A method and apparatus for determining a user security level based on a user service change

Abstract

The application provides a user security level determination method and device based on user service change, which comprises the following steps: inputting user feature data of candidate users in a candidate user group into a trained user screening model to obtain a first user score for representing change probability of specific business service of the user; determining a specific user in the candidate user group according to the first user score; determining a second user score according to the first user score and the user feature data of the specific user; and determining a security level of the specific user based on the first user score and the second user score. The application determines possible change conditions of services that the user can enjoy from the condition that the user enjoys changed services, effectively quantifies change condition information of the user's demand for internet services from the dimension of change condition information of the specific business service from nothing to something, and can accurately determine the security level of the specific user and change condition information thereof.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of computer information processing, and in particular to a user security level determination method and device based on user service changes. BACKGROUND

[0002] In internet-based application technology, service exchange is often required between different participants. The service referred to here refers to any material, information, money, time, etc. that can be utilized. Information services include computing data and various types of data services. Data services include various specialized data in various fields. In the process of providing user services, the service configuration rights of the user often need to be authenticated, and different services are allocated to different users. The service configuration rights refer to an authentication of whether the user has the right to obtain the service, which can be authenticated by a specific management organization or authenticated by the party owning the service.

[0003] For an institution that provides user services, a comprehensive and in-depth analysis of the user can help to provide better services for the user. However, in many cases, the user information submitted by the user is not sufficient, especially for users who have just landed registration. The user service institution can only obtain simple user information and cannot know the security situation or risk situation of the user. In this case, integrating scattered user data has become an important trend.

[0004] For individual users, due to the individual differences of individual users, internet service companies are almost unable to predict the plans and times of individual users' internet service demands in advance, and the amount of internet services that can be used or utilized at different time periods is also changing. Therefore, how to effectively quantify the demand changes of individual users for internet services that change over time, how to determine the change information of the user security level, how to provide more reasonable and better internet services for different users, and how to effectively optimize the internet service providing process are difficult problems faced by today's internet service companies.

[0005] Therefore, it is necessary to provide a user security level determination method based on user service changes. SUMMARY

[0006] In order to solve the technical problems in the prior art that the demand changes of each user for internet services that change over time cannot be accurately quantified, how to determine the change information of the user security level, how to provide more reasonable and better internet services for different users, and how to optimize the providing process of internet services.

[0007] The first aspect of the present application provides a user security level determination method based on user service changes, comprising: obtaining sample feature data of a sample user and change information of whether the sample user holds a specific service; adding a training label to the sample user according to the change information of whether the sample user holds the specific service; training a user screening model according to the sample feature data of the sample user and the training label; inputting user feature data of a candidate user in a candidate user group into the trained user screening model to obtain a first user score representing a change probability of a specific service of the user, and determining a specific user in the candidate user group according to the first user score; inputting the first user score and user feature data of the specific user into a user security identification model to obtain a second user score; and determining a security level of the specific user based on the first user score and the second user score.

[0008] According to an optional implementation, determining the security level of the specific user based on the first user score and the second user score comprises: determining a coefficient determination curve according to the first user score and the security level of a historical user, the coefficient determination curve being used to determine a first calculation coefficient of the security level of the user; obtaining the first calculation coefficient a of the specific user according to the first user score of the specific user and the coefficient determination curve; determining a second calculation coefficient b corresponding to the second user score; using a product of the first calculation coefficient a and the second calculation coefficient b as an adjustment coefficient of the specific user; determining a pending security level of the specific user according to the first user score and the second user score; and adjusting the pending security level according to the adjustment coefficient to obtain the security level of the specific user.

[0009] According to an optional implementation, determining the pending security level of the specific user according to the first user score and the second user score comprises: determining a corresponding pending security level partition from preset security level partitions according to the first user score and the second user score of the specific user; and determining the pending security level of the specific user according to the pending security level partition.

[0010] According to an optional implementation, adjusting the pending security level according to the adjustment coefficient to obtain the security level of the specific user comprises: determining a corresponding security level partition from the pending security level partition according to the adjustment coefficient; and determining the security level of the specific user according to the security level partition.

[0011] According to an optional implementation, determining the security level of a specific user based on the security level segmentation interval includes: dividing the security level segmentation interval into at least one security level score judgment interval according to a preset security level score threshold range; calculating the sum of each security level score in each security level score judgment interval; and determining the security level of the specific user based on the preset security level score threshold range to which the security level score judgment interval with the largest sum of security level scores belongs.

[0012] According to an optional implementation, determining a specific user in the candidate user group based on the first user score includes: when the first user score of any candidate user is within a set threshold range, determining the candidate user as a specific user, wherein the specific user is a candidate user identified as ranging from not holding a specific business service to holding a specific business service.

[0013] According to an optional implementation, adding training labels to the sample users based on information about changes in whether the sample users hold specific business services includes: if the sample user changes from never holding a specific business service to holding a specific business service within a specific time period, then adding a first label to the sample user; if the sample user does not hold a specific business service within a specific time period, then adding a second label to the sample user; the specific business service includes issuing discount cards, coupons, and / or vouchers that can be used within a specified time.

[0014] Furthermore, a second aspect of the present invention provides a user security level determination device based on changes in user services, comprising: a data acquisition module for acquiring sample feature data of sample users and information on changes in whether sample users hold specific business services; a model training module for adding training labels to the sample users based on the information on changes in whether sample users hold specific business services; and training a user screening model based on the sample feature data of the sample users and the training labels; a first processing module for inputting user feature data of candidate users in a candidate user group into the trained user screening model to obtain a first user score characterizing the probability of changes in a user's specific business services, and determining a specific user in the candidate user group based on the first user score; inputting the first user score and user feature data of the specific user into a user security identification model to obtain a second user score; and a second processing module for determining the security level of the specific user based on the first user score and the second user score.

