CN111105043A - Method for implementing banking case and operation risk prevention and control based on index dimension - Google Patents
Method for implementing banking case and operation risk prevention and control based on index dimension Download PDFInfo
- Publication number
- CN111105043A CN111105043A CN201911321039.2A CN201911321039A CN111105043A CN 111105043 A CN111105043 A CN 111105043A CN 201911321039 A CN201911321039 A CN 201911321039A CN 111105043 A CN111105043 A CN 111105043A
- Authority
- CN
- China
- Prior art keywords
- indexes
- index
- risk
- machine learning
- rule
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Software Systems (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Technology Law (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method for implementing prevention and control of banking cases and operation risks based on index dimensionality. The metrics are further used for rule configuration, risk profiling, machine learning model training, and intelligent rule output. The invention uses the indexes in the field of case prevention of banks for the first time, and realizes monitoring, portraying and early warning of accounts with related risk characteristics, suspicious transactions or abnormal business operations.
Description
Technical Field
The invention belongs to the field of banking case risk prevention, and particularly relates to a method for implementing banking case and operation risk prevention and control based on index dimension.
Background
In recent years, cases in the banking industry of China have multiple and high-occurrence situations, cases with huge involved amounts are successively exposed by multiple banks, criminals are rampant, economic losses are huge, cases are complex, and case manipulation is high-clear and secret, so that the whole industry is frightened, and the cases also bring extremely serious influence to the involved banks. How to effectively prevent and control the case risk in different scenes in the banking industry becomes a problem which needs to be solved urgently.
At present, most banks mainly use traditional 'general branch inspection', 'after-the-fact supervision and risk monitoring', 'audit' and the like as main means to carry out case risk prevention and control, cannot accurately position operation risks, cannot comprehensively control case clues, and is difficult to meet the requirements of current case prevention. Although a small number of banks begin to build a large data platform, the following problems still exist: 1) the rule logic is complex, business personnel can not deeply participate in the configuration of the rule, only SQL configuration rules can be written by scientific and technical personnel, the working efficiency is low, and the rule effect can not be known and targeted improvement can not be carried out in time; 2) the fine granularity of the rule needs to be improved, and the original rule lacks a flexible configuration means for the relation between each parameter and condition; 3) the case rules are complex, the rules run repeatedly, the running time is long, the performance is poor, and even if a large data platform is used, the cost of consumed hardware is still high; 4) the traditional SQL technology is based on the full data execution rule, and has poor performance and long time consumption.
Disclosure of Invention
The invention aims to provide a method for implementing the case and operation risk prevention and control of the banking industry based on index dimensionality, aiming at solving the problem of case prevention and control of different scenes of the banking industry. The indexing is the core concept of the method, so that the rules can be freely combined, sub-packaged and managed, reused and visually configured, and the data mining and analyzing capability is greatly improved through the indexed rules, machine learning and knowledge maps.
The purpose of the invention is realized by the following scheme: a method for implementing banking case and operation risk prevention and control based on index dimension comprises the following steps:
step 1, scene analysis: extracting typical risk scenes from banking cases and operation risk events by using expert experience;
step 2, feature extraction: summarizing typical risk characteristics in the typical risk scene in the step 1, corresponding to the operation mode and the operation flow of the line, judging whether a bug or a defect exists, describing the expression form of the bug and the defect, and using the expression form as a risk rule corresponding to the risk characteristic dimension;
step 3, rule logical: corresponding the rule analyzed in the step 2 to data, and forming the expression of the loophole and the defect into a conditional logic expression which comprises indexes, operators and threshold values;
step 4, index extraction: removing operators and threshold values of the conditional logic expression in the step 3, and keeping indexes;
step 5, configuring a template, writing a script, establishing mapping, processing and generating relevant indexes;
and 6, realizing the configuration of the indexing rules, risk portrayal, machine learning model training and intelligent rule output through the generated indexes.
Further, the steps 1 to 4 are performed by a service person, and the step 5 is performed by a technician.
