CN115203641A - Training method, device, equipment and storage medium for card-raising recognition model - Google Patents
Training method, device, equipment and storage medium for card-raising recognition model Download PDFInfo
- Publication number
- CN115203641A CN115203641A CN202110391185.3A CN202110391185A CN115203641A CN 115203641 A CN115203641 A CN 115203641A CN 202110391185 A CN202110391185 A CN 202110391185A CN 115203641 A CN115203641 A CN 115203641A
- Authority
- CN
- China
- Prior art keywords
- data
- card
- model
- agent
- raising
- 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
- 238000000034 method Methods 0.000 title claims abstract description 76
- 238000012549 training Methods 0.000 title claims abstract description 72
- 230000002159 abnormal effect Effects 0.000 claims abstract description 56
- 238000012423 maintenance Methods 0.000 claims abstract description 48
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 34
- 238000012417 linear regression Methods 0.000 claims abstract description 17
- 230000006399 behavior Effects 0.000 claims description 135
- 230000003542 behavioural effect Effects 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 3
- 239000003795 chemical substances by application Substances 0.000 description 145
- 238000004891 communication Methods 0.000 description 22
- 238000010586 diagram Methods 0.000 description 16
- 238000011161 development Methods 0.000 description 15
- 230000008569 process Effects 0.000 description 13
- 230000006870 function Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004634 feeding behavior Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression 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
- 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
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Computational Mathematics (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Quality & Reliability (AREA)
- Mathematical Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a training method, a device, equipment and a storage medium for a card raising identification model, wherein the training method for the card raising identification model comprises the following steps: acquiring a training sample set; determining abnormal value data from the business data of the first historical agent; inputting the abnormal value data into a linear regression algorithm, and outputting target abnormal value data of a high risk value; inputting the target abnormal value data into a quartile algorithm of a target model, and outputting a plurality of model threshold values; determining a plurality of model thresholds as parameters of a preset model to obtain an original model; inputting the service data of the historical agent into an original model and outputting result data; on the premise that the hit rate of the result data and the card maintenance behavior tag data which are consistent is smaller than a preset threshold value, adjusting parameters of the original model, returning to input the business data of the historical agent into the original model until the hit rate is not smaller than the preset threshold value; the method and the device for identifying the card-holding behavior can solve the problem that an existing method for identifying the card-holding behavior is low in identification precision.
Description
Technical Field
The application belongs to the technical field of communication, and particularly relates to a training method, device, equipment and storage medium for a card-raising recognition model.
Background
With the development of the telecommunication industry, commission rule management processes become more and more complex, commission settlement complexity is continuously improved, meanwhile, due to the fact that part of marketing policies are unreasonable in business design, not standard in operation, incomplete in system control, not in-place in enterprise management and the like, business risks are increased, and some agents utilize loopholes and weak links existing in the operation management processes and maliciously utilize rules to raise cards and arbitrage.
For the behavior that the agent violates the rule of card care, the telecom operator also has some methods for identifying the card care behavior in the daily operation process, but the existing method for identifying the card care behavior has lower identification precision.
Disclosure of Invention
The embodiment of the application provides a training method, a training device, equipment and a storage medium for a card raising recognition model, and can solve the problem that the existing method for recognizing the card raising behavior is low in recognition accuracy.
In a first aspect, an embodiment of the present application provides a training method for a card raising recognition model, including:
acquiring a training sample set, wherein the training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card-raising behavior label data corresponding to the business data of the historical agent;
determining abnormal value data which accords with a preset abnormal value judgment rule from the business data of the first historical agent, wherein the business data of the first historical agent represents the business data of the historical agent corresponding to the tag data of the card maintenance behavior with the card maintenance behavior;
inputting the abnormal value data into a preset linear regression algorithm for risk estimation, and outputting target abnormal value data with a high risk value;
inputting the target abnormal value data into a quartile algorithm of a target model to perform model threshold calculation, and outputting a plurality of model thresholds;
determining a plurality of model thresholds as parameters of a preset model to obtain an original model;
inputting the business data of the historical agent into an original model, and outputting result data representing whether card maintenance behaviors exist or not;
and on the premise that the hit rate of the result data and the card maintenance behavior tag data is consistent and is less than a preset threshold value, adjusting parameters of the original model, returning to input the business data of the historical agent into the original model until the hit rate is not less than the preset threshold value.