[0015] Furthermore, the present invention also provides a computer device including a processor and a memory, the memory being used to store a computer executable program, wherein when the computer program is executed by the processor, the processor executes the user security level determination method based on user service changes as described in the present invention.

[0016] In addition, the present invention also provides a computer program product storing a computer executable program, which, when executed, implements the user security level determination method based on user service changes as described in the present invention.

[0017] Beneficial effects

[0018] Compared with existing technologies, this invention adds training labels to sample users by using the feature of "information on changes in whether sample users hold specific business services". A user screening model is trained based on the sample user feature data and the training labels. User feature data of candidate users in the candidate user group are input into the trained user screening model to obtain a first user score representing the probability of changes in a user's specific business services. Specific users in the candidate user group are determined based on the first user score. The first user score and user feature data of the specific user are input into a user security identification model to obtain a second user score. Based on the first user score and the second user score, starting from the user situation where the services enjoyed have changed, the possible changes in the services a user can enjoy are determined. This effectively quantifies the changes in users' demand for internet services from the dimension of information on changes in specific business services from non-existent to available, and can accurately determine the security level of the specific user and its changes. Furthermore, it can provide more reasonable and higher-quality internet services to different users based on their security levels.

[0019] Furthermore, by determining the curve based on the first user rating and security level fitting coefficient of the historical users, a more accurate coefficient determination curve can be obtained; by calculating the first calculation coefficient based on the determined coefficient determination curve, the calculated first user rating can be corrected, and a more accurate first user rating can be obtained.

[0020] Furthermore, by using the product of the first calculation coefficient a and the second calculation coefficient b as the adjustment coefficient for a specific user, the security level of the specific user can be re-determined based on the changes in each user's demand for Internet services over time. This allows for a more accurate determination of the security level of the specific user, further optimizes the user security level determination method, and further optimizes the Internet service provision process. Attached Figure Description

[0021] To make the technical problems solved by the present invention, the technical means adopted, and the technical effects achieved clearer, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, it should be noted that the drawings described below are merely drawings of exemplary embodiments of the present invention. Those skilled in the art can obtain drawings of other embodiments based on these drawings without any creative effort.

[0022] Figure 1 This is a flowchart illustrating an example of the user security level determination method based on changes in user services according to the present invention.

[0023] Figure 2 This is a flowchart of another example of the user security level determination method based on changes in user services according to the present invention.

[0024] Figure 3 This is a flowchart illustrating yet another example of the user security level determination method based on changes in user services according to the present invention.

[0025] Figure 4 This is a schematic structural block diagram of an example of the user security level determination device based on changes in user services according to the present invention.

[0026] Figure 5 This is a schematic structural block diagram of another example of the user security level determination device based on changes in user services according to the present invention.

[0027] Figure 6 This is a structural block diagram of an exemplary embodiment of a computer device according to the present invention.

[0028] Figure 7 This is a structural block diagram of an exemplary embodiment of a computer program product according to the present invention. Detailed Implementation

[0029] Exemplary embodiments of the invention will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limiting the invention to the embodiments set forth herein. Rather, these exemplary embodiments are provided to make the invention more comprehensive and complete, and to facilitate a full communication of the inventive concept to those skilled in the art. The same reference numerals in the drawings denote the same or similar elements, components, or parts, and therefore repeated descriptions of them will be omitted.

[0030] Subject to the technical concept of this invention, the features, structures, characteristics or other details described in a particular embodiment may be combined in one or more other embodiments in a suitable manner.

[0031] In the description of specific embodiments, the features, structures, characteristics, or other details described in this invention are intended to enable those skilled in the art to fully understand the embodiments. However, it is not excluded that those skilled in the art can practice the technical solutions of this invention without one or more of the specific features, structures, characteristics, or other details.

[0032] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0033] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0034] It should be understood that although terms such as "first," "second," "third," etc., indicating designations, may be used herein to describe various devices, elements, components, or parts, this should not be limited by these terms. These terms are used to distinguish one from another. For example, a first device may also be referred to as a second device without departing from the essential technical solution of this invention.

[0035] The terms “and / or” or “and / or” include any one or more of the listed items in relation to each other.

[0036] In view of the above problems, the present invention provides a method for determining user security levels based on changes in user services. This method adds training labels to sample users using the feature of "information on changes in whether a sample user holds specific business services," and trains a user screening model based on the sample user feature data and the training labels. The user feature data of candidate users in a candidate user group are input into the trained user screening model to obtain a first user score representing the probability of changes in a user's specific business services. Based on the first user score, specific users in the candidate user group are determined. The first user score and user feature data of the specific user are input into a user security identification model to obtain a second user score. Based on the first user score and the second user score, starting from the user situation where the services enjoyed have changed, the possible changes in the services a user can enjoy are determined. This effectively quantifies the changes in user demand for internet services from the dimension of information on changes in specific business services from non-existent to available, and can accurately determine the security level of the specific user and its changes. Furthermore, it can provide more reasonable and higher-quality internet services to different users based on their security levels. Compared to existing models that only consider static user scenarios, this solution performs targeted analysis for users who may experience specific situations (service changes), improving the refined management of different users and making the services provided more tailored to user needs.

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. For convenience, this invention uses Internet services as an example to illustrate the implementation of the user security level determination method; however, those skilled in the art should understand that this invention can also be used to redetermine quotas for other services.

[0038] Example 1

[0039] Below, we will refer to Figures 1 to 3 This invention describes an embodiment of the user security level determination method based on changes in user services.

[0040] Figure 1 This is a flowchart illustrating an example of the user security level determination method based on changes in user services according to the present invention.

[0041] like Figure 1 As shown, the method for determining the user's security level includes the following steps.

[0042] Step S101: Obtain sample feature data of sample users and information on changes in whether sample users hold specific business services.

