy variables that are the appropriate degree of transaction restriction corresponding to each input situation. A transaction restriction fuzzy control device for an automatic transaction machine, comprising fuzzy inference control means for controlling fuzzy inference to the most appropriate degree of transaction restriction.