Further, in the step 2, typical risk characteristics in the typical risk scenario in the step 1 are summarized according to different levels, including service products, customer information, operation operations, transaction behaviors, employee behaviors, and business institution levels;
corresponding to the operation mode and operation flow of the local line, judging whether the loopholes of the same system, flow and post responsibility exist or not, or the local line possibly has execution defects in the business processing and management although corresponding control measures are taken.
Further, in the step 4, the indexes include a basic index, a statistical index, a graph index, and a machine learning model index;
the basic indexes are derived from names of corresponding columns of a source data table formed by a business system;
the statistical indicator is obtained by statistical calculation of source data and comprises at least three attributes: the subject, the time range and the counted object, and can also comprise a filtering condition;
the graph indexes are used for describing graph characteristics after a knowledge graph is established by utilizing data of mining analysis;
the machine learning model index is a risk probability value obtained by a machine learning model through training prediction.
Further, in the step 4, an index library is formed through accumulation, so that the same indexes among different risk cases can not be processed repeatedly, and sharing and resource saving are realized.
Further, in the step 5, the processing and generating of the relevant indexes are specifically as follows:
establishing a one-to-one mapping relation between the field names and the source data of the service system by the basic index in a probe configuration mode;
the statistical indexes are directly processed through standard template configuration elements or through a script compiling mode;
the graph indexes need to construct a knowledge graph and are compiled and processed through a graph query language;
machine learning model indexes need to be trained through a machine learning model, and entry parameters and assignments are carried out on function variables.
Further, in the step 5, the statistical index and the graph index can be calculated in real time through an index calculation engine and a streaming calculation technology, wherein the real time incremental calculation is performed on the mass data through dynamic sliding of a time window, so that the operation efficiency and the calculation performance are greatly improved, and the timeliness of the case and defense work is improved.
Further, in step 6, the index is used for rule configuration, and specifically the following is used: service personnel use the existing indexes to flexibly configure operators and thresholds to obtain conditions, the conditions are freely combined by using ' AND ' or ' to form single-point rules, the single-point rules can be nested and quoted to form combined rules, the combined rules can be managed in a sub-package mode, the formed rule packages correspondingly monitor different risk scenes, historical data and incremental data are mined and analyzed, and strategy making-up and expanded application are achieved; and (4) performing online operation on the rule packet, namely mining and analyzing historical data and incremental data related to the risk scene, and outputting a result of the hit data.
Further, in step 6, the index is directly applied to machine learning model training and intelligent rule output, specifically as follows: carrying out index importance analysis on the indexes through a machine learning algorithm, and screening out important indexes; based on the important indexes, an operator and an optimal threshold value are automatically generated through a machine learning algorithm, and then an intelligent rule is generated; based on the important indexes, different machine learning algorithms are combined, after machine learning model training is carried out, risks can be identified more accurately, high-risk accounts can be predicted, after manual analysis and judgment, suspicious features are further refined, and indexes and rules are continuously optimized.
Further, in step 6, the index is directly applied to the risk image, specifically as follows: dividing the value of the index into different intervals, corresponding to different labels, and directly describing suspicious characteristics of customers, suspicious transactions of accounts and suspicious business operations of employees.
The invention has the beneficial effects that:
1. the index can provide convenience for the research of new business scenes
In the traditional case research process, each case is independent, the processing logic of the case is not cross-case, namely, when a banking professional starts to research a new business scene, the case must be researched from the beginning, and even for some very general rules, repeated operation is required. In order to reduce repeated workload, different case scenes are cut in by adopting the angle of indexes, the service characteristics and rule logics of the case scenes are indexed, and complete case rules are configured in an index combination mode. For different cases, the indexes have universality, and as long as the indexes of the current case scene are met, the related indexes can form specific rules in a combined configuration mode. Therefore, the indexes not only optimize the regular processing process, but also facilitate the research of banking experts on new business scenes due to the function of supporting various index combinations under different cases.
2. The index service has clear meaning and is convenient to understand and use
Each index is processed according to the existing case rule, and the processing logic of the index can be kept consistent with the rule logic to a great extent, so that each index has a very obvious business meaning, and is convenient for a user to understand and use. Meanwhile, indexes are extracted from the processing logic of the rules, and the generated indexes can completely describe the existing rules in various permutation and combination modes, so that the smooth conversion from the rule development work to the index development work is realized.