Further, in one embodiment, the business data of the historical brokers comprises:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data.
Further, in one embodiment, the preset outlier determination rule includes a preset 95-quantile algorithm;
determining abnormal value data which accords with a preset abnormal value judgment rule from service data of a historical agent corresponding to the tag data of the card maintenance behavior representing the existence of the card maintenance behavior according to the preset abnormal value judgment rule, wherein the abnormal value data comprises the following steps:
and inputting the service data of the historical agent corresponding to the tag data of the card-raising behavior representing the existence of the card-raising behavior into a preset 95-quantile algorithm to determine abnormal value data, and outputting the abnormal value data.
Further, in one embodiment, the predetermined model is selected as a decision tree model.
In a second aspect, an embodiment of the present application provides a method for calling a card-maintenance identification model to identify a card-maintenance behavior, where the card-maintenance identification model is obtained by training according to a claimed method, and the method includes:
acquiring service data of a target agent, wherein the service data of the target agent comprises at least one of the following items: behavior data of the target agent, user behavior data of the target agent and charge-out data of the target agent;
and inputting the business data of the target agent into the card maintenance identification model, and outputting a card maintenance identification result representing whether the card maintenance behavior exists or not.
In a third aspect, an embodiment of the present application provides a training apparatus for a card raising recognition model, including:
the acquisition module is used for acquiring a training sample set, wherein the training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card-raising behavior label data corresponding to the business data of the historical agent;
the determining module is used for determining abnormal value data which accord with a preset abnormal value judgment rule from the business data of the first historical agent, and the business data of the first historical agent represents the business data of the historical agent corresponding to the tag data of the card raising behavior with the card raising behavior;
the output module is used for inputting the abnormal value data into a preset linear regression algorithm for risk estimation and outputting target abnormal value data with a high risk value;
the output module is also used for inputting the target abnormal value data into a quartile algorithm of a target model to carry out model threshold calculation and outputting a plurality of model thresholds;
the determining module is further used for determining the plurality of model thresholds as parameters of a preset model to obtain an original model;
the output module is also used for inputting the business data of the historical agent into the original model and outputting result data representing whether the card-holding behavior exists or not;
and the adjusting module is used for adjusting the parameters of the original model on the premise that the hit rate of the result data and the card-raising behavior tag data is consistent to be less than a preset threshold value, and returning to input the business data of the historical agent into the original model until the hit rate is not less than the preset threshold value.
Further, in one embodiment, the business data of the historical brokers comprises:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data.
Further, in one embodiment, the preset outlier determination rule includes a preset 95-quantile algorithm;
a determination module specifically configured to:
and inputting the service data of the historical agent corresponding to the tag data of the card-raising behavior representing the existence of the card-raising behavior into a preset 95-quantile algorithm to determine abnormal value data, and outputting the abnormal value data.
Further, in one embodiment, the predetermined model is selected as a decision tree model.
In a fourth aspect, an embodiment of the present application provides an apparatus for calling a card maintenance identification model to identify a card maintenance behavior, where the card maintenance identification model is obtained through training by an apparatus according to the claims, and the apparatus includes:
the acquisition module is used for acquiring the service data of the target agent, and the service data of the target agent comprises at least one of the following items: the behavior data of the target agent, the user behavior data of the target agent and the charge-out data of the target agent;
and the output module is used for inputting the business data of the target agent into the card-keeping identification model and outputting a card-keeping identification result representing whether card-keeping behaviors exist or not.
In a fifth aspect, an embodiment of the present application provides a training device for a card raising recognition model, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program implementing the above-mentioned method when executed by the processor.
In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, on which an implementation program for information transfer is stored, and when the implementation program is executed by a processor, the method is implemented.