[0043] Step S102: Add training labels to the sample users based on the information on changes in whether the sample users hold specific business services; and train a user screening model based on the sample feature data of the sample users and the training labels.

[0044] Step S103: Input the user feature data of the candidate users in the candidate user group into the trained user screening model to obtain a first user score that represents the probability of change of a user’s specific business service, and determine the specific user in the candidate user group based on the first user score.

[0045] Step S104: Input the first user score and user feature data of the specific user into the user security identification model to obtain the second user score; determine the security level of the specific user based on the first user score and the second user score.

[0046] It should be noted that, in this invention, internet services include services such as shopping, transportation, maps, food delivery, and shared bicycles, provided by user devices upon request to an internet service platform. For example, these internet services include usage services, allocation services, fundraising services, guarantee services or mutual aid services, group buying, and transportation services, etc. Here, "service" refers to any usable material, information, or time; information data includes computational data and various types of service data. Service data includes various specialized data in various fields. The user security level determination method of this invention is particularly applicable to these internet services. The following uses data related to internet service allocation as an example to specifically illustrate the method of this invention.

[0047] First, in step S101, sample characteristic data of sample users and information on changes in whether sample users hold specific business services are obtained.

[0048] For example, sample characteristic data of sample users and information on changes in whether sample users hold specific business services can be obtained from historical data on internet service usage and allocation.

[0049] Specifically, the sample user's sample feature data includes publicly available user information on the internet or user information obtained with user authorization, as well as user identifiers. These user identifiers are, for example, 112**0, or user identifiers represented by letters and / or numbers. The user identifier is used only to identify the user, and information that can identify the user is replaced by the user identifier. The data processing in this scheme is performed only on user information that cannot identify the user, such as age, education, and household registration, to protect user privacy. Alternatively, user privacy can be protected by deleting or anonymizing information that can identify the user in the user information. Anonymization can be achieved by processing the data using encryption methods.

[0050] It should be noted that the above is only an example and should not be construed as a limitation of the present invention.

[0051] Next, in step S102, training labels are added to the sample users based on information about changes in whether the sample users hold specific business services; and a user screening model is trained based on the sample feature data of the sample users and the training labels.

[0052] In order to effectively quantify the changing needs of users for Internet services from the feature dimension of user service changes, this invention adds training labels to the sample users based on the information on whether the sample users hold specific business services to construct a user screening model.

[0053] Specifically, a user screening model is established using a logistic regression model or a deep neural network. This user screening model is used to screen specific users, and the first user rating calculated by the user screening model is a value between 0 and 1.

[0054] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other examples, logistic regression models and deep neural networks can also be used to build user screening models.

[0055] The construction of the user screening model includes establishing a training dataset, which involves defining good and bad samples. Specifically, training labels are added to the sample users using the feature of "information on changes in whether a sample user holds a specific service" to define good and bad samples. The user screening model is then trained based on the sample user feature data and the training labels. That is, the training label is "the probability that a user has changed from never holding a specific service to holding one," and the label value is specified as 0 or 1. The first user rating is a value between 0 and 1, where 1 indicates that the user has changed from never holding a specific service to holding one (or, if the sample user changed from never holding a specific service to holding one within a specific time period, then a first label is added to the sample user), and 0 indicates that the user has not changed from never holding a specific service to holding one (or, if the sample user did not hold a specific service within a specific time period, then a second label is added to the sample user). Generally, a higher first user rating indicates a greater probability that the user has gone from never having held a specific service to holding one, and thus a greater likelihood that the user has changed from a bad user to a good user. Conversely, a lower first user rating indicates a greater probability that the user has not gone from never having held a specific service to holding one, and thus a less likely that the user has changed from a bad user to a good user.

[0056] It should be noted that the above are merely examples for illustration and should not be construed as limiting the present invention. In other embodiments, the feature of "whether a user holds a specific service within a specific time period" can be used to add training labels to sample users to define good and bad samples, i.e., the label is "the probability that a user has changed from never holding to holding a specific service within a specific time period," and the label value is specified as 0 or 1. The specific time period includes 6 months, 8 months, 10 months, 12 months, or 18 months, etc. In addition, in other embodiments, multiple parameters such as the number of times and the value of specific services such as coupons or discount coupons are issued within a specific time period can be used to determine the probability of the change. The above are all described as optional examples and should not be construed as limiting the present invention.

[0057] More specifically, the specific business service includes issuing cards, coupons, and / or discount vouchers with a service limit that can be used within a specified period. For example, the specified period may include one month, two months, three months, six months, or one year. For example, the service limit may be 1,000 yuan, 2,000 yuan, or 5,000 yuan, or a range of 1,000 yuan to 5,000 yuan, 1,000 yuan to 3,000 yuan, or 2,000 yuan to 10,000 yuan.

[0058] As one specific implementation, the training dataset includes historical users and their user characteristic data labeled with probability tags indicating changes from never having held a specific business service to holding one. The user characteristics include publicly available user information on the network or user information obtained with user authorization, as well as user identifiers, such as 112**0, or user identifiers represented by letters and / or numbers.

[0059] In another implementation, the training dataset includes historical users and their user characteristics data labeled with probability tags indicating changes from never holding to holding a specific business service, as well as service quota data.

[0060] Furthermore, the service quota data includes the type of Internet service selected by the user, the service quota corresponding to each Internet service type, the service availability time, whether the service is unused, whether the service return is overdue or in breach of contract, etc.

[0061] In another embodiment, the service quota data includes the type of Internet service selected by the user, the service quota corresponding to each Internet service type, the service usage time, the service return time, the number of times the service has been returned, the amount not yet returned, overdue return data, default data, and fraud data, etc.

[0062] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other embodiments, the service quota data may also include whether it includes multiple Internet service types, the number of service types, and whether there is fraudulent data, etc.

