CN115860904A - Credit abnormal group mining method and device - Google Patents
Credit abnormal group mining method and device Download PDFInfo
- Publication number
- CN115860904A CN115860904A CN202211418296.XA CN202211418296A CN115860904A CN 115860904 A CN115860904 A CN 115860904A CN 202211418296 A CN202211418296 A CN 202211418296A CN 115860904 A CN115860904 A CN 115860904A
- Authority
- CN
- China
- Prior art keywords
- credit
- node
- abnormal
- representing
- community
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a credit abnormity group mining method, which comprises the following steps: acquiring credit users in a preset time period and credit data thereof; constructing a same composition with the credit user as a node according to the credit user and the credit data thereof; carrying out community division on the same composition by using a Louvain algorithm to obtain a plurality of communities; calculating credit abnormal group risk evaluation indexes of the communities, and judging whether the communities are credit abnormal groups or not according to the credit abnormal group risk evaluation indexes. The invention also provides a credit abnormity group mining device. The credit exception group identification method and the credit exception group identification device can accurately identify credit exception groups.
Description
Technical Field
The invention relates to the technical field of computers. More specifically, the invention relates to a credit exception group mining method and device.
Background
The credit abnormal group refers to a person who provides corresponding service by receiving a commission of the client loan in the process of the client loan, and in the process of the client application, the credit degree of the client per se is packaged and promoted by various illegal means, so that the client loans to financial institutions such as banks and the like, and then earns a spread price, and the client finally takes a part of money and owes a huge amount of loan of the banks, thereby causing property loss to both the banks and the client. The behavior recognition has higher concealment due to the fact that the behavior recognition has higher difficulty due to the characteristics of the behavior recognition, and the real information of the client and the false information after the intermediary package. For the financial institution, whether the current application operation is the customer behavior cannot be judged, and the false information of the package cannot be effectively judged due to limited data resources.
Therefore, there is a need to design a technical solution that can overcome the above-mentioned drawbacks to some extent.
Disclosure of Invention
The invention aims to provide a method and a device for mining a credit abnormal group, which can accurately identify the credit abnormal group.
To achieve these objects and other advantages and in accordance with the purpose of the invention, as embodied and broadly described herein, there is provided a credit exception population mining method including: acquiring credit users in a preset time period and credit data thereof; constructing a same composition with the credit user as a node according to the credit user and the credit data thereof; carrying out community division on the isomorphic graph by using a Louvain algorithm to obtain a plurality of communities; calculating credit abnormal group risk evaluation indexes of the communities, and judging whether the communities are credit abnormal groups or not according to the credit abnormal group risk evaluation indexes.
Further, constructing an original graph from at least the credit user, address, location, contact phone, device and IP and their associations; removing invalid nodes in the original graph; and extracting the association relationship between the credit user nodes, and constructing a same composition taking the credit user as the node.
Further, if two credit users are associated to the same non-credit user node in the original graph, then there is an associated edge for the two credit user nodes in the same graph.
Further, still include: evaluating the weight value of the associated edge in the same composition, and eliminating the associated edge with the weight value lower than a first preset value, wherein the weight value is obtained by evaluation according to the SDK characteristic information of the credit user node and the attribute information of the associated edge, and the SDK characteristic information comprises registration operation information, abnormal login operation information and account opening operation information.
Further, calculating a weight value W;
W＝Weight (a,b) *Weight(edge)
p (a,i) i-th SDK characteristic information, p, for a credit user node (b,i) The ith SDK characteristic information of the neighbor node, n is the total number of the SDK characteristic information of the credit user node, r (a,b) Is the total number of relationships, r, between the credit user node and the neighbor nodes max The maximum value of the relationship sum in the credit user node and the neighbor node;
w i weight, m, representing the ith class relationship of the associated edge i Representing the number of i-th class relationships of the associated edge, m num Representing the risk quantity of the ith type relation of the associated edge, wherein when the neighbor node is an abnormal credit node, the relation between the credit user node and the neighbor node has risk, and the abnormal credit node is determined according to a preset rule; edge i Represents an i-class edge, f (edge) i ) An attribute computation function representing an edge;
wherein (x) 1 ,...,x m ) And (y) 1 ,...,y m ) Respectively, the attribute feature values of the associated credit user nodes.