According to the training method, device, equipment and storage medium of the card-raising identification model, 95-quantile, quartile and linear regression algorithms are adopted to process business data of a historical agent corresponding to card-raising behavior tag data representing card-raising behavior, parameters of an original model are obtained, the parameters of the original model are enabled to be more consistent with the business data characteristics of the historical agent corresponding to the card-raising behavior tag data representing card-raising behavior, training of the card-raising identification model is conducted on the basis of all business data of the historical agent on the basis of the original model, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of result data representing whether card-raising behavior exists and the card-raising behavior tag data, the card-raising identification model meeting expected requirements is trained finally, and whether card-raising behavior exists in business data of a target agent or not can be accurately identified on the basis of the card-raising identification model.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments of the present application will be briefly described below, and for those skilled in the art, other drawings may be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart diagram illustrating a method for training a card-based recognition model according to an embodiment of the present application;
FIG. 2 is a schematic flow chart diagram illustrating a method for identifying card maintenance behavior by invoking a card maintenance identification model according to an embodiment of the present application;
FIG. 3 is a schematic structural diagram of a training apparatus for a card feeding recognition model according to an embodiment of the present application;
FIG. 4 is a schematic structural diagram of an apparatus for recognizing a card-holding behavior by calling a card-holding recognition model according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a training device for a card feeding recognition model according to an embodiment of the present application.
Detailed Description
Features and exemplary embodiments of various aspects of the present application will be described in detail below, and in order to make objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of and not restrictive on the broad application. It will be apparent to one skilled in the art that the present application may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present application by illustrating examples thereof.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising 8230; \8230;" comprises 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
Aiming at the behavior of the agent for violating the card care arbitrage, the telecom operator also has some methods for identifying the card care behavior in the daily operation process, but in view of the fact that the agent can continuously change the behavior mode to avoid being monitored by the operator, a fixed card care identification model is difficult to monitor and accurately identify the behavior for a long time, and the existing card care auditing model cannot be dynamically and continuously adjusted and optimized automatically according to the change of the behavior, so that the identification accuracy of the existing method for identifying the card care behavior is low.
In order to solve the problems of the prior art, the embodiment of the application provides a training method, a training device, equipment and a storage medium for a card raising recognition model. According to the method, the business data of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain the parameters of the original model, so that the parameters of the original model are more consistent with the business data characteristics of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior, training of a card-raising identification model is carried out on the basis of the original model based on all the business data of the historical agents, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of the result data representing the card-raising behavior and the tag data of the card-raising behavior, the card-raising identification model meeting the expected requirement is trained finally, and whether the card-raising behavior exists in the business data of the target agents or not can be accurately identified based on the card-raising identification model. First, a training method of the card-raising recognition model provided in the embodiment of the present application is described below.
Fig. 1 shows a flowchart of a training method of a card-feeding recognition model according to an embodiment of the present application. As shown in fig. 1, the method may include the steps of:
and S110, acquiring a training sample set.
The training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card-raising behavior label data corresponding to the business data of the historical agent. The service data of the historical agent can be obtained from the agent big data platform, and the card maintenance behavior tag data represents whether card maintenance behavior exists or not and can be obtained from the operator big data platform.
In one embodiment, the historical agent's business data includes:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data, wherein:
historical agent behavior data, including: the method comprises the steps of enabling an agent to account for the concentrated development amount in the current month, enabling the agent to develop a unified channel access network of a customer and use the same International Mobile Equipment Identity (IMEI) number, enabling the agent to develop the same channel access network of the customer and use the same IMEI number, enabling the agent to develop the same customer name consistency number, enabling the agent to develop the same customer number and enabling the agent to account for the high-frequency similar services in the current month.
The agent develops a customer uniform channel to access the network, and uses the number of the same International Mobile Equipment Identity (IMEI), and the number can be marked as I _ val; the agent develops a customer using the same IMEI customer name consistent number, which may be denoted as M _ val.
User behavior data for a historical agent, comprising: the agent develops the number of the clients communicating only in the same base station, the agent develops the proportion of the number of all call forwarding telephone bills of the clients to the total telephone bill amount, the agent develops the communication cost of the clients, the agent develops the communication times of the clients (including short message receiving and sending, calling and called), and the agent develops the charging flow of the clients.
The agent develops the number of times of communication of the client (including short message receiving and sending, calling and called), and the value of the number can be recorded as C _ val; the agent develops the customer charging flow, and the value can be recorded as E _ val, unit million (M); the agent develops the communication cost of the client, and the value can be recorded as F _ val, unit (element); the agent develops the proportion of the number of all call ticket to the total ticket amount of the client, and the proportion can be recorded as H _ val.