[0063] Furthermore, the established user screening model is trained using the training dataset.

[0064] Optionally, the model training process also includes optimizing the model parameters, which include weight parameters and bias parameters between layers of the deep neural network.

[0065] It should be noted that the above is merely an example and should not be construed as limiting the invention. In other examples, model parameters may include the number of layers, the number of iterations, and the learning rate of the deep neural network.

[0066] Specifically, for example, the prior probabilities of the model parameters are set to follow a standard normal distribution. The distributions of the weight and bias parameters are sampled multiple times using the MCMC method to obtain a parameter combination set. More specifically, when testing using the parameter combination set, the same user feature data is forward-propagated multiple times to calculate the mean and statistical variance of each model parameter in order to optimize the model parameters. Finally, the optimized neural network is output. This optimizes the model parameters of the user selection model and improves its accuracy.

[0067] Therefore, by defining good and bad samples using the feature of "whether a user holds a specific business service" to build a training dataset, and using this training dataset to train a user screening model, it is possible to effectively quantify the information on changes in user demand for Internet services from the dimension of the change features of a specific business service from non-existence to availability.

[0068] It should be noted that the above is only an example and should not be construed as a limitation of the present invention.

[0069] Next, in step S103, the user feature data of the candidate users in the candidate user group are input into the trained user screening model to obtain a first user score that characterizes the probability of change of a user's specific business service, and the specific user in the candidate user group is determined based on the first user score.

[0070] Specifically, a candidate user group is determined, and candidates whose internet service quotas (also referred to as service quotas or service allocations in this invention) are selected from this candidate user group.

[0071] It should be noted that, in this invention, the candidate users refer to the group of users who have already used the Internet services provided by the Internet service.

[0072] In one implementation, in a scenario where a service allocation with increased service quotas (e.g., financial services) is selected, a candidate user group within a specified service quota range is determined. For example, the specified service quota range may be, for instance, 1000-2000 yuan, 2000-5000 yuan, etc.

[0073] Specifically, user characteristic data of the candidate user group is obtained, such as user identifiers.

[0074] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other embodiments, user feature data may also be user information publicly available on the Internet or user information obtained with user authorization.

[0075] Next, the user characteristic data of the candidate user group (i.e., the user characteristic data of each candidate user) is input into the trained user screening model to calculate the first user score for each candidate user. The first user score is used to characterize the probability of change in a specific business service. For example, the first user score can characterize the probability of changing from never holding a specific business service to holding a specific business service within a specific time period (e.g., 6 months, 12 months, etc.), or it can characterize the probability of no change from never holding a specific business service to not holding a specific business service within a specific time period (e.g., 6 months, 12 months, etc.).

[0076] In this implementation, specific users are selected based on the first user rating. The specific users are those who have changed from never having a specific service to having a specific service, and the specific users are those whose internet service quota needs to be increased.

[0077] Specifically, the change characteristics include the change from never having held a specific business service to having held a specific business service within a specific time period. The specific time period includes 3 months, 6 months, 9 months, and 12 months, etc.

[0078] Therefore, by inputting the user characteristic data of the candidate user group into the trained user screening model, calculating the first user score, and screening specific users based on the first user score, it is possible to accurately screen specific users whose needs have changed or whose service quotas need to be adjusted.

[0079] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other embodiments, the specific user may also be a user whose Internet service quota is to be reduced.

[0080] In another embodiment, such as Figure 2 As shown, Figure 1 Step S102 is broken down into step S102 and step S201.

[0081] In step S201, a fitting coefficient determination curve is used to determine the first calculated coefficient a.

[0082] It should be noted that, due to Figure 2 Steps S101, S102, and S103 in the process are similar to... Figure 1 Steps S101, S102, and S103 are largely the same, so the descriptions of the identical parts are omitted.

[0083] Specifically, user characteristic data, overdue probability, or resource return probability of historical users of different Internet services are obtained, and the first user score of each historical user is calculated.

[0084] In one embodiment, a curve is determined based on a first user rating and a security level fitting coefficient of a historical user, the coefficient determination curve being used to determine a first calculation coefficient for the user's security level.

[0085] Specifically, the first calculated coefficient 'a' for the specific user is obtained based on the first user rating of the specific user and the coefficient determination curve.

[0086] In another embodiment, a curve is determined based on the first user rating and default probability or resource return probability fitting coefficient of the historical user, and the coefficient determination curve is used to determine the first calculation coefficient 'a' of the increase (or decrease) of the user's service quota.

[0087] Specifically, the coefficients are calculated using the following formula to determine the curve.

[0088] y = Ax 2 +Bx+C (1)

[0089] Where y represents the security level or historical probability of delinquency, probability of default, or probability of resource return; x represents the first calculation coefficient a; and A, B, and C are curve parameters, respectively.

[0090] It should be noted that in this invention, statistical data (e.g., the maximum, minimum, and average values ​​of overdue probability, default probability, or resource return probability, and the proportion of defaulting or overdue users among the total number of users) for different types of internet services over different historical time periods are set. Furthermore, the coefficient determination curve for a specific time period can also be a piecewise function composed of multiple line segments and / or curves. The above is merely illustrative and should not be construed as limiting the invention.

[0091] Therefore, by fitting the curve based on the first user rating and security level, or default probability (or default probability), or service return probability of the historical user, a more accurate coefficient determination curve can be obtained.

[0092] Specifically, based on the user characteristic data of the candidate users and the corresponding Internet service types, a coefficient determination curve is selected for each candidate user, and the first calculated coefficient 'a' for each candidate user is calculated using the selected coefficient determination curve.

[0093] Furthermore, the product of the calculated first calculation coefficient 'a' and the first user score calculated in step S102 is taken as the final first user score for the candidate user. Thus, by calculating the first calculation coefficient based on the determined coefficient curve, a correction process for the calculated first user score can be achieved, resulting in a more accurate first user score.