3. The index processing logic is independent, and the subsequent maintenance is convenient
The processing logic of each index is independent and does not depend on the processing of other indexes, so that the indexes are supported to be adjusted at any time in the later-stage index development process, and the experience of banking business experts is added into the indexes in time. Meanwhile, the whole process of the index processing comprises obvious processing process grades, and the data processing and the index processing are clearly and hierarchically distinguished, so that maintenance personnel can quickly and accurately find the processing logic of the index, and the maintenance cost is reduced.
Drawings
FIG. 1 is a flow chart of a method for implementing banking case and operational risk prevention and control based on an index dimension in accordance with the present invention;
FIG. 2 is a graph illustrating graph metrics according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating machine learning model metrics in an embodiment of the present invention;
FIG. 4 is a diagram illustrating an application of metrics to machine learning model training and intelligent rule output in an embodiment of the present invention.
Detailed Description
The following detailed description of the present invention will be made with reference to the accompanying drawings and examples, which are provided for illustration of the present invention and are not intended to limit the scope of the present invention.
The invention provides a method for implementing banking case and operation risk prevention and control based on index dimension, which efficiently monitors different risk scenes by using an indexing rule and accurately figures suspicious customers, suspicious transactions and suspicious employees by using an indexing label, and comprises the following steps of:
step 1, scene analysis (service personnel): extracting typical risk scenes such as internal and external appropriation of appropriated deposit, illegal collection of resources, illegal storage, telecommunication fraud, financial management flyer bill and the like from banking cases and operation risk events by using expert experience;
typical cases of banking industry such as a bank financial fraud case, an inner Mongolia Youya resource collection fraud case, a bank fund theft case and the like.
The operation risk events comprise a violation payment-based event of a certain branch in the row, a false pledge deception credit-attempted event of a certain branch in the row, a regulatory re-fine event of illegal operation of a certain bank bill in the row and the like.
Step 2, feature extraction (service personnel): summarizing typical risk characteristics in a typical risk scene in the step 1 according to different levels (service products, customer information, operation, transaction behavior, employee behavior and business institutions), corresponding to the operation mode and operation flow of the local line, judging whether vulnerabilities of the same system, flow and post responsibility exist or whether the local line possibly generates execution defects although corresponding control measures are adopted in the service processing and management, and describing the expression forms of the vulnerabilities and the defects to be used as risk rules corresponding to the risk characteristic dimension;
risk characteristics such as abnormal advance drawing of the periodic deposit, fund drawing of stolen account into lawbreaker control account, abnormal drawing of related employee account, abnormal account checking result of suspicious account and the like. The risk profile is shown in table 1 and the profile refinement rules are shown in table 2.
TABLE 1 characteristics table
TABLE 2 refinement of features into rules
Step 3, rule logical (service personnel): the rules analyzed in the step 2 are corresponding to data, whether bugs or defects have different data traces and specific expressions, and then the expressions form conditional logic expressions (including indexes, operators and threshold values).
Conditional logic expressions are as follows: account type ═ fair;
and the name of the client is not equal to the name of the account of the counterparty or the transaction date is the account opening date;
and the transaction amount is more than or equal to 500 ten thousand.
The operators are as follows: not ≠ gtoreq, belonged, unaddressed, included, excluded, etc.
The threshold values are as follows: "on a regular basis", 500 ten thousand, etc.
Step 4, index extraction (service personnel): and (3) removing operators and threshold values of the conditional logic expression in the step (3), wherein the remained operators are indexes, and the indexes comprise basic indexes, statistical indexes, graph indexes, machine learning model indexes and the like.
The basic index is derived from the names of corresponding columns of a source data table formed by a business system, such as: account type, customer name, purchase mode, impression result, etc.
The statistical indicator is obtained by statistical calculation of source data and comprises at least three attributes: the subject, the time range and the object to be counted, and may further include filtering conditions and the like. For example, "the main account accounts have a single debit in the last year for more than 500 ten thousand accumulated transaction amounts": the main body of the statistical index is a main account, the time range is 'the last year', the object to be counted is 'accumulated transaction amount', and the filtering condition is 'debit, single stroke more than 500 ten thousand'.