The charge-out data of the historical agent comprises: the method comprises the steps that an agent development client generates communication only in one base station cell and accesses the network through the same channel, the agent development client generates charge for paying out the communication only in one base station cell, the agent development client generates communication times only in one base station cell, the agent development client generates communication only in one base station cell and uses the number of the same international mobile equipment identification code, the agent development client generates communication times which are all call forwarding lists in the voice call lists of the client, the agent development client generates charge for paying out the call forwarding lists, the agent development client uses charge corresponding to the same IMEI number, and the agent development client uses charge corresponding to the same client name.
The agent develops a charge for the customer to generate communication in only one base station cell, and the charge can be marked as D _ val.
And S120, determining abnormal value data which accords with a preset abnormal value judgment rule from the service data of the first historical agent.
And the business data of the first historical agent represents the business data of the historical agent corresponding to the tag data of the card-holding behavior with the card-holding behavior.
The card-raising behavior tag data with the card-raising behavior can be determined based on the number of times that the agent develops the customer communication, for example, the business data of the historical agent with the number of times that the agent develops the customer communication being less than 10 can be regarded as the card-raising behavior tag data corresponding to the business data as the card-raising behavior.
In one embodiment, the predetermined outlier determination rule comprises a predetermined 95-quantile algorithm; s120 may include:
and inputting the service data of the first historical agent into a preset 95-bit algorithm to determine abnormal value data, and outputting the abnormal value data.
The 95 quantile algorithm, namely 95th percentile, refers to the number of more than 95% of the given data set, and more than 5% of the data are determined as abnormal value data, so that the abnormal value data are prevented from influencing the model precision.
And S130, inputting the abnormal value data into a preset linear regression algorithm for risk estimation, and outputting target abnormal value data with a high risk value.
Linear Regression (Linear Regression) is a Regression analysis that models the relationship between one or more independent and dependent variables using a least squares function called the Linear Regression equation. A regression line formed after modeling represents a middle value of the current group, data on the upper side of the regression line represents that the commission payment corresponding to the sample is larger than the average value under the same condition, the more exceeding is, the higher the cost is, the higher the risk is, and the target abnormal value data with the high risk value can be selected from the data on the upper side of the regression line.
In one embodiment, the risk estimation for the user behavior data of the historical agent meets any one of the following conditions, and it can be determined that the card-raising behavior is suspected to exist:
1. i _ val is greater than or equal to 5;
2. 2-woven I \ u val-woven 5 and C _ val <5;
3. m _ val is greater than or equal to 3;
4. c _ val <5 and D _ val is greater than or equal to 1;
6. 2-woven I _val-woven 5 with D _ val greater than or equal to 1;
7. c _ val <5 and E _ val <50m.
And (3) aiming at the risk estimation of the charge data of the historical agent, if any one of the following conditions is met, the suspected card-raising behavior can be judged to be:
1. i _ val is greater than or equal to 10;
2. i _ val is greater than 2 and less than 10 and C _ val is less than 5;
3. d _ val is smaller than 1-element client;
4. the number of times that the agent development client generates communication in only one base station cell is less than 5;
5. the charge of all call forwarding telephone bills in the agent development customer voice telephone bills is less than 1 yuan;
6. the communication times of all call forwarding telephone bills in the agent development client voice telephone bills are less than 5;
7. the agent development customers only generate communication in a base station cell and the number of the network access numbers of the same channel is more than 2 and less than 10;
8. the I _ val is larger than 2 and smaller than 10, and the charge-out cost corresponding to the same IMEI number used by the agent development client is smaller than 1 yuan;
9. the agent develops the customer to enter the network in the same channel, and the number of the consistent customer names is greater than or equal to 10.
And S140, inputting the target abnormal value data into a quartile algorithm of a target model to calculate a model threshold value, and outputting a plurality of model threshold values.
S150, determining a plurality of model thresholds as parameters of a preset model to obtain an original model.
In one embodiment, the default model may be selected as a decision tree model.
And S160, inputting the service data of the historical agent into the original model, and outputting result data representing whether the card-holding behavior exists.
S170, on the premise that the hit rate of the result data and the card raising behavior tag data is consistent and is smaller than a preset threshold value, adjusting parameters of the original model, and returning to input the business data of the historical agent into the original model until the hit rate is not smaller than the preset threshold value.
In one embodiment, the card keeping identification model can be deployed in an operator system, a calculation point is set to identify whether card keeping behaviors exist or not in time after the user accounts out and before commission is issued, and interception is carried out before the commission is issued, so that risk advance control is realized.
According to the method, the business data of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain the parameters of the original model, so that the parameters of the original model are more consistent with the business data characteristics of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior, training of a card-raising identification model is carried out on the basis of the original model based on all the business data of the historical agents, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of the result data representing the card-raising behavior and the tag data of the card-raising behavior, the card-raising identification model meeting the expected requirement is trained finally, and whether the card-raising behavior exists in the business data of the target agents or not can be accurately identified based on the card-raising identification model.
The training process of the card-raising recognition model is introduced above, and the application process of the card-raising recognition model is introduced below.
The embodiment of the application provides a method for calling a card maintenance identification model to identify a card maintenance behavior, and fig. 2 is a flow diagram illustrating the method for calling the card maintenance identification model to identify the card maintenance behavior. As shown in fig. 2, the method may include the steps of:
s210, acquiring the service data of the target agent.
The traffic data of the target agent includes at least one of: the behavior data of the target agent, the user behavior data of the target agent and the charge-out data of the target agent;
s220, inputting the business data of the target agent into the card maintenance identification model, and outputting a card maintenance identification result representing whether the card maintenance behavior exists or not.
According to the method, the business data of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain the parameters of the original model, so that the parameters of the original model are more consistent with the business data characteristics of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior, training of a card-raising identification model is carried out on the basis of the original model based on all the business data of the historical agents, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of the result data representing the card-raising behavior and the tag data of the card-raising behavior, the card-raising identification model meeting the expected requirement is trained finally, and whether the card-raising behavior exists in the business data of the target agents or not can be accurately identified based on the card-raising identification model.
Fig. 1-2 illustrate a training method of a card-raising recognition model, and the following describes an apparatus provided by an embodiment of the present application with reference to fig. 3-5.
Fig. 3 shows a schematic structural diagram of a training device for a card-raising recognition model according to an embodiment of the present application, and each module in the device shown in fig. 3 has a function of implementing each step in fig. 1, and can achieve its corresponding technical effect. As shown in fig. 3, the apparatus may include:
an obtaining module 310 is configured to obtain a training sample set.
The training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card feeding behavior label data corresponding to the business data of the historical agent;
the determining module 320 is configured to determine, from the service data of the first history agent, abnormal value data that meets a preset abnormal value determination rule.
The service data of the first historical agent represents the service data of the historical agent corresponding to the tag data of the card maintenance behavior with the card maintenance behavior;
and the output module 330 is configured to input the abnormal value data into a preset linear regression algorithm to perform risk estimation, and output target abnormal value data with a high risk value.
The output module 330 is further configured to input the target abnormal value data into a quartile algorithm of the target model to perform model threshold calculation, and output a plurality of model thresholds.
The determining module 320 is further configured to determine a plurality of model thresholds as parameters of a preset model, so as to obtain an original model.
The output module 330 is further configured to input the service data of the historical agent into the original model, and output result data representing whether the card-holding behavior exists.
And the adjusting module 340 is configured to adjust parameters of the original model on the premise that the hit rate of the result data and the card maintenance behavior tag data are consistent is smaller than a preset threshold, and return to input the service data of the historical agent into the original model until the hit rate is not smaller than the preset threshold.
According to the method, the business data of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain the parameters of the original model, so that the parameters of the original model are more consistent with the business data characteristics of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior, training of a card-raising identification model is carried out on the basis of the original model based on all the business data of the historical agents, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of the result data representing the card-raising behavior and the tag data of the card-raising behavior, the card-raising identification model meeting the expected requirement is trained finally, and whether the card-raising behavior exists in the business data of the target agents or not can be accurately identified based on the card-raising identification model.
In one embodiment, the business data of the history agent comprises:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data.
In one embodiment, the predetermined outlier determination rule comprises a predetermined 95-quantile algorithm;
the determining module 320 is specifically configured to:
and inputting the service data of the first historical agent into a preset 95-bit algorithm to determine abnormal value data, and outputting the abnormal value data.
In one embodiment, the predetermined model is selected as a decision tree model.
According to the method, the service data of the historical agents corresponding to the tag data of the card raising behaviors with the card raising behaviors are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain parameters of an original model, so that the parameters of the original model better accord with the service data characteristics of the historical agents corresponding to the tag data of the card raising behaviors with the card raising behaviors, training of a card raising recognition model is carried out on the basis of the original model on the basis of all the service data of the historical agents, whether the parameters of the original model need to be adjusted is judged according to the hit rate of the consistency of the result data representing whether the card raising behaviors exist and the tag data of the card raising behaviors, the card raising recognition model meeting the expected requirements is finally trained, and whether the card raising behaviors exist in the service data of target agents can be accurately recognized on the basis of the card raising recognition model.
The training device of the card-raising recognition model is introduced above, and the application device of the card-raising recognition model is introduced below.
Fig. 4 is a schematic structural diagram illustrating an apparatus for calling a card maintenance identification model to identify a card maintenance behavior according to an embodiment of the present application, where each module in the apparatus shown in fig. 4 has a function of implementing each step in fig. 2, and can achieve its corresponding technical effect. As shown in fig. 4, the apparatus may include:
an obtaining module 410, configured to obtain service data of the target agent.
The service data of the target agent comprises at least one of the following items: behavior data of the target agent, user behavior data of the target agent and charge-out data of the target agent;
and the output module 420 is used for inputting the service data of the target agent into the card maintenance identification model and outputting a card maintenance identification result representing whether the card maintenance behavior exists or not.
According to the method, the business data of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior are processed by adopting a 95-quantile, quartile and linear regression algorithm to obtain the parameters of the original model, so that the parameters of the original model are more consistent with the business data characteristics of the historical agents corresponding to the tag data of the card-raising behavior representing the card-raising behavior, training of a card-raising identification model is carried out on the basis of the original model based on all the business data of the historical agents, whether the parameters of the original model need to be adjusted or not is judged according to the hit rate of consistency of the result data representing the card-raising behavior and the tag data of the card-raising behavior, the card-raising identification model meeting the expected requirement is trained finally, and whether the card-raising behavior exists in the business data of the target agents or not can be accurately identified based on the card-raising identification model.
Fig. 5 is a schematic structural diagram illustrating a training device for a card feeding recognition model according to an embodiment of the present application. As shown in fig. 5, the apparatus may include a processor 501 and a memory 502 storing computer program instructions.
Specifically, the processor 501 may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement the embodiments of the present Application.
In one example, the Memory 502 may be a Read Only Memory (ROM). In one example, the ROM can be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically Erasable PROM (EEPROM), electrically Alterable ROM (EAROM), or flash memory, or a combination of two or more of these.
The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement the method in the embodiment shown in fig. 1-2, and achieve the corresponding technical effects achieved by the embodiment shown in fig. 1-2 executing the method, which are not described herein again for brevity.
In one example, the training device of the card-based recognition model may also include a communication interface 503 and a bus 510. As shown in fig. 5, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 to complete communication therebetween.
The communication interface 503 is mainly used for implementing communication between modules, apparatuses, units and/or devices in the embodiments of the present application.
The training device of the card feeding identification model can execute the training method of the card feeding identification model in the embodiment of the application, so that the corresponding technical effects of the training method of the card feeding identification model described in the figures 1-2 are realized.
In addition, in combination with the training method of the card-raising recognition model in the foregoing embodiment, the embodiment of the present application may provide a computer storage medium to implement the method. The computer storage medium having computer program instructions stored thereon; the computer program instructions, when executed by the processor, implement any of the above embodiments of the training method for a card-based recognition model.
It is to be understood that the present application is not limited to the particular arrangements and instrumentalities described above and shown in the attached drawings. A detailed description of known methods is omitted herein for the sake of brevity. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications, and additions or change the order between the steps after comprehending the spirit of the present application.
The functional blocks shown in the above-described structural block diagrams may be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic Circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the present application are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link. A "machine-readable medium" may include any medium that can store or transfer information. Examples of a machine-readable medium include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio Frequency (RF) links, and so forth. The code segments may be downloaded via computer networks such as the internet, intranets, etc.
It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, may be performed in an order different from the order in the embodiments, or may be performed simultaneously.
Aspects of the present application are described above with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions/acts specified in the flowchart and/or block diagram block or blocks. Such a processor may be, but is not limited to, a general purpose processor, a special purpose processor, an application specific processor, or a field programmable logic circuit. It will also be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware for performing the specified functions or acts, or combinations of special purpose hardware and computer instructions.
As described above, only the specific embodiments of the present application are provided, and it can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system, the module and the unit described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. It should be understood that the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the present application, and these modifications or substitutions should be covered within the scope of the present application.
Claims (12)
1. A training method of a card raising recognition model is characterized by comprising the following steps:
acquiring a training sample set, wherein the training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card-raising behavior label data corresponding to the business data of the historical agent;
determining abnormal value data which accords with a preset abnormal value judgment rule from the business data of a first historical agent, wherein the business data of the first historical agent represents the business data of the historical agent corresponding to the tag data of the card-raising behavior with the card-raising behavior;
inputting the abnormal value data into a preset linear regression algorithm for risk estimation, and outputting target abnormal value data of a high risk value;
inputting the target abnormal value data into a quartile algorithm of a target model to perform model threshold calculation, and outputting a plurality of model thresholds;
determining the plurality of model thresholds as parameters of a preset model to obtain an original model;
inputting the business data of the historical agent into the original model, and outputting result data representing whether card maintenance behaviors exist or not;
and on the premise that the hit rate of the result data and the card-raising behavior tag data is consistent with a preset threshold value, adjusting the parameters of the original model, and returning to input the business data of the historical agent into the original model until the hit rate is not less than the preset threshold value.
2. The method for training the card-raising recognition model according to claim 1, wherein the business data of the historical agent comprises:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data.
3. The training method of the card-raising recognition model according to claim 1, wherein the preset outlier determination rule comprises a preset 95-decitex algorithm;
the method for determining abnormal value data meeting a preset abnormal value judgment rule from the business data of the first historical agent comprises the following steps:
and inputting the service data of the first historical agent into a preset 95-bit algorithm to determine abnormal value data, and outputting the abnormal value data.
4. The method for training a card feeding recognition model according to claim 1, wherein the predetermined model is selected as a decision tree model.
5. A method for calling a card-raising recognition model to recognize a card-raising behavior, wherein the card-raising recognition model is trained by the method of claim 1, the method comprising:
acquiring service data of a target agent, wherein the service data of the target agent comprises at least one of the following items: behavior data of the target agent, user behavior data of the target agent and charge-out data of the target agent;
and inputting the service data of the target agent into the card maintenance identification model, and outputting a card maintenance identification result representing whether card maintenance behaviors exist or not.
6. The utility model provides a training device of card raising recognition model which characterized in that includes:
the acquisition module is used for acquiring a training sample set, wherein the training sample set comprises a plurality of training samples, and each training sample comprises business data of a historical agent and card-raising behavior label data corresponding to the business data of the historical agent;
the determining module is used for determining abnormal value data which accords with a preset abnormal value judging rule from service data of a first historical agent, and the service data of the first historical agent represents service data of the historical agent corresponding to the tag data of the card maintenance behavior with the card maintenance behavior;
the output module is used for inputting the abnormal value data into a preset linear regression algorithm for risk estimation and outputting target abnormal value data of a high risk value;
the output module is also used for inputting the target abnormal value data into a quartile algorithm of a target model to carry out model threshold calculation and outputting a plurality of model thresholds;
the determining module is further configured to determine the plurality of model thresholds as parameters of a preset model to obtain an original model;
the output module is further used for inputting the service data of the historical agent into the original model and outputting result data representing whether card maintenance behaviors exist or not;
and the adjusting module is used for adjusting the parameters of the original model on the premise that the hit rate of the result data and the card-raising behavior tag data is consistent is less than a preset threshold value, and returning to input the business data of the historical agent into the original model until the hit rate is not less than the preset threshold value.
7. The apparatus for training card-keeping recognition model as claimed in claim 6, wherein the business data of the historical agent comprises:
historical agent's behavioral data, historical agent's user behavioral data and historical agent's charge-out data.
8. The training device of the card-raising recognition model as claimed in claim 6, wherein the preset abnormal value judgment rule comprises a preset 95-quantile algorithm;
the determining module is specifically configured to:
and inputting the business data of the historical agent corresponding to the tag data of the card-raising behavior representing the existence of the card-raising behavior into a preset 95-quantile algorithm to determine abnormal value data, and outputting the abnormal value data.
9. The training device for the card-raising recognition model as claimed in claim 6, wherein the predetermined model is selected as a decision tree model.
10. An apparatus for invoking a card maintenance identification model to identify card maintenance behavior, wherein the card maintenance identification model is trained by the apparatus of claim 6, the apparatus comprising:
an obtaining module, configured to obtain service data of a target agent, where the service data of the target agent includes at least one of the following: behavior data of the target agent, user behavior data of the target agent and charge-out data of the target agent;
and the output module is used for inputting the service data of the target agent into the card maintenance identification model and outputting a card maintenance identification result representing whether card maintenance behaviors exist or not.
11. The utility model provides a training equipment of card raising recognition model which characterized in that includes: memory, processor and computer program stored on the memory and executable on the processor, which when executed by the processor implements the method of any one of claims 1 to 5.
12. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a program for implementing information transfer, which program, when executed by a processor, implements the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110391185.3A CN115203641A (en) | 2021-04-12 | 2021-04-12 | Training method, device, equipment and storage medium for card-raising recognition model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110391185.3A CN115203641A (en) | 2021-04-12 | 2021-04-12 | Training method, device, equipment and storage medium for card-raising recognition model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115203641A true CN115203641A (en) | 2022-10-18 |
Family
ID=83570427
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110391185.3A Pending CN115203641A (en) | 2021-04-12 | 2021-04-12 | Training method, device, equipment and storage medium for card-raising recognition model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115203641A (en) |
-
2021
- 2021-04-12 CN CN202110391185.3A patent/CN115203641A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104778583A (en) | Security verification method for mobile payment | |
| CN104794616A (en) | Safety verification method for mobile phone payment | |
| CN108810290B (en) | Method and system for recognizing fraud calls | |
| CN110347566B (en) | Method and device for evaluating effectiveness of registration wind control model | |
| CN113011899A (en) | Enterprise credit evaluation method, device, equipment and computer storage medium | |
| CN109063965A (en) | A kind of network payment methods of risk assessment, device and server | |
| CN111091408A (en) | User identification model creating method and device and identification method and device | |
| CN114549136A (en) | Method, device and equipment for evaluating credit of supply chain user and computer storage medium | |
| CN117252429A (en) | Risk user identification method and device, storage medium and electronic equipment | |
| CN115203641A (en) | Training method, device, equipment and storage medium for card-raising recognition model | |
| CN111353015B (en) | Crowd-sourced question recommendation method, device, equipment and storage medium | |
| CN108874619B (en) | Information monitoring method, storage medium and server | |
| CN117252678A (en) | Agricultural loan request processing method, device, equipment and readable storage medium | |
| CN113051601A (en) | Sensitive data identification method, device, equipment and medium | |
| CN111340062A (en) | Mapping relation determining method and device | |
| CN113052422A (en) | Wind control model training method and user credit evaluation method | |
| CN114417830A (en) | Risk evaluation method, device, equipment and computer readable storage medium | |
| CN109873836A (en) | A kind of methods of risk assessment and device of data | |
| CN114677139A (en) | Method and device for determining loan amount, equipment, product and readable storage medium | |
| CN112819613B (en) | Loan information processing method, device and storage medium | |
| CN112488825B (en) | Object transaction method and device based on blockchain | |
| CN110570301B (en) | Risk identification method, device, equipment and medium | |
| CN115130577A (en) | Method and device for identifying fraudulent number and electronic equipment | |
| CN109993181B (en) | Abnormal behavior pattern recognition method, device, equipment and medium | |
| CN112839335A (en) | Number identification method, device, equipment and medium |
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 |