KR20070111255A - System and method for detecting accident in banking - Google Patents

Abstract

A system and a method for detecting banking accidents are provided to detect the banking accident by storing all risk factors for the banking accidents, automatically calculating weight of each risk factor, and automatically calculating an accident score based on the weight. A risk factor managing module(10) includes a banking database(110) storing the risk factors affecting the banking accidents and manages the risk factors stored in the banking database. A banking accident data extracting module(20) extracts banking accident data from each banking accident risk factor stored in the banking database when banking transaction occurs in a banking facility. A banking accident determining module(30) determines probability of the banking accident by automatically calculating the weight of each extracted risk factor data and automatically calculating the accident score of each risk factor based on the weight. The banking database stores an employee database section, a branch database section, a transaction database section, and an account database section.

Description

System and method for detecting accident in banking

1 is a block diagram of a financial accident detection system according to the present invention.

2 is a flow chart of a financial accident detection method according to the present invention.

3 is a detailed flowchart of 100 steps of FIG.

Figure 4 is an illustration of a description file generated in the present invention.

5 is a schematic explanatory diagram of a weight calculation.

The present invention detects financial accidents early in domestic financial institutions (banks, insurance, second financial sectors, etc.) and prevents possible accidents in advance, thereby reducing the loss of financial accidents that occur annually, thereby improving the stability and transparency of the domestic financial system. A financial incident detection system and method developed to establish.

Conventionally, in the early detection and prevention methods of financial accidents in Korea, the development of mathematical models through analysis of personal information and transaction data or mining solutions is dominant, and technology development for complex financial transaction accidents is insufficient. In particular, technology development for the Basel II Capital Convention (Basel II) plays a major role in overseas companies.

At present, domestic related companies are trying to develop solutions for financial accident prevention technologies based on mining algorithms, but they have the following problems.

1) Poor data control technology for complex financial transaction incident analysis

2) Lack of standardized meta management technology that can apply developed technology to large financial institutions

3) Lack of report production and monitoring technology to utilize risk analysis results

Accordingly, the present invention provides an automated financial accident detection system and method to solve the above problems.

The present invention provides a system and method for detecting a financial accident by storing all risk factors of the financial accident as data, automatically calculating weights for each risk factor, and automatically calculating an accident score based on the weights. The purpose.

<Overview of invention>

In Figure 1, the financial accident detection system according to the present invention is largely composed of financial accident risk factor management module 10, financial accident data extraction module 20, financial accident determination module 30.

(1) Financial Accident Risk Management Module

The first and foremost necessary for establishing a risk detection system for financial accidents is to identify all possible risk factors that can affect financial accidents. In other words, the risk factors of financial accidents that can occur in banks can be extracted from the characteristics of employees, branches, or the characteristics of transactions and accounts. Therefore, the extraction of all such possible risk factors is the first and most important process in the financial accident detection system of the present invention.

Risk factors used as data in the present invention are shown in Table 1 below. Risk categories were set up by categorizing them into categories of employees, branches, transactions, and account risks, and each risk has attributes. The attributes of the risk factors are shown in Tables 2a to c below.

category property Employee Risk Gender, Age, Marriage, Education, Position, Longevity, Obligation, Salary Foreclosure, Discipline Type, Deposit, Delinquent Amount, Card Limit, Cash Service Amount, Card Payment, Card Limit, Ratio of Third Party Cards, I Number of solicitation deposits, amount of self-reservation deposit, number of solicitation loans, amount of solicitation loans Branch risk Location type, management evaluation grade, preliminary ratio, delinquency ratio, number of permanent audit records, years of service distribution, loan increase, loan decrease, deposit increase, deposit decrease, delinquency increase, delinquency, foreign exchange performance increase, new credit delinquency amount, credit Card member number skyrockets Transaction risk Separate deposit, cancellation of deposit, non-payment payment transaction, non-payment deposit cancellation transaction, withdrawal and withdrawal payment, balance before payment zero, repayment after loan repayment, payment without payment passbook, cash transaction after closing, large transaction of rest time, withdrawal immediately after big payment Large transaction before 10 o'clock, large transfer of small account Account risk Long-term dormant accounts, self-granting accounts, self-loaning loans, minor accounts, relatives, senior citizens' accounts, self-accounts, same account

Financial accident risk factor management module 10 is connected to the financial database 110, the financial database 110, employee database area 111, branch database area 112 that stores the data related to the risk factors individually , The transaction database area 113 and the account database area 114 are included.