The graph index is used for describing the graph features after a knowledge graph is established by utilizing mining and analyzing data, for example, whether a main account collects funds and then turns to a stock fund type account through multiple layers or not, and the index describes three graph features of collection, multiple layers of transition and flow to a specific type of account and is used for carrying out pattern matching. As shown in fig. 2.
The machine learning model index is a risk probability value predicted by a machine learning model through training, and is generally represented by a value from 0 to 1, and the closer to 1, the higher the possibility of violation or crime is. Such as: the likelihood that the primary account is an illicit funding account, the likelihood that the primary account is controlled by a lawbreaker, and the like. As shown in fig. 3. The regular extraction index is shown in table 3.
TABLE 3 regular extraction index
Furthermore, through accumulation to form an index library, the same indexes among different risk cases can not be processed repeatedly, and sharing is realized, so that resources are saved. With the increasing number of the analysis cases, the margin is decreased after the index number reaches a certain scale, and the reuse rate is gradually increased. Management can be performed according to classification of different dimensions, so that the search and the call are convenient. For example, the classification is performed according to risk scenes, the classification is performed according to a statistical calculation mode, the classification is performed according to subjects such as clients, accounts, institutions and employees, the classification is performed according to business types such as deposit, loan, payment settlement and settlement.
And 5, processing and generating related indexes (technical personnel): the method comprises the steps of template configuration, script writing and mapping establishment.
The basic index establishes a one-to-one mapping relation between the field names and the source data of the service system in a probe configuration mode.
The statistical index can be configured with the main body, the object to be counted, the time range, the filtering condition, the calculating method and other elements through the standard template, and can be directly processed through writing a script.
The map indexes need to construct a knowledge map and are compiled and processed through a graph query language such as Cypher and the like.
Machine learning model indexes need to be trained through a machine learning model, and entry parameters and assignments are carried out on function variables.
Preferably, the statistical indexes and the graph indexes can be calculated in real time through an index calculation engine and a streaming calculation technology, wherein the real time incremental calculation is performed on the mass data through dynamic sliding of a time window, so that the operation efficiency and the calculation performance are greatly improved, and the timeliness of the case and defense work is improved.
And 6, aiming at the generated indexes, configuring other different operators and thresholds to form flexible combinations of flexibly-set conditions and various logics, visually configuring, generating combination rules, performing sub-package management, and mining and analyzing historical data and incremental data corresponding to different risk scenes. Three application modes are given below, but not limited thereto:
6.1 indexing rule configuration (hit mechanism);
service personnel use the existing indexes to flexibly configure operators and thresholds to obtain conditions, the conditions are freely combined by using ' AND ' or ' to form single-point rules, the single-point rules can be nested and quoted to form combined rules, the combined rules can be managed in a sub-package mode, the formed rule packages correspondingly monitor different risk scenes, historical data and incremental data are mined and analyzed, and strategy making and expanded application are achieved. And (4) performing online operation on the rule packet, namely mining and analyzing historical data and incremental data related to the risk scene, and outputting a result of the hit data. The indexing is a precondition for realizing flexible adjustment, free combination and visual configuration of rules.
6.2 direct application based on metrics (machine learning);
the indexes can be subjected to index importance analysis through a machine learning algorithm, and important indexes are screened out. Based on the important indexes, an operator and an optimal threshold value are automatically generated through a machine learning algorithm, and then an intelligent rule is generated; based on the important indexes, different machine learning algorithms are combined, after machine learning model training is carried out, risks can be identified more accurately, high-risk accounts can be predicted, after manual analysis and judgment, suspicious features are further refined, and indexes and rules are continuously optimized. As shown in fig. 4.
6.3 direct application (risk profile) based on the indicators;
dividing the value of the index into different intervals, corresponding to different labels, and directly describing suspicious characteristics of customers, suspicious transactions of accounts and suspicious business operations of employees. Such as: the transaction times of the main account are accumulated in three months, the corresponding label of <10 times is an inactive account, and the corresponding label of >360 times is a high-frequency transaction account; the number of times that the employee changes the user login password in the last year, the label corresponding to <3 times is 'the password is not changed for a long time', and the label corresponding to >50 times is 'the password is frequently changed'. And judging the risk condition of the main body according to different labels with different dimensions.
The method of the invention analyzes the banking case or operation risk event by using expert experience, restores the case scene, backtracks the operation and management defects in the business process, and then determines and generates relevant indexes for positioning relevant risks and monitoring relevant characteristics. The invention uses the indexes in the field of case defense of banks for the first time, and uses the indexing rules to early warn accounts, transactions or business operations with related risk characteristics, and carry out verification and implementation. The advantages include:
1) multiplexing common rules: based on the service index, the condition is combined through the threshold value, and the common rule can be combined through the logical AND or relationship on the basis of the condition. If the rule is in accordance with a plurality of service scenes, the rule can be nested and used by the service scenes, and the reusability of the rule is greatly improved.
2) Service scene subpackage management: in risk prevention and control, different service scenes exist and correspond to different service rules, so that the data of the different service scenes can subscribe the corresponding rule packages through sub-package management according to the sub-package management rules of the service scenes to trigger the corresponding rules, and the accuracy and the coverage rate of the rules are improved.
3) The business personnel can carry out visual configuration based on the indexes: based on the service index, the bank service personnel can independently realize visual flexible configuration, training and online of the indexing rule without depending on technical personnel, so that the service personnel can better participate in the daily case risk prevention and control work, the communication cost with technical personnel is reduced, and the rule configuration work efficiency is improved.
The above embodiments are only for illustrating the invention and are not to be construed as limiting the invention, and those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention, therefore, all equivalent technical solutions also fall into the scope of the invention, and the scope of the invention is defined by the claims.
Claims (10)
1. A method for implementing banking case and operation risk prevention and control based on index dimension is characterized by comprising the following steps:
step 1, scene analysis: extracting typical risk scenes from banking cases and operation risk events by using expert experience;
step 2, feature extraction: summarizing the typical risk characteristics in the typical risk scene in the step 1, corresponding to the operation mode and the operation flow of the line, judging whether a vulnerability or a defect exists, describing the expression form of the vulnerability or the defect, and using the vulnerability or the defect as a risk rule corresponding to the risk characteristic dimension.
Step 3, rule logical: and (3) corresponding the rule analyzed in the step (2) to data, and forming a conditional logic expression comprising indexes, operators and threshold values according to the expression of the bugs and the defects.
Step 4, index extraction: and (4) removing the operator and the threshold of the conditional logic expression in the step (3) and keeping the index.
And 5, configuring a template, writing a script, establishing mapping, processing and generating a relevant index.
And 6, realizing the configuration of the indexing rules, risk portrayal, machine learning model training and intelligent rule output through the generated indexes.
2. The method for implementing banking case and operation risk prevention and control based on index dimension as claimed in claim 1, wherein said steps 1-4 are performed by service personnel and said step 5 is performed by technical personnel.
3. The method for implementing banking case and operation risk prevention and control based on index dimension as claimed in claim 1, wherein in step 2, typical risk features in the typical risk scenario in step 1 are summarized according to different levels, including business products, customer information, operation, transaction behavior, employee behavior, and business organization levels;
corresponding to the operation mode and operation flow of the local line, judging whether the loopholes of the same system, flow and post responsibility exist or not, or the local line possibly has execution defects in the business processing and management although corresponding control measures are taken.
4. The method for implementing banking case and operation risk prevention and control based on index dimension as claimed in claim 1, wherein in said step 4, said indexes include basic indexes, statistical indexes, graph indexes, machine learning model indexes;
the basic indexes are derived from names of corresponding columns of a source data table formed by a business system;
the statistical indicator is obtained by statistical calculation of source data and comprises at least three attributes: the subject, the time range and the counted object, and can also comprise a filtering condition;
the graph indexes are used for describing graph characteristics after a knowledge graph is established by utilizing data of mining analysis;
the machine learning model index is a risk probability value obtained by a machine learning model through training prediction.
5. The method for implementing prevention and control of banking cases and operational risks based on index dimensionality as claimed in claim 1, wherein in step 4, an index library is formed by accumulation, so that the same indexes between different risk cases can not be processed repeatedly, thereby realizing resource saving by sharing.
6. The method for implementing prevention and control of banking cases and operational risks based on index dimensions as claimed in claim 1, wherein in the step 5, the processing and generating of relevant indexes are specifically as follows:
establishing a one-to-one mapping relation between the field names and the source data of the service system by the basic index in a probe configuration mode;
the statistical indexes are directly processed through standard template configuration elements or through a script compiling mode;
the graph indexes need to construct a knowledge graph and are compiled and processed through a graph query language;
machine learning model indexes need to be trained through a machine learning model, and entry parameters and assignments are carried out on function variables.
7. The method for implementing prevention and control of banking case and operational risk based on index dimension as claimed in claim 1, wherein in said step 5, statistical indexes and graph indexes can be calculated by an index calculation engine through real-time incremental calculation of dynamic sliding of time window performed on mass data by a stream type calculation technique, thereby greatly improving operation efficiency and calculation performance, and further improving timeliness of case and defense work.
8. The method for implementing prevention and control of banking cases and operational risks based on index dimension as claimed in claim 1, wherein in step 6, the index is used for rule configuration, specifically as follows: service personnel use the existing indexes to flexibly configure operators and thresholds to obtain conditions, the conditions are freely combined by using ' AND ' or ' to form single-point rules, the single-point rules can be nested and quoted to form combined rules, the combined rules can be managed in a sub-package mode, the formed rule packages correspondingly monitor different risk scenes, historical data and incremental data are mined and analyzed, and strategy making-up and expanded application are achieved; and (4) performing online operation on the rule packet, namely mining and analyzing historical data and incremental data related to the risk scene, and outputting a result of the hit data.
9. The method for implementing banking case and operation risk prevention and control based on index dimension as claimed in claim 1, wherein in said step 6, the index is directly applied to machine learning model training and intelligent rule output, specifically as follows: carrying out index importance analysis on the indexes through a machine learning algorithm, and screening out important indexes; based on the important indexes, an operator and an optimal threshold value are automatically generated through a machine learning algorithm, and then an intelligent rule is generated; based on the important indexes, different machine learning algorithms are combined, after machine learning model training is carried out, risks can be identified more accurately, high-risk accounts can be predicted, after manual analysis and judgment, suspicious features are further refined, and indexes and rules are continuously optimized.
10. The method for implementing banking case and operation risk prevention and control based on index dimension as claimed in claim 1, wherein in said step 6, the index is directly applied to the risk profile, specifically as follows: dividing the value of the index into different intervals, corresponding to different labels, and directly describing suspicious characteristics of customers, suspicious transactions of accounts and suspicious business operations of employees.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911321039.2A CN111105043B (en) | 2019-12-19 | 2019-12-19 | Method for implementing banking case and operation risk prevention and control based on index dimension |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911321039.2A CN111105043B (en) | 2019-12-19 | 2019-12-19 | Method for implementing banking case and operation risk prevention and control based on index dimension |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111105043A true CN111105043A (en) | 2020-05-05 |
| CN111105043B CN111105043B (en) | 2023-09-05 |
Family
ID=70423698
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911321039.2A Active CN111105043B (en) | 2019-12-19 | 2019-12-19 | Method for implementing banking case and operation risk prevention and control based on index dimension |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111105043B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113223200A (en) * | 2021-05-07 | 2021-08-06 | 广州天长信息技术有限公司 | Road stealing and escaping intelligent prevention and control method, storage medium and system based on index dimension |
| CN117252446A (en) * | 2023-11-15 | 2023-12-19 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545360A (en) * | 2017-07-28 | 2018-01-05 | 浙江邦盛科技有限公司 | A kind of air control intelligent rules deriving method and system based on decision tree |
| CN109389486A (en) * | 2018-08-27 | 2019-02-26 | 深圳壹账通智能科技有限公司 | Loan air control rule adjustment method, apparatus, equipment and computer storage medium |
| CN109767067A (en) * | 2018-12-13 | 2019-05-17 | 平安医疗健康管理股份有限公司 | Method and Related product based on more evaluative dimensions evaluation hospital |
| CN110009174A (en) * | 2018-12-13 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Risk identification model training method, device and server |
| CN110264336A (en) * | 2019-05-28 | 2019-09-20 | 浙江邦盛科技有限公司 | A kind of anti-system of intelligent case based on big data |
-
2019
- 2019-12-19 CN CN201911321039.2A patent/CN111105043B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545360A (en) * | 2017-07-28 | 2018-01-05 | 浙江邦盛科技有限公司 | A kind of air control intelligent rules deriving method and system based on decision tree |
| CN109389486A (en) * | 2018-08-27 | 2019-02-26 | 深圳壹账通智能科技有限公司 | Loan air control rule adjustment method, apparatus, equipment and computer storage medium |
| CN109767067A (en) * | 2018-12-13 | 2019-05-17 | 平安医疗健康管理股份有限公司 | Method and Related product based on more evaluative dimensions evaluation hospital |
| CN110009174A (en) * | 2018-12-13 | 2019-07-12 | 阿里巴巴集团控股有限公司 | Risk identification model training method, device and server |
| CN110264336A (en) * | 2019-05-28 | 2019-09-20 | 浙江邦盛科技有限公司 | A kind of anti-system of intelligent case based on big data |
Non-Patent Citations (1)
| Title |
|---|
| 蒋韬: "大数据和人工智能视角下的银行业风险防控" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113223200A (en) * | 2021-05-07 | 2021-08-06 | 广州天长信息技术有限公司 | Road stealing and escaping intelligent prevention and control method, storage medium and system based on index dimension |
| CN117252446A (en) * | 2023-11-15 | 2023-12-19 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
| CN117252446B (en) * | 2023-11-15 | 2024-02-13 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111105043B (en) | 2023-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10607228B1 (en) | Dynamic rule strategy and fraud detection system and method | |
| Ogwueleka | Data mining application in credit card fraud detection system | |
| Rieke et al. | Fraud detection in mobile payments utilizing process behavior analysis | |
| CN110866820A (en) | Real-time monitoring system, method, equipment and storage medium for banking business | |
| CN111177250A (en) | Abnormal transaction monitoring method, system and storage medium | |
| CN110348528A (en) | Method is determined based on the user credit of multidimensional data mining | |
| CN108388969A (en) | Inside threat personage's Risk Forecast Method based on personal behavior temporal aspect | |
| CN111461216A (en) | Case risk identification method based on machine learning | |
| CN111105043A (en) | Method for implementing banking case and operation risk prevention and control based on index dimension | |
| CN107103479A (en) | The real-time anti-fake system of financial transaction | |
| CN106681980B (en) | A kind of refuse messages analysis method and device | |
| CN114140123B (en) | Method and system for tracing two-layer network transaction of Ethernet workshop | |
| Madhuri et al. | Big-data driven approaches in materials science for real-time detection and prevention of fraud | |
| CN113709170A (en) | Asset safe operation system, method and device | |
| Torres et al. | A proposal for online analysis and identification of fraudulent financial transactions | |
| Palshikar et al. | Financial security against money laundering: A survey | |
| WO2011025689A1 (en) | Integrated fraud platform | |
| CN105204869B (en) | A kind of data processing method of communication system | |
| Zainal et al. | A review on computer technology applications in fraud detection and prevention | |
| CN115719281A (en) | Capital income and expenditure management system | |
| CN113592499B (en) | Internet money laundering countermeasure method and device | |
| CN114493858A (en) | Illegal fund transfer suspicious transaction monitoring method and related components | |
| CN113162891B (en) | Attack flow rapid identification system, method, computer readable medium and equipment | |
| CN110096532A (en) | A kind of safety in production big data analysis method for digging and system | |
| CN114723548A (en) | Data processing method, apparatus, device, medium, and program product |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room ABCD, 17th floor, building D, Paradise Software Park, No.3 xidoumen Road, Xihu District, Hangzhou City, Zhejiang Province, 310012 Applicant after: Zhejiang Bangsheng Technology Co.,Ltd. Address before: Room ABCD, 17th floor, building D, Paradise Software Park, No.3 xidoumen Road, Xihu District, Hangzhou City, Zhejiang Province, 310012 Applicant before: ZHEJIANG BANGSUN TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |