CN109360099A - A kind of anti-fraud method of finance based on k- nearest neighbor algorithm - Google Patents
A kind of anti-fraud method of finance based on k- nearest neighbor algorithm Download PDFInfo
- Publication number
- CN109360099A CN109360099A CN201811230707.6A CN201811230707A CN109360099A CN 109360099 A CN109360099 A CN 109360099A CN 201811230707 A CN201811230707 A CN 201811230707A CN 109360099 A CN109360099 A CN 109360099A
- Authority
- CN
- China
- Prior art keywords
- data
- fraud
- database
- client
- nearest neighbor
- 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.)
- Pending
Links
Classifications
-
- 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
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
-
- 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
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing
Abstract
The present invention provides a kind of anti-fraud method of finance based on k- nearest neighbor algorithm, comprising the following steps: the database B for establishing and store the database A of client's general information, obtain the Transaction Information of client in real time obtains the Transaction Information of client in real time；Transaction Information is normalized；Calculate the distance between data in current data point and database A；Resulting all data will be calculated to arrange according to incremental sequence, choose and current point is apart from the smallest k point；The number of arm's length dealing event and fraud, calculates the probability of two kinds of events appearance where k point of Iterative statistical；Whether belong to the result of fraud as the financial transaction behavior.A kind of anti-fraud method of finance based on k- nearest neighbor algorithm provided by the invention, classifies to client trading information data using k- nearest neighbor algorithm, machine learning is combined with financial field, financial fraud discrimination is effectively improved, reduces financial risks.
Description
Technical field
The present invention relates to financial anti-fraud fields, more particularly to a kind of anti-fraud of finance based on k- nearest neighbor algorithm
Method.
Background technique
As financial industry develops, by various channels carry out financial services transactions the case where it is more more and more universal,
Simultaneously also along with many criminals using the various frauds of loophole progress of each platform, but due to financial product can
Selectivity is excessively abundant, but corresponding Risk-recovery measure is no in place, and client is for various financial products
Solution is not deep enough, so the event of financial fraud frequently occurs, causes huge damage to banking establishments and financial product client
It loses.Therefore, it is particularly significant to carry out the anti-air control work cheated of finance.
The anti-fraud problems of finance are directed to, the most common risk censorship is the base by manually carrying out under normal circumstances
Although good effect can be obtained in financial anti-fraud by carrying out the anti-fraud of finance in this method, manual review has not really
Qualitative and unstability can not be real moreover the quantity of financial transaction daily is very huge, needs to expend huge human resources
Applicable market.In order to overcome this difficulty, previously it has been proposed that a kind of anti-fraudulent party of finance based on model-naive Bayesian
Method, but this method needs to know prior probability, and prior probability many times depends on it is assumed that the model because of hypothesis has not
Certainty, therefore prediction effect can be caused bad because of the reason of prior model in certain circumstances, it is counter to finance to take advantage of
The effect of swindleness is poor, it is difficult to which put goods on the market use.
Summary of the invention
The present invention be overcome the anti-fraud method of existing finance there are prediction effects it is bad, to the financial anti-effect cheated compared with
Difference, it is difficult to which put goods on the market the technological deficiency used, provides a kind of anti-fraud method of finance based on k- nearest neighbor algorithm.
In order to solve the above technical problems, technical scheme is as follows:
A kind of anti-fraud method of finance based on k- nearest neighbor algorithm, comprising the following steps:
S1: establishing the database A of storage client's general information, will be in the recent Financial Information deposit database A of client；
S2: it establishes database B and stores collected client trading information, obtain the Transaction Information of client in real time；
S3: the Transaction Information that will acquire is normalized；
S4: according to the data after normalized, the distance between data in current data point and database A are calculated；
S5: will calculate resulting all data and arrange according to incremental sequence, chooses and current point is apart from the smallest k
Point；
S6: the number of arm's length dealing event and fraud where k point of Iterative statistical calculates what two kinds of events occurred
Probability size；
S7: multilevel iudge returns to the big person of probability as currently judging as a result, whether belonging to as the financial transaction behavior
In the result of fraud.
Wherein, in step sl, the recent Financial Information of the client includes being transferred within nearest three months the amount of money, three months nearest
Produce the amount of money, nearest three months maximum spending amounts, nearest three months minimum spending amounts.
Wherein, in step s 2, data acquisition service is carried out by a variety of collecting methods, passes data to distribution
Formula message queue introduces Storm frame and carries out the cleaning of data, statistics, the data handled are stored in distributed SQL data
In library, i.e. in database B.
Wherein, the step S3 specifically:
S31: the data taken out in database B are matched one by one with the data in database A；
S32: the data after matching are normalized, specific formula are as follows:
Wherein, CalValue is the data after normalization；Value is the data for calculating point；Minvalue is all data
In minimum value；Maxvalue is the maximum value in all data；Any value range numerical value is gone between the 0-1 of section.
Wherein, in the step S5, using the method for bubble sort method, resulting all data will be calculated according to incremental
Sequence line up.
Wherein, in the step S5, the k number that will sort evidence is denoted as: k1,k2,…,ki, i is the number of iterations；
In step s 6, it is assumed that belonging to fraud number is a, and arm's length dealing number is b；Initialization the number of iterations i value be
1, fraud number a is 0, and arm's length dealing number b is 1；
Judge kiWhich kind of event belonged to, wherein i is the integer no more than 20；If kiFor fraud, then a=a+1；If ki
For normal transaction event, then b=b+1；Until iteration is completed, the value of a, b are recorded.
Wherein, the specific steps of the step S7 are as follows:
S71: it is by the probability that Classical Probability Spaces probability calculation formula calculating current data point belongs to fraudThe probability for belonging to arm's length dealing event is
S72: comparing the size of P (a), P (b), if P (a) > P (b), judges that the transaction situation belongs to fraud；
If P (a) < P (b), judge that the transaction belongs to arm's length dealing behavior.
In above scheme, if it is determined that the transaction situation belongs to fraud, then it is different to determine that the finance account to give belongs to
Often, and by this situation client is fed back, is cheated with achieving the purpose that finance is counter.If it is determined that the transaction situation belongs to arm's length dealing feelings
Condition does not take any measure then, which is normally carried out.
Compared with prior art, the beneficial effect of technical solution of the present invention is:
A kind of anti-fraud method of finance based on k- nearest neighbor algorithm provided by the invention, by establishing client's general information number
On the basis of library A and real-time acquisition client trading information, all data are normalized, using k- nearest neighbor algorithm pair
Client trading information data is classified, and the probability that this finance expenditure belongs to arm's length dealing with trades extremely is calculated separately, from
And determines whether this financial transaction belongs to and abnormal achieve the purpose that finance is counter and cheat.The present invention is by machine learning and financial field
It combines, effectively improves financial fraud discrimination, reduce financial risks.
Detailed description of the invention
Fig. 1 is invention flow diagram.
Fig. 2 is the financial transaction information schematic diagram that the present invention obtains client in real time.
Fig. 3 is data normalization processing schematic of the present invention.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent；
In order to better illustrate this embodiment, the certain components of attached drawing have omission, zoom in or out, and do not represent actual product
Size；
To those skilled in the art, it is to be understood that certain known features and its explanation, which may be omitted, in attached drawing
's.
The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
Embodiment 1
As shown in Figure 1, Figure 2, Figure 3 shows, a kind of anti-fraud method of finance based on k- nearest neighbor algorithm, comprising the following steps:
S1: establishing the database A of storage client's general information, will be in the recent Financial Information deposit database A of client；
S2: it establishes database B and stores collected client trading information, obtain the Transaction Information of client in real time；
S3: the Transaction Information that will acquire is normalized；
S4: according to the data after normalized, the distance between data in current data point and database A are calculated；
S5: will calculate resulting all data and arrange according to incremental sequence, chooses and current point is apart from the smallest k
Point；
S6: the number of arm's length dealing event and fraud where k point of Iterative statistical calculates what two kinds of events occurred
Probability size；
S7: multilevel iudge returns to the big person of probability as currently judging as a result, whether belonging to as the financial transaction behavior
In the result of fraud.
More specifically, in step sl, the recent Financial Information of the client includes being transferred to the amount of money, nearest three in nearest three months
Produce within a month the amount of money, nearest three months maximum spending amounts, nearest three months minimum spending amounts.
More specifically, in step s 2, data acquisition service is carried out by a variety of collecting methods, passed data to
Distributed Message Queue introduces Storm frame and carries out the cleaning of data, statistics, the data handled are stored in distributed SQL
In database, i.e. in database B.
More specifically, the step S3 specifically:
S31: the data taken out in database B are matched one by one with the data in database A；
S32: the data after matching are normalized, specific formula are as follows:
More specifically, CalValue is the data after normalization；Value is the data for calculating point；Minvalue is all
Minimum value in data；Maxvalue is the maximum value in all data；By any value range numerical value go to section 0-1 it
Between.
More specifically, in the step S5, using the method for bubble sort method, will calculate resulting all data according to
Incremental sequence is lined up.
More specifically, in the step S5, the k number that will sort evidence is denoted as: k1,k2,…,ki, i is the number of iterations；
In step s 6, it is assumed that belonging to fraud number is a, and arm's length dealing number is b；Initialization the number of iterations i value be
1, fraud number a is 0, and arm's length dealing number b is 1；
Judge kiWhich kind of event belonged to, wherein i is the integer no more than 20；If kiFor fraud, then a=a+1；If ki
For normal transaction event, then b=b+1；Until iteration is completed, the value of a, b are recorded.
More specifically, the specific steps of the step S7 are as follows:
S71: it is by the probability that Classical Probability Spaces probability calculation formula calculating current data point belongs to fraudThe probability for belonging to arm's length dealing event is
S72: comparing the size of P (a), P (b), if P (a) > P (b), judges that the transaction situation belongs to fraud；
If P (a) < P (b), judge that the transaction belongs to arm's length dealing behavior.
In the specific implementation process, the relevant information of client is collected first, and collected information is analyzed, and is determined
Relationship between information characteristics, to define database structure type, and then among the A of storing data library.
In the specific implementation process, if it is determined that the transaction situation belongs to fraud, then determine the finance account to give
Belong to exception, and this situation is fed back into client, is cheated with achieving the purpose that finance is counter.If it is determined that the transaction situation belongs to normally
Trading situation does not take any measure then, which is normally carried out.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (7)
1. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm, which comprises the following steps:
S1: establishing the database A of storage client's general information, will be in the recent Financial Information deposit database A of client；
S2: it establishes database B and stores collected client trading information, obtain the Transaction Information of client in real time；
S3: the Transaction Information that will acquire is normalized；
S4: according to the data after normalized, the distance between data in current data point and database A are calculated；
S5: resulting all data will be calculated and arranged according to incremental sequence, choose and current point is apart from the smallest k point；
S6: the number of arm's length dealing event and fraud where k point of Iterative statistical calculates the probability of two kinds of events appearance
Size；
S7: multilevel iudge returns to the big person of probability takes advantage of as currently judging as a result, whether belonging to as the financial transaction behavior
The result of swindleness behavior.
2. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 1, which is characterized in that in step
In S1, the recent Financial Information of client includes being transferred within nearest three months the amount of money, producing within nearest three months the amount of money, is three months nearest
Maximum spending amount, nearest three months minimum spending amounts.
3. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 1, which is characterized in that in step
In S2, data acquisition service is carried out by a variety of collecting methods, passes data to Distributed Message Queue, is introduced
Storm frame carries out the cleaning of data, statistics, the data handled is stored in distributed SQL database, i.e. database B
In.
4. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 1, which is characterized in that the step
Rapid S3 specifically:
S31: the data taken out in database B are matched one by one with the data in database A；
S32: the data after matching are normalized, specific formula are as follows:
Wherein, CalValue is the data after normalization；Value is the data for calculating point；Minvalue is in all data
Minimum value；Maxvalue is the maximum value in all data；Any value range numerical value is gone between the 0-1 of section.
5. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 1, it is characterised in that: described
In step S5, using the method for bubble sort method, resulting all data will be calculated and lined up according to incremental sequence.
6. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 5, it is characterised in that: described
In step S5, the k number to have sorted evidence is denoted as: k1,k2,…,ki, i is the number of iterations；
In step s 6, it is assumed that belonging to fraud number is a, and arm's length dealing number is b；The value for initializing the number of iterations i is 1, is taken advantage of
Cheating number a is 0, and arm's length dealing number b is 1；
Judge kiWhich kind of event belonged to, wherein i is the integer no more than 20；If kiFor fraud, then a=a+1；If kiIt is positive
Normal transaction event, then b=b+1；Until iteration is completed, the value of a, b are recorded.
7. a kind of anti-fraud method of finance based on k- nearest neighbor algorithm according to claim 6, it is characterised in that: the step
The specific steps of rapid S7 are as follows:
S71: it is by the probability that Classical Probability Spaces probability calculation formula calculating current data point belongs to fraud
The probability for belonging to arm's length dealing event is
S72: comparing the size of P (a), P (b), if P (a) > P (b), judges that the transaction situation belongs to fraud；If P
(a) < P (b) then judges that the transaction belongs to arm's length dealing behavior.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811230707.6A CN109360099A (en) | 2018-10-22 | 2018-10-22 | A kind of anti-fraud method of finance based on k- nearest neighbor algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811230707.6A CN109360099A (en) | 2018-10-22 | 2018-10-22 | A kind of anti-fraud method of finance based on k- nearest neighbor algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109360099A true CN109360099A (en) | 2019-02-19 |
Family
ID=65346231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811230707.6A Pending CN109360099A (en) | 2018-10-22 | 2018-10-22 | A kind of anti-fraud method of finance based on k- nearest neighbor algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109360099A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993538A (en) * | 2019-02-28 | 2019-07-09 | 同济大学 | Identity theft detection method based on probability graph model |
| CN111833175A (en) * | 2020-06-03 | 2020-10-27 | 百维金科(上海)信息科技有限公司 | Internet financial platform application fraud behavior detection method based on KNN algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105138641A (en) * | 2015-08-24 | 2015-12-09 | 河海大学 | Angle-based high dimensional data outlier detection method |
| CN105894038A (en) * | 2016-04-22 | 2016-08-24 | 天云融创数据科技（北京）有限公司 | Credit card fraud prediction method based on signal transmission and link mode |
| CN106372938A (en) * | 2015-07-21 | 2017-02-01 | 华为技术有限公司 | Abnormal account identification method and system |
| CN108229963A (en) * | 2016-12-12 | 2018-06-29 | 阿里巴巴集团控股有限公司 | The Risk Identification Method and device of user's operation behavior |
| CN108305166A (en) * | 2018-04-04 | 2018-07-20 | 淮阴师范学院 | A kind of negative dealing fraud method towards financial field |
-
2018
- 2018-10-22 CN CN201811230707.6A patent/CN109360099A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372938A (en) * | 2015-07-21 | 2017-02-01 | 华为技术有限公司 | Abnormal account identification method and system |
| CN105138641A (en) * | 2015-08-24 | 2015-12-09 | 河海大学 | Angle-based high dimensional data outlier detection method |
| CN105894038A (en) * | 2016-04-22 | 2016-08-24 | 天云融创数据科技（北京）有限公司 | Credit card fraud prediction method based on signal transmission and link mode |
| CN108229963A (en) * | 2016-12-12 | 2018-06-29 | 阿里巴巴集团控股有限公司 | The Risk Identification Method and device of user's operation behavior |
| CN108305166A (en) * | 2018-04-04 | 2018-07-20 | 淮阴师范学院 | A kind of negative dealing fraud method towards financial field |
Non-Patent Citations (1)
| Title |
|---|
| 张德舜: "《高等数学》", 30 June 1996, 中国医药科技出版社 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993538A (en) * | 2019-02-28 | 2019-07-09 | 同济大学 | Identity theft detection method based on probability graph model |
| CN111833175A (en) * | 2020-06-03 | 2020-10-27 | 百维金科(上海)信息科技有限公司 | Internet financial platform application fraud behavior detection method based on KNN algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106709800A (en) | Community partitioning method and device based on characteristic matching network | |
| US7296734B2 (en) | Systems and methods for scoring bank customers direct deposit account transaction activity to match financial behavior to specific acquisition, performance and risk events defined by the bank using a decision tree and stochastic process | |
| CN108053318B (en) | Method and device for identifying abnormal transactions | |
| CN111476660B (en) | Intelligent wind control system and method based on data analysis | |
| Li et al. | Identifying the signs of fraudulent accounts using data mining techniques | |
| US20090222323A1 (en) | Opportunity segmentation | |
| US20070226129A1 (en) | System and method of detecting mortgage related fraud | |
| US20130124393A1 (en) | Connecting decisions through customer transaction profiles | |
| CN105095238A (en) | Decision tree generation method used for detecting fraudulent trade | |
| CN109034194A (en) | Transaction swindling behavior depth detection method based on feature differentiation | |
| EP2946355A1 (en) | System and method for characterizing financial messages | |
| CN108446291A (en) | The real-time methods of marking and points-scoring system of user credit | |
| CN109509082A (en) | The monitoring method and device of bank application system | |
| CN109934700A (en) | A kind of method and device of arbitrage detecting | |
| CN110458576A (en) | The network trading that detects is counter in a kind of fusion ex ante forecasting and thing cheats method | |
| CN109360099A (en) | A kind of anti-fraud method of finance based on k- nearest neighbor algorithm | |
| CN112232950A (en) | Loan risk assessment method and device, equipment and computer-readable storage medium | |
| CN109993538A (en) | Identity theft detection method based on probability graph model | |
| Akyildirim et al. | Applications of signature methods to market anomaly detection | |
| He et al. | An efficient solution to detect common topologies in money launderings based on coupling and connection | |
| Dastidar et al. | The importance of future information in credit card fraud detection | |
| CN107679862A (en) | A kind of characteristic value of fraudulent trading model determines method and device | |
| CN117114689A (en) | Fraud detection model construction method, fraud detection model construction device, fraud detection model construction equipment and storage medium | |
| CN110706111A (en) | Method and device for identifying suspicious transaction account, storage medium and server | |
| CN110458684A (en) | A kind of anti-fraud detection method of finance based on two-way shot and long term Memory Neural Networks |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190219 |