(2) Financial Incident Data Extraction Module

Financial incident data should be extracted from each financial accident risk stored in the database as above. When a transaction occurs in a financial institution, the risk factor of the transaction is transmitted to the financial accident detection system according to the present invention in real time to request the use of a database. The daemon program is operated in the account account (account, transaction) server and the financial accident detection server of the present invention to transmit data on financial transactions at regular intervals. This will be described later in detail.

(3) Financial Incident Decision Module

In the financial accident judgment module, first, the degree of influence on the occurrence of an accident for each risk factor is automatically calculated by weight. To this end, 1) each subcategory of the risk factor is automatically calculated as a weight that affects the occurrence of an accident, and 2) the weight of each risk factor is automatically calculated through the average of each subitem weight of the risk factor. Next, the risk level (accident score) is calculated by accumulating the risk level of each transaction, employee, and branch. This will be described in detail in the following examples.

<Specific Configuration and Action of Embodiments According to the Present Invention>

In the following, the configuration and operation of the present invention will be described through the specific embodiments actually implemented, in addition to the above-described outline of the present invention.

Referring to FIG. 2, first, data is extracted from the database 110 [100]. This procedure can be done as shown in FIG. First, the financial database 110 is accessed [120] and data is extracted from each data area (employee, branch, transaction, account) [122]. Join the extracted data and create a data file. In FIG. 3, a data file 40 named "training.dat" is generated.

2, a description file is generated from the generated file [200]. A description file is a file that describes the attributes of a file. An example of a description file is shown in FIG. 4. The description file of FIG. 4 has a file name of "training.des". The following conditions were applied to generate the description file shown in FIG. 1) The first pool must declare the target value, 2) List the attribute values with "attrX:" from the second pool ("X" is the number of the attribute), 3) Between the attribute value and the attribute value Must be separated by comma (4), 4) Insert a period (.) After one declaration, 5) File name must be the same as the file name of the data file, 6) Extension must be set to "xxx.des".

As a next step, in Fig. 2, an object file is executed to calculate the weight [300]. Specifically, as shown in FIG. 5, the description file (training.des) generated above is loaded, and the weight for each attribute is calculated from the description file and the data file (training.dat). To do this, the number of attributes, attribute values, number of data, target values, etc. of the data file are extracted, and the weights of the attribute values according to the target values are calculated.

That is, this step is a step of automatically calculating the degree of influence on the occurrence of an accident for each risk factor as a weight. In order to implement a system having both accuracy and descriptiveness, the present invention is implemented using the following detailed description. 1) For each subcategory of the risk factor, the degree of impact on the occurrence of an accident is automatically calculated by weight, and 2) The weight for each risk factor is automatically calculated through the average of each subitem weight of the risk factor.

Step 1) may be performed by using Equation 1 below.

Where D _i is the weight of the _subitem of the i th risk factor and p _i is the probability of occurrence of the i th subcategory value.

In addition, step 2) may be performed using Equation 2.

Where F _k is the weight of the k th risk factor and N is the number of subitems of the k th risk factor.

Next, the weight for each attribute calculated as described above is stored in each table [400]. Each weight table is named as follows. -T_EMP_WGT (Employment Weight Table), T_BRANCH_WGT (Point Weight Table), T_ACC_WGT (Account Weight Table), T_TRAN_WGT (Transaction Weight Table).

Next, returning to FIG. 2 again, the accident score is calculated and stored [500]. First, the weight is read from the weight table stored as a file by the above step. Scoring is performed for each tuple based on the calculated weight value. After the scoring is complete, the data is loaded into the score table of the database.

Next, the risk level (accident score) is calculated by accumulating the risk level of each transaction, employee, and branch. This is a process of assigning a score based on the likelihood of accident risk for each account, employee, branch and transaction. This will be described in detail in the following examples.

(Where S is the degree of risk and k is defined as k = {employee, transaction, account, branch})

Next, the program ends [600]. This is done by deleting the data extracted from the DB.

Although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the present invention is not limited to the specific embodiments of the present invention without departing from the spirit of the present invention as claimed in the claims. Anyone skilled in the art can make various modifications, as well as such modifications are within the scope of the claims. In addition, the financial accident detection method according to the present invention can be implemented as a computer program. In this case, the scope of the present invention extends to a computer recording medium storing a computer program.

-Technical effect

Manual detection of financial incidents based on the experience of auditors cannot respond quickly to increasing frequency of financial transactions. The technology defines risk factors for financial accidents and automatically calculates their weights to quickly and accurately identify factors that affect accidents in large volumes of transaction data, and also provides automatic scoring to enable potential transactions. It can effectively detect employees, branches, etc.

i. Use as Basel II risk management solution

ii. Used for financial institution audit and internal control system (banks, insurance companies, securities companies, second financial institutions (mutual savings banks, etc.))

iii. Used for public institution audit and internal control system (construction, government-invested institution, etc.)

iv. Use as a private company audit and internal control system

Economic effect

The company is able to create a risk-controlled market that has not been activated due to complex data mining techniques and excessive cost of introduction, and import substitution that can replace the market share of foreign products with domestic products.

On the other hand, this technology is not limited to insurance and credit cards, but is used for general financial and public institutions, and can be used as a risk management solution for Basel II. It is a specialized business area that can open up overseas markets through the provision of Asian financial solutions.

Considering the start of the preoccupation activities of foreign companies in the relevant field, it seems that the amount of 100 billion won of the market size estimated at 100 billion won can be secured by the competition. The effect is high.

Claims (10)

A financial accident risk factor management module, which includes a financial database that stores risk factors affecting financial accidents, and manages financial accident risk factors stored in the financial database. A financial accident data extraction module for extracting financial accident data from each financial accident risk factor stored in a financial database of the financial accident risk factor management module when a transaction occurs at a financial institution; A financial accident detection system comprising a financial accident determination module for calculating a weight for each risk factor data extracted from the financial accident data extraction module, and determining whether there is a financial accident possibility by calculating an accident score for each risk factor. The financial database of claim 1, wherein the financial database of the financial accident risk factor management module is A financial accident detection system comprising an employee database area, a branch database area, a transaction database area, and an account database area. The method of claim 1, wherein the financial accident determination module Means for automatically calculating, by weight, the degree of impact on the occurrence of an accident for each subitem of the risk factor; A financial incident detection system comprising means for automatically calculating risk-specific weights through an average of the weights of each sub-item of the risk. Establishing a financial database that stores risk factors affecting financial incidents; When a transaction occurs in a financial institution, extracting financial accident data from each financial accident risk factor stored in a financial database of the financial accident risk factor management module, Calculating a weight for each risk factor data extracted by the financial accident data extraction module, and calculating an accident score for each risk factor to determine whether there is a possibility of a financial accident. The method of claim 4, wherein the financial database is A financial accident detection method comprising an employee database area, a branch database area, a transaction database area, and an account database area. The method of claim 4, wherein the calculating of weights for the financial accident risk factor comprises: 1) automatically calculating, by weight, the degree of impact on the occurrence of an accident for each subitem of the risk factor; 2) A method of detecting a financial accident, comprising automatically calculating a weight for each risk factor through an average of weights for each subitem of the risk factor. The method of claim 6, wherein the weight of step 1) is

(Where, D _i is the weight of the _subitem of the i th risk factor, p _i is the probability of occurrence of the i th subitem value). The method of claim 6, wherein the weight of step 2)

Wherein F _k is the weight of the k th risk factor and N is the number of subitems of the k th risk factor. The method of claim 4, wherein the accident score in the financial accident determination step is A financial incident detection method, characterized in that the scoring per tuple based on the weight value calculated in the previous step. The method of claim 9, wherein the accident score is

Wherein S is the degree of risk and k is defined as k = {employee, transaction, account, branch}.

Images