Further, still include: calculating a risk value f (node) of the credit user node, and eliminating the credit user node with the risk value lower than a second preset value;
wherein z is i The value is determined according to the degree of association,
v i weight, q, representing the ith class relationship of the associated edge i Representing the number of i-th class relationships of the associated edge, q num Representing the risk quantity of the ith type relation of the associated edge; l i Representing the number of connected anomalous credit nodes.
Further, the method also comprises the step of calculating the community structure retention degree index
Evaluating the stability of the agglomeration;
wherein λ is 1 And λ 2 Respectively are stability weighted values; s represents the number of common nodes, s 1 The number of nodes representing the community p at time t-1;
Further, a credit abnormal group risk assessment index Score is calculated according to the following formula, and the community of which the credit abnormal group risk assessment index is larger than a community abnormal threshold value is considered as a credit abnormal group.
Score＝(1+log 10 M)*(1+blackper)
Wherein M is the number of nodes in the community, and blacker is the abnormal proportion in the community.
According to another aspect of the present invention, there is also provided credit exception population mining apparatus comprising: a processor; a memory storing executable instructions; wherein the processor is configured to execute the executable instructions to perform the credit exception population mining method.
The invention at least comprises the following beneficial effects:
according to the method, the isomorphic graph with the credit users as the nodes is constructed according to the credit users and the credit data of the credit users, then the community division is carried out on the isomorphic graph by using the Louvain algorithm, the risk evaluation index of the abnormal credit group of the community is obtained through calculation, whether the community is the abnormal credit group is identified according to the risk evaluation index of the abnormal credit group, and the identification efficiency and the accuracy are high.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The present invention is further described in detail below with reference to the attached drawings so that those skilled in the art can implement the invention by referring to the description text.
It will be understood that terms such as "having," "including," and "comprising," as used herein, do not preclude the presence or addition of one or more other elements or groups thereof.
As shown in fig. 1, an embodiment of the application provides a credit exception crowd mining method, including:
s1: acquiring credit users in a preset time period and credit data thereof; optionally, selecting a history application with the calendar of nearly 3-5, rejecting a user through the user and the history application, and selecting in-line data information of the user, wherein the in-line data information comprises financial asset data, SDK (software development kit) end operation behavior data, credit application data and the like;
s2: constructing a same composition with the credit user as a node according to the credit user and the credit data thereof; the credit user nodes have an incidence relation, and the incidence relation is represented by an incidence edge, so that a same composition is established;
s3: carrying out community division on the isomorphic graph by using a Louvain algorithm to obtain a plurality of communities; the Louvain algorithm is a community discovery algorithm based on modularity, and the basic idea is that nodes in a network try to traverse community labels of all neighbors, and select a community label which maximizes modularity increment, after the modularity is maximized, each community is regarded as a new node, and the steps are repeated until the modularity is not increased any more;
s4: calculating credit abnormal group risk evaluation indexes of the communities, and judging whether the communities are credit abnormal groups or not according to the credit abnormal group risk evaluation indexes, wherein optionally, the credit abnormal group risk evaluation indexes can be determined according to the number of abnormal credit user nodes in the communities.
In other embodiments, the original graph is constructed from at least the credit user (represented by the certificate number), address, location, contact phone, equipment and IP and their associations, and may also include the credit user's custom user tags, see table 1; removing invalid nodes in the original graph; extracting the association relationship between the credit user nodes, and constructing a same composition taking the credit user as the node;
TABLE 1 node classes and attributes in the same graph
In other embodiments, if two credit users are associated to the same non-credit user node in the original graph, then there is an association edge between the two credit user nodes in the same graph, and the association relationship may be customer-operating device, customer-contact phone, customer-address, customer-IP, etc., customer-GPS, see table 2.
TABLE 2 Association edge attribute characteristics
In other embodiments, further comprising: removing invalid nodes, namely removing abnormal and invalid non-client nodes in the original graph; step 1, defining invalid node range: the number of associated nodes with the degree greater than a threshold value (100, or 40% of the number of network credit nodes); step 2, invalid node evaluation: a. the node is of an IP type, and the median difference average value of the latest login time of the associated client node is greater than a threshold value (such as 2 weeks); b. the node is an address type, and the mean value of the difference values of the most recent input time medians of the associated client node is greater than a threshold (such as 2 weeks); c. the node is the equipment type, and the median difference average value of the latest login time of the associated client node is greater than a threshold value (such as 2 weeks); d. the nodes are address types and are non-standard addresses (e.g., lack street/community dimensional information).
In other embodiments, further comprising: evaluating the weight value of the associated edge in the same composition, and eliminating the associated edge with the weight value lower than a first preset value, wherein the weight value is obtained by evaluation according to the SDK characteristic information of the credit user node and the attribute information of the associated edge, and the SDK characteristic information comprises registration operation information, abnormal login operation information and account opening operation information; for example, inefficiently associated edges having weight values lower than 0.2 are culled based on the evaluation value of the associated edge weight for the composition graph.
In other embodiments, a weight value W is calculated;
W＝Weight (a,b) *Weight(edge)
p (a,i) i-th SDK characteristic information, p, for a credit user node (b,i) The ith SDK characteristic information of the neighbor node, n is the total number of the SDK characteristic information of the credit user node, r (a,b) For credit user node to neighbor node relationshipTotal number, r max The maximum value is the maximum value in the relation sum of the credit user node and the neighbor node;
w i weight representing the i-th class of relationship of the associated edge, i.e., the association of two credit user nodes, see edge class, m in Table 2 i Number of i-th class relationships representing associated edges, m num Representing the risk quantity of the ith type relation of the associated edge, when the neighbor node is an abnormal credit user node, the relation between the credit user node and the neighbor node has risks, the abnormal credit node is determined according to a preset rule, the preset rule can be label definition, firstly, according to overdue performance of a historical credit client in a row, the user is defined as type 2 abnormity, abnormity 1, the number of first credits is over 30+, the current overdue number is equal to the current repayment number, abnormity 2, the first credits are not overdue, the second overdue is over 60+, and the current overdue number is equal to the current repayment number-1%; secondly, a rejection client label is established, the historical application is rejected, no application passes through the record, and the abnormal credit user node comprises an abnormality 1 and an abnormality 2;
edge i indicating a class i edge, f (edge) i ) An attribute computation function representing an edge;
wherein (x) 1 ，...，x m ) And (y) 1 ，...，y m ) Attribute feature values representing the associated credit user node, i.e., the values of the node attributes in table 1, such as the current number of days of expiration, time, etc., respectively.
For example, according to the flow operation of a client in a bank APP loan application, the SDK characteristic information including basic equipment characteristic information, position characteristic information, loan application flow characteristic information and sensitive operation characteristic information is substituted into the formula;
(1) Basic device class information features
Constructing basic equipment information characteristics through the equipment information acquired by the SDK at the APP terminal;
TABLE 3 basic device class characteristic information
| Frequency of operation of the apparatus | Device replacement behavior | Basic information of equipment |
| Boot time | Number of equipment login users | Whether to brush the machine or not |
| Frequency of device application/login/transfer in last week | Number of login failures after device replacement | Model of the device |
| Last week equipment early morning login frequency | Equipment IMEI associated user number | Model class |
(2) Location class feature information
Extracting IP and GPS position data through the equipment information acquired by the SDK at the APP terminal, and constructing position information characteristics;
TABLE 4 location class feature information
| IP dimension | GPS dimension |
| IP drift | Whether or not to be missing |
| IP number of device login in one week | Last week device log-in GPS number |
| Associating number of clients with IP | Associating a number of clients with a 6PS address |
| Number of devices associated with IP | Associating device numbers with GPS addresses |
(3) Loan application process class characteristic information
And (4) combing operation links according to a time sequence, including borrowing clicking, borrowing information filling confirmation, credit granting authorization, certificate photo uploading, face recognition, submission completion and the like. Constructing a characteristic system by calculating the failure times before success of various operation links and the operation duration of adjacent operation links in the process of successful application of the latest time;
TABLE 5 loan application Process class characteristic information
| Number of failures class characteristics | Duration of operation class characteristics |
| Number of times of failure of borrowing information filling confirmation | Borrowing click-borrowing information filling confirmation operation duration |
| Number of credit authorization failures | Borrowing information filling confirmation-credit authorization operation duration |
| Number of failed uploads of certificate photo | Authorization of credit-operation duration of uploading certificate photo |
| Number of face recognition failures | Uploading certificate photo-face recognition operation duration |
| Face recognition-submission completion operation duration |
(4) Sensitive operation class feature information
Extracting associated information characteristics through client operation behavior information obtained by an APP (application) end SDK, wherein the client operation behavior information comprises registration operation, abnormal login operation, account opening operation and the like;
TABLE 6 sensitive operation class characteristics information
| Registration operations | Abnormal login operation | Opening an account |
| Registration device association number | Continuous entry trend | Correlation operation statistics after account opening |
| Registration device associated IP number | Device switching aggregation distribution | Same trend and same time interval account opening trend |
| Duration of registration operation | Password modification frequency | |
| Simultaneous segment password setting similarity |
In other embodiments, further comprising: calculating a risk value f (node) of the credit user node, and eliminating the credit user node with the risk value lower than a second preset value;
wherein z is i The value is determined according to the degree of association,
v i weight, q, representing the ith class relationship of the associated edge i Representing the number of i-th class relationships of the associated edge, q num Representing the risk quantity of the ith type relation of the associated edge; l i An exceptional credit node number representing a connection; alternatively, z i Values are classified according to degrees of association, for example, one degree, 0.1 degree, two degrees, 0.05 degree, three degrees, and 0.02 degree, for example, the relationship of credit user nodes a, b, c, and d is a-b-c-d, then the one degree neighbor of a is b, the two degree neighbor of a is c, and the three degree neighbor of a is d. />
In other embodiments, the method further comprises calculating a community structure retention index
Evaluating the stability of the agglomeration;
wherein λ is 1 And λ 2 Respectively taking the stability weighted values, preferably all taking 0.5; s denotes the number of common nodes, s 1 The number of nodes representing the community p at time t-1;
representing the importance of the node; e denotes the number of common edges, e 1 Represents the number of sides of the community p at time t-1, is>
d represents the degree of the ith node in the community p, and>
representing the total degree of the nodes in the community p; the interval between the t-1 moment and the t moment can be one month or more, namely whether the clustering is stable or not is judged through newly added credit data of one month, and the accuracy of credit abnormal group mining is improved.
In other embodiments, the credit abnormal group risk assessment indicator Score is calculated according to the following formula, and the community with the credit abnormal group risk assessment indicator larger than the community abnormal threshold value is considered as the credit abnormal group.
Score＝(1+log 10 M)*(1+blackper)
Wherein M is the number of nodes in the community, and blacker is the abnormal proportion in the community, which means that the number of abnormal credit user nodes in the community is divided by the total number of nodes in the community; the community anomaly threshold is determined empirically or statistically.
Embodiments of the present application further provide a credit exception crowd mining apparatus, including: a processor; a memory storing executable instructions; wherein the processor is configured to execute the executable instructions to perform the credit exception population mining method; the processor and the memory can be selected from a PC, a server and a cloud server, and the establishment program executes the above credit exception group mining method.
The number of apparatuses and the scale of the process described herein are intended to simplify the description of the present invention. Applications, modifications and variations of the credit exception population mining method and apparatus of the present invention will be apparent to those skilled in the art.
While embodiments of the invention have been described above, it is not limited to the applications set forth in the description and the embodiments, which are fully applicable in various fields of endeavor to which the invention pertains, and further modifications may readily be made by those skilled in the art, it being understood that the invention is not limited to the details shown and described herein without departing from the general concept defined by the appended claims and their equivalents.
Claims (9)
1. The credit abnormality group mining method is characterized by comprising the following steps:
acquiring credit users in a preset time period and credit data thereof;
constructing a same composition with the credit user as a node according to the credit user and the credit data thereof;
carrying out community division on the same composition by using a Louvain algorithm to obtain a plurality of communities;
calculating credit abnormal group risk evaluation indexes of the communities, and judging whether the communities are credit abnormal groups or not according to the credit abnormal group risk evaluation indexes.
2. The credit exception population mining method of claim 1 wherein the original graph is constructed from at least credit users, addresses, locations, contact phones, devices and IPs and their associations; removing invalid nodes in the original graph; and extracting the association relationship between the credit user nodes, and constructing a same composition taking the credit user as the node.
3. The credit exception crowd mining method of claim 2, where if in the original graph two credit users are associated to the same non-credit user node, then in the same graph there is an associated edge for the two credit user nodes.
4. The credit exception population mining method of claim 3 further comprising:
evaluating the weight value of the associated edge in the same composition, and eliminating the associated edge with the weight value lower than a first preset value, wherein the weight value is obtained by evaluation according to the SDK characteristic information of the credit user node and the attribute information of the associated edge, and the SDK characteristic information comprises registration operation information, abnormal login operation information and account opening operation information.
5. The credit exception population mining method of claim 4, wherein a weight value W is calculated;
W＝Weight (a,b) *Weight(edge)
p (a，i) i-th SDK characteristic information, p, for credit user node (b，i) The ith SDK characteristic information of the neighbor node, n is the total number of the SDK characteristic information of the credit user node, r (a,b) For credit user node and neighbour nodeTotal number of relationships, r max The maximum value of the relationship sum in the credit user node and the neighbor node;
w i weight, m, representing the ith class relationship of the associated edge i Representing the number of i-th class relationships of the associated edge, m num Representing the risk quantity of the ith type relation of the associated edge, wherein when the neighbor node is an abnormal credit node, the relation between the credit user node and the neighbor node has risk, and the abnormal credit node is determined according to a preset rule; edge i Indicating a class i edge, f (edge) i ) An attribute computation function representing an edge;
wherein (x) 1 ,...,x m ) And (y) 1 ,...,y m ) Respectively representing attribute feature values of the associated credit user node.
6. The credit exception population mining method of claim 5 further comprising:
calculating a risk value f (node) of the credit user node, and eliminating the credit user nodes with the risk values lower than a second preset value;
wherein z is i The value is determined according to the degree of association,
v i weight, q, representing the ith class relationship of the associated edge i Representing the number of i-th class relationships of the associated edge, q num Representing the risk quantity of the ith type relation of the associated edge; l. the i Representing the number of connected anomalous credit nodes.
7. The credit exception population mining method of claim 4 further comprising calculating a community structure retention index
Evaluating the stability of the agglomeration;
wherein λ is 1 And λ 2 Respectively are stability weighted values; s represents the number of common nodes, s 1 The number of nodes representing the community p at time t-1;
representing the importance of the node; e denotes the number of common edges, e 1 Representing the number of edges of the community p at time t-1,
d represents the degree of the ith node in the community p, and>
representing the total number of nodes in the community p.
8. The method of claim 4, wherein the credit abnormal population mining method is characterized in that a credit abnormal population risk assessment indicator Score is calculated according to the following formula, and the communities with the credit abnormal population risk assessment indicator larger than a community abnormality threshold are considered as a credit abnormal population.
Score＝(1+log 10 M)*(1+blackper)
Wherein M is the number of nodes in the community, and blacker is the abnormal proportion in the community.
9. Credit anomaly population mining apparatus, comprising:
a processor;
a memory storing executable instructions;
wherein the processor is configured to execute the executable instructions to perform the credit exception population mining method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211418296.XA CN115860904A (en) | 2022-11-14 | 2022-11-14 | Credit abnormal group mining method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211418296.XA CN115860904A (en) | 2022-11-14 | 2022-11-14 | Credit abnormal group mining method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115860904A true CN115860904A (en) | 2023-03-28 |
Family
ID=85663322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211418296.XA Pending CN115860904A (en) | 2022-11-14 | 2022-11-14 | Credit abnormal group mining method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115860904A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107194623A (en) * | 2017-07-20 | 2017-09-22 | 深圳市分期乐网络科技有限公司 | A kind of discovery method and device of clique's fraud |
| CN109214617A (en) * | 2017-06-29 | 2019-01-15 | 格局商学教育科技（深圳）有限公司 | A kind of internet financial risks qualitative assessment auditing system |
| CN111831923A (en) * | 2020-07-14 | 2020-10-27 | 北京芯盾时代科技有限公司 | Method, device and storage medium for identifying associated specific account |
| CN112037038A (en) * | 2020-09-02 | 2020-12-04 | 中国银行股份有限公司 | Bank credit risk prediction method and device |
| CN112348659A (en) * | 2020-10-21 | 2021-02-09 | 上海淇玥信息技术有限公司 | User risk identification strategy allocation method and device and electronic equipment |
| CN114331665A (en) * | 2021-11-11 | 2022-04-12 | 中科聚信信息技术（北京）有限公司 | Training method and device for credit judgment model of predetermined applicant and electronic equipment |
-
2022
- 2022-11-14 CN CN202211418296.XA patent/CN115860904A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214617A (en) * | 2017-06-29 | 2019-01-15 | 格局商学教育科技（深圳）有限公司 | A kind of internet financial risks qualitative assessment auditing system |
| CN107194623A (en) * | 2017-07-20 | 2017-09-22 | 深圳市分期乐网络科技有限公司 | A kind of discovery method and device of clique's fraud |
| CN111831923A (en) * | 2020-07-14 | 2020-10-27 | 北京芯盾时代科技有限公司 | Method, device and storage medium for identifying associated specific account |
| CN112037038A (en) * | 2020-09-02 | 2020-12-04 | 中国银行股份有限公司 | Bank credit risk prediction method and device |
| CN112348659A (en) * | 2020-10-21 | 2021-02-09 | 上海淇玥信息技术有限公司 | User risk identification strategy allocation method and device and electronic equipment |
| CN114331665A (en) * | 2021-11-11 | 2022-04-12 | 中科聚信信息技术（北京）有限公司 | Training method and device for credit judgment model of predetermined applicant and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108009915B (en) | Marking method and related device for fraudulent user community | |
| US20240086930A1 (en) | Blockchain transaction safety | |
| CN108648071B (en) | Resource value evaluation method and device based on block chain | |
| Cortes et al. | Communities of interest | |
| WO2017133456A1 (en) | Method and device for determining risk evaluation parameter | |
| CN109063966A (en) | The recognition methods of adventure account and device | |
| CN110009365B (en) | User group detection method, device and equipment for abnormally transferring electronic assets | |
| Cortes et al. | Communities of interest | |
| CN111309788A (en) | Community structure discovery method and system for bank customer transaction network | |
| CN114187112A (en) | Training method of account risk model and determination method of risk user group | |
| CN109670933A (en) | Identify method, user equipment, storage medium and the device of user role | |
| CN116450951A (en) | Service recommendation method and device, storage medium and electronic equipment | |
| CN109271415B (en) | Data processing method and device for credit investigation database | |
| CN111797942A (en) | User information classification method and device, computer equipment and storage medium | |
| CN112749974A (en) | Transaction data processing method, device, equipment and storage medium | |
| CN114820219B (en) | Complex network-based fraud community identification method and system | |
| CN116645134A (en) | Method, device, equipment and medium for recommending credit card in stages | |
| CN115860904A (en) | Credit abnormal group mining method and device | |
| CN113870021B (en) | Data analysis method and device, storage medium and electronic equipment | |
| CN115375480A (en) | Abnormal virtual coin wallet address detection method based on graph neural network | |
| CN110570301B (en) | Risk identification method, device, equipment and medium | |
| CN113079135B (en) | Block chain phishing fraud address detection method, device, terminal and medium | |
| CN111461863A (en) | Data processing method and device, computer equipment and storage medium | |
| US11270230B1 (en) | Self learning machine learning transaction scores adjustment via normalization thereof | |
| CN116882998B (en) | Virtual currency fund analysis method |
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 |