[0094] It should be noted that the above is only an example and should not be construed as a limitation of the present invention.

[0095] Next, based on the calculated first user rating, specific users are selected from the candidate users.

[0096] Specifically, when the first user score of any candidate user is within a set threshold range, the candidate user is determined to be a specific user, and the specific user is identified as a candidate user who has not held a specific business service to the point where he / she holds a specific business service.

[0097] It should be noted that in this embodiment, the set threshold is a value calculated based on historical users over a certain period of time, but it is not limited to this; it can also be the product of the average of all statistical data and the corresponding first calculation coefficient. The above is only for illustrative purposes and should not be construed as a limitation of the present invention.

[0098] In another embodiment, when the calculated first user score of a candidate user is within a set threshold range, the candidate user is filtered as a specific user, and the specific user is a user who has the characteristic of changing from never holding a specific business service to holding a specific business service.

[0099] Therefore, based on the first user rating, a specific user in the candidate user group can be accurately identified.

[0100] It should be noted that the above is illustrative and should not be construed as a limitation of the present invention.

[0101] Next, in step S104, the first user score and user feature data of the specific user are input into the user security identification model to obtain a second user score; the security level of the specific user is determined based on the first user score and the second user score.

[0102] Specifically, for the selected specific users, the first user score and user feature data of the specific user are input into the user security identification model to calculate the second user score.

[0103] Specifically, for example, an Xgboost model and / or a deep neural network are used to build a user security identification model, and the user security identification model is trained using a training dataset, wherein the training dataset includes user characteristic data of historical users and the probability of default or delinquency during service use.

[0104] For the aforementioned user security identification model, for example, good and bad samples are defined using "whether there was a breach of contract or overdue payment during service use (i.e., the probability of breach of contract or the probability of overdue payment)". That is, the label for "there was a breach of contract or overdue payment during service use" is 1, while the label for "there was a breach of contract or overdue payment during service use" is 0. Thus, the user's good or bad status is characterized by the probability of breach of contract or the probability of overdue payment.

[0105] Specifically, a second user score for the specific user is calculated using a trained user security identification model.

[0106] Optionally, the security level score interval is divided into at least one security level score judgment interval according to a preset security level score threshold range. For each security level score judgment interval, the sum of all security level scores in the security level judgment interval is calculated.

[0107] For the preset security level classification threshold, for example, the preset security level classification threshold range is 0~1, 1~2, 2~3.....N-1~N; if the security level classification interval is 0~1, then the security level is level 1, if it is 1~2, then the security level is level 2; and so on, the security level corresponding to the preset security level classification threshold range.

[0108] Furthermore, the security level of the specific user is determined by judging the range of preset security level score thresholds to which the security level score with the largest sum of security level scores belongs.

[0109] Optionally, it also includes pre-setting multiple service quota ranges to determine the preset service quota ranges corresponding to different security levels.

[0110] In a specific implementation, the service quota is divided into multiple intervals using any value between 0.01 and 0.10. For example, with an interval of 0.03, a second user score of 0.97 to 1.00 corresponds to 8,000 to 10,000 yuan, a security level score of 0.94 to 0.97 corresponds to 6,000 to 8,000 yuan, and so on.

[0111] In another embodiment, such as Figure 3 As shown, it also includes step S301 of determining the adjustment factor for the service quota.

[0112] In step S301, an adjustment coefficient for a specific user is determined, and the security level of the specific user is re-determined based on the adjustment coefficient.

[0113] It should be noted that, due to Figure 3 Steps S101, S102, and S103 in the process are similar to... Figure 1 Steps S101, S102, and S103 are largely the same, so the descriptions of the identical parts are omitted.

[0114] exist Figure 3 In the embodiment shown, the undetermined security level of the specific user is determined based on the first user rating and the second user rating.

[0115] Specifically, determining the adjustment coefficient for a specific user includes determining a second calculation coefficient, wherein a second calculation coefficient corresponding to the second user score is determined based on the calculated second user score.

[0116] Optionally, statistical calculations are performed on historical users within a specific time period, such as 12 months prior to the current time. The preset service quota interval with the largest number of users is selected, or the service quota interval with the number of users greater than the set value or the top three (ranked from largest to smallest) in terms of the number of users are selected. The selected service quota interval is used as a reference service quota interval to determine the corresponding second calculation coefficient.

[0117] Further, it is determined whether the second user score of the specific user is within the range corresponding to the reference service quota interval. If it is within the range corresponding to the reference service quota interval, the second calculation coefficient corresponding to the specific user is determined. If it is not within the range corresponding to the reference service quota interval, the calculated second score is used to determine the corresponding pre-service quota interval, so as to further determine the service quota to be increased (or adjusted). Thus, the service quota of the specific user can be re-determined.

[0118] For the adjustment coefficient, the product of the first calculated coefficient a and the second calculated coefficient b is used as the adjustment coefficient for the specific user.

[0119] Specifically, the undetermined security level is adjusted according to the adjustment coefficient to obtain the security level of the specific user.

[0120] More specifically, based on the adjustment coefficient, a corresponding security level interval is determined from the undetermined security level interval (i.e., the aforementioned preset security level threshold range). Further, the security level of the specific user is determined based on the security level interval.

[0121] In another embodiment, the product of the first calculation coefficient a and the second calculation coefficient b (i.e., the adjustment coefficient) is multiplied by the determined service quota range to redetermine the service quota of the specific user, thereby enabling a more accurate determination of the service quota of the specific user.

[0122] Therefore, by using the product of the first calculation coefficient a and the second calculation coefficient b as the adjustment coefficient for a specific user, the security level of the specific user can be re-determined based on the changes in each user's demand for Internet services over time. This allows for a more accurate determination of the security level of the specific user, further optimizes the user security level determination method, and further optimizes the Internet service provision process.

[0123] The above-described method is for illustrative purposes only, and there are no particular limitations on the order and number of steps. Furthermore, the steps in the above method can be broken down into two or three steps, or some steps can be combined into one step, depending on the specific example.

[0124] Compared with existing technologies, this invention adds training labels to sample users by using the feature of "whether the sample user holds information on changes in specific business services". A user screening model is trained based on the sample user feature data and the training labels. The user feature data of candidate users in the candidate user group are input into the trained user screening model to obtain a first user score representing the probability of changes in a user's specific business services. A specific user in the candidate user group is determined based on the first user score. The first user score and user feature data of the specific user are input into a user security identification model to obtain a second user score. Based on the first user score and the second user score, the security level and its changes of the specific user can be accurately determined, enabling the provision of more reasonable and higher-quality internet services to different users and optimizing the internet service delivery process.

[0125] Furthermore, by determining the curve based on the first user rating and security level fitting coefficient of the historical users, a more accurate coefficient determination curve can be obtained; by calculating the first calculation coefficient based on the determined coefficient determination curve, the calculated first user rating can be corrected, and a more accurate first user rating can be obtained.

[0126] Furthermore, by using the product of the first calculation coefficient a and the second calculation coefficient b as the adjustment coefficient for a specific user, the security level of the specific user can be re-determined based on the changes in each user's demand for Internet services over time. This allows for a more accurate determination of the security level of the specific user, further optimizes the user security level determination method, and further optimizes the Internet service provision process.

[0127] Those skilled in the art will understand that all or part of the steps of the above embodiments are implemented as a program (computer program) executed by a computer data processing device. When the computer program is executed, the method provided by the present invention can be implemented. Moreover, the computer program can be stored in a computer-readable storage medium, which can be a disk, optical disk, ROM, RAM, or other readable storage medium, or a storage array composed of multiple storage media, such as a disk or magnetic tape storage array. The storage medium is not limited to centralized storage; it can also be distributed storage, such as cloud storage based on cloud computing.

[0128] The following describes embodiments of the apparatus of the present invention, which can be used to perform the method embodiments of the present invention. The details described in the apparatus embodiments of the present invention should be considered as supplements to the above method embodiments; details not disclosed in the apparatus embodiments of the present invention can be implemented with reference to the above method embodiments.

[0129] Example 2

[0130] Reference Figure 4 and Figure 5 The user security level determination device of Embodiment 2 of the present invention will be described.

[0131] like Figure 4 As shown, the present invention also provides a user security level determination device based on changes in user services. The user security level device 400 includes: a data acquisition module 401, used to acquire sample feature data of sample users and information on changes in whether sample users hold specific business services; a model training module 402, used to add training labels to the sample users based on the information on changes in whether sample users hold specific business services; and to train a user screening model based on the sample feature data of the sample users and the training labels; a first processing module 403, used to input user feature data of candidate users in a candidate user group into the trained user screening model to obtain a first user score representing the probability of changes in a user's specific business services, and to determine a specific user in the candidate user group based on the first user score; to input the first user score and user feature data of the specific user into a user security identification model to obtain a second user score; and a second processing module 404, used to determine the security level of the specific user based on the first user score and the second user score.

[0132] In order to effectively quantify the changing needs of users for Internet services from the feature dimension of user service changes, this invention adds training labels to the sample users based on the information on whether the sample users hold specific business services to construct a user screening model.

[0133] Specifically, a user screening model is established using a logistic regression model or a deep neural network. This user screening model is used to screen specific users, and the first user rating calculated by the user screening model is a value between 0 and 1.

[0134] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other examples, logistic regression models and deep neural networks can also be used to build user screening models.

[0135] The construction of the user screening model includes establishing a training dataset, which involves defining good and bad samples. Specifically, training labels are added to the sample users using the feature of "information on changes in whether a sample user holds a specific service" to define good and bad samples. The user screening model is then trained based on the sample user feature data and the training labels. That is, the training label is "the probability that a user has changed from never holding a specific service to holding one," and the label value is specified as 0 or 1. The first user rating is a value between 0 and 1, where 1 indicates that the user has changed from never holding a specific service to holding one (or, if the sample user changed from never holding a specific service to holding one within a specific time period, then a first label is added to the sample user), and 0 indicates that the user has not changed from never holding a specific service to holding one (or, if the sample user did not hold a specific service within a specific time period, then a second label is added to the sample user). Generally, a higher first user rating indicates a greater probability that the user has gone from never having held a specific service to holding one, and thus a greater likelihood that the user has changed from a bad user to a good user. Conversely, a lower first user rating indicates a greater probability that the user has not gone from never having held a specific service to holding one, and thus a less likely that the user has changed from a bad user to a good user.

[0136] It should be noted that the above are merely examples for illustration and should not be construed as limiting the present invention. In other embodiments, the feature of "whether a user holds a specific service within a specific time period" can be used to add training labels to sample users to define good and bad samples, i.e., the label is "the probability that a user has changed from never holding to holding a specific service within a specific time period," and the label value is specified as 0 or 1. The specific time period includes 6 months, 8 months, 10 months, 12 months, or 18 months, etc. In addition, in other embodiments, multiple parameters such as the number of times and the value of specific services such as coupons or discount coupons are issued within a specific time period can be used to determine the probability of the change. The above are all described as optional examples and should not be construed as limiting the present invention.

[0137] More specifically, the specific business service includes issuing cards, coupons, and / or discount vouchers with a service limit that can be used within a specified period. For example, the specified period may include one month, two months, three months, six months, or one year. For example, the service limit may be 1,000 yuan, 2,000 yuan, or 5,000 yuan, or a range of 1,000 yuan to 5,000 yuan, 1,000 yuan to 3,000 yuan, or 2,000 yuan to 10,000 yuan.

[0138] As one specific implementation, the training dataset includes historical users and their user characteristic data labeled with probability tags indicating changes from never having held a specific business service to holding one. The user characteristics include publicly available user information on the network or user information obtained with user authorization, as well as user identifiers, such as 112**0, or user identifiers represented by letters and / or numbers.

[0139] It should be noted that the above is only an example and should not be construed as a limitation of the present invention. In other embodiments, the service quota data may also include whether it includes multiple Internet service types, the number of service types, and whether there is fraudulent data, etc.

[0140] Furthermore, the established user screening model is trained using the training dataset.

[0141] Therefore, by defining good and bad samples using the feature of "whether a user holds a specific business service" to build a training dataset, and using this training dataset to train a user screening model, it is possible to effectively quantify the information on changes in user demand for Internet services from the dimension of the change features of a specific business service from non-existence to availability.

[0142] It should be noted that the above is only an example and should not be construed as a limitation of the present invention.

[0143] In another embodiment, such as Figure 5 As shown, it also includes a determining module 501, which is... Figure 4 The second processing module 404 is divided into a determination module 501 and a second processing module 404, wherein the determination module 501 is used to determine the adjustment coefficient.

[0144] Specifically, a curve is determined based on the first user rating and security level fitting coefficient of historical users, and the coefficient determination curve is used to determine the first calculation coefficient of the user's security level; based on the first user rating of the specific user and the coefficient determination curve, the first calculation coefficient 'a' of the specific user is obtained.

[0145] Specifically, the coefficients are calculated using the following formula to determine the curve.

[0146] y = Ax2 +Bx+C (1)

[0147] Where y represents the security level or historical probability of delinquency, probability of default, or probability of resource return; x represents the first calculation coefficient a; and A, B, and C are curve parameters, respectively.

[0148] It should be noted that in this invention, statistical data (e.g., the maximum, minimum, and average values ​​of overdue probability, default probability, or resource return probability, and the proportion of defaulting or overdue users among the total number of users) for different types of internet services over different historical time periods are set. Furthermore, the coefficient determination curve for a specific time period can also be a piecewise function composed of multiple line segments and / or curves. The above is merely illustrative and should not be construed as limiting the invention.

[0149] Therefore, by fitting the curve based on the first user rating and security level, or default probability (or default probability), or service return probability of the historical user, a more accurate coefficient determination curve can be obtained.

[0150] Further, a second calculation coefficient b corresponding to the second user rating is determined.

[0151] Specifically, the product of the first calculation coefficient a and the second calculation coefficient b is used as the adjustment coefficient for the specific user.

[0152] Optionally, based on the first user rating and the second user rating of the specific user, a corresponding undetermined security level interval is determined from a preset security level interval. The undetermined security level of the specific user is then determined based on the undetermined security level interval.

[0153] First, the pending security level of the specific user is determined based on the first user rating and the second user rating.

[0154] Specifically, based on the adjustment coefficient, a corresponding security level interval is determined from the undetermined security level intervals; and the security level of the specific user is re-determined based on the security level intervals.

[0155] More specifically, determining the adjustment coefficient for a specific user includes determining a second calculation coefficient, wherein a second calculation coefficient corresponding to the second user score is determined based on the calculated second user score.

[0156] Optionally, statistical calculations are performed on historical users within a specific time period, such as 12 months prior to the current time. The preset service quota interval with the largest number of users is selected, or the service quota interval with the number of users greater than the set value or the top three (ranked from largest to smallest) in terms of the number of users are selected. The selected service quota interval is used as a reference service quota interval to determine the corresponding second calculation coefficient.

[0157] Further, it is determined whether the second user score of the specific user is within the range corresponding to the reference service quota interval. If it is within the range corresponding to the reference service quota interval, the second calculation coefficient corresponding to the specific user is determined. If it is not within the range corresponding to the reference service quota interval, the calculated second score is used to determine the corresponding pre-service quota interval, so as to further determine the service quota to be increased (or adjusted). Thus, the service quota of the specific user can be re-determined.

[0158] Compared with existing technologies, this invention adds training labels to sample users by using the feature of "whether the sample user holds information on changes in specific business services". A user screening model is trained based on the sample user feature data and the training labels. The user feature data of candidate users in the candidate user group are input into the trained user screening model to obtain a first user score representing the probability of changes in a user's specific business services. A specific user in the candidate user group is determined based on the first user score. The first user score and user feature data of the specific user are input into a user security identification model to obtain a second user score. Based on the first user score and the second user score, the security level and its changes of the specific user can be accurately determined, enabling the provision of more reasonable and higher-quality internet services to different users and optimizing the internet service delivery process.

[0159] Furthermore, by determining the curve based on the first user rating and security level fitting coefficient of the historical users, a more accurate coefficient determination curve can be obtained; by calculating the first calculation coefficient based on the determined coefficient determination curve, the calculated first user rating can be corrected, and a more accurate first user rating can be obtained.

[0160] Furthermore, by using the product of the first calculation coefficient a and the second calculation coefficient b as the adjustment coefficient for a specific user, the security level of the specific user can be re-determined based on the changes in each user's demand for Internet services over time. This allows for a more accurate determination of the security level of the specific user, further optimizes the user security level determination method, and further optimizes the Internet service provision process.

[0161] Those skilled in the art will understand that the modules in the above-described device embodiments can be distributed throughout the device as described, or they can be modified accordingly and distributed in one or more devices different from the above embodiments. The modules in the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0162] Example 3

[0163] The following describes embodiments of the computer device of the present invention, which can be considered as specific implementations of the methods and apparatus embodiments of the present invention described above. Details described in the computer device embodiments of the present invention should be considered as supplements to the methods or apparatus embodiments described above; details not disclosed in the computer device embodiments of the present invention can be implemented with reference to the methods or apparatus embodiments described above.

[0164] Figure 6 This is a structural block diagram of an exemplary embodiment of a computer device according to the present invention. Referring below... Figure 6 To describe the computer device 200 according to this embodiment of the present invention. Figure 6 The computer device 200 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0165] like Figure 6 As shown, the computer device 200 is presented in the form of a general-purpose computing device. The components of the computer device 200 may include, but are not limited to: at least one processing unit 210, at least one storage unit 220, a bus 230 connecting different device components (including storage unit 220 and processing unit 210), a display unit 240, etc.

[0166] The storage unit stores program code that can be executed by the processing unit 210, causing the processing unit 210 to perform the steps described in the processing method section of the computer device described above, according to various exemplary embodiments of the present invention. For example, the processing unit 210 can perform, for example... Figure 1 The steps are shown.

[0167] The storage unit 220 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 2201 and / or a cache storage unit 2202, and may further include a read-only memory unit (ROM) 2203.

[0168] The storage unit 220 may also include a program / utility 2204 having a set (at least one) program module 2205, such program module 2205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0169] Bus 230 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0170] Computer device 200 can also communicate with one or more external devices 300 (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with the computer device 200, and / or any device that enables the computer device 200 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 250. Furthermore, computer device 200 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 260. Network adapter 260 can communicate with other modules of computer device 200 via bus 230. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with computer device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0171] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described in this invention can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the method described above according to this invention. When the computer program is executed by a data processing device, the computer program product enables the implementation of the method described above according to this invention.

[0172] like Figure 7As shown, the computer program can be stored on one or more computer program products. The computer program product can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer program products (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0173] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer program product may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0174] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0175] In summary, this invention can be implemented in hardware, or as a software module running on one or more processors, or a combination thereof. Those skilled in the art will understand that in practice, general-purpose data processing devices such as microprocessors or digital signal processors (DSPs) can be used to implement some or all of the functions of some or all of the components according to the embodiments of the invention. This invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such a program implementing the invention can be stored on a computer program product, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0176] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or electronic device, and various general-purpose devices can also implement the present invention. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for determining user security levels based on changes in user services, characterized in that, Extract sample characteristic data of sample users and information on changes in whether sample users hold specific business services from historical data of Internet services; Training labels are added to sample users based on information about changes in whether they hold specific business services. The labels include: a first label indicating that the user has changed from never holding to holding specific business services within a specific time period, and a second label indicating that the user has not changed from never holding to holding specific business services within a specific time period. Furthermore, a user screening model is trained based on the sample user feature data and training labels to obtain a trained user screening model. From users who have already used Internet services, a candidate user group is identified. The user characteristic data of the candidate users in the candidate user group is input into the trained user screening model to obtain a first user score that represents the probability of change of a user’s specific business service. Based on the first user score, a specific user in the candidate user group is identified. The specific user has the change characteristic of never having held a specific business service to having held a specific business service within a specific time period. The first user score and user feature data of a specific user are input into the trained user security identification model to obtain the second user score. Determining the security level of a specific user based on a first user rating and a second user rating includes: determining the corresponding pending security level interval from a preset security level interval based on the first user rating and the second user rating of the specific user to determine the pending security level of the specific user, and then adjusting the pending security level according to an adjustment coefficient to determine the security level corresponding to the specific user.

2. The method according to claim 1, characterized in that, Determining a specific user's security level based on the first user rating and the second user rating also includes: A curve is determined based on the first user rating and security level fitting coefficient of historical users. The coefficient determination curve is used to determine the first calculation coefficient for the user's security level; and a second calculation coefficient corresponding to the second user rating is determined. The product of the first calculated coefficient and the second calculated coefficient is used as the adjustment coefficient for a specific user.

3. The method according to claim 2, characterized in that, Also includes: Based on the first user rating of a specific user and the selection of corresponding coefficients, the curve is determined to obtain the first calculated coefficient 'a' of the specific user. A second calculation coefficient b is determined based on the second user rating of a specific user; the product of the first calculation coefficient a and the second calculation coefficient b is used as the adjustment coefficient for the specific user to adjust the pending security level and obtain the security level of the specific user.

4. The method according to claim 1, characterized in that, The pending security level is adjusted based on an adjustment factor to determine the security level corresponding to a specific user, including: Based on the adjustment coefficient, determine the corresponding safety level interval from the undetermined safety level interval; The security level of a specific user is determined based on the corresponding security level range.

5. The method according to claim 4, characterized in that, The security level of a specific user is determined based on the corresponding security level range, including: Based on the preset security level score threshold range, the security level score interval is divided into at least one security level score judgment interval; For each security level score range, calculate the sum of all security level scores within that range. The security level of a specific user is determined by identifying the preset security level threshold range to which the range with the highest sum of security level scores belongs.

6. The method according to claim 1, characterized in that, Specific users within the candidate user group are identified based on the first user rating, including: When the first user rating of any candidate user is within a set threshold range, the candidate user is identified as a specific user. The specific user is identified as a candidate user who does not hold specific business services to a candidate user who holds specific business services.

7. The method according to claim 1 or 6, characterized in that, Training labels are added to sample users based on information regarding changes in whether they hold specific business services. Specifically, these labels include: If a sample user goes from never having a specific business service to having a specific business service within a specific time period, then a first tag is added to the sample user. If a sample user does not hold a specific business service within a specific time period, a second tag is added to the sample user. Specific business services include issuing discount cards, coupons, and / or vouchers that can be used within a specified period.

8. A device for determining user security level based on changes in user services, characterized in that, include: The data acquisition module is used to acquire sample characteristic data of sample users and information on changes in whether sample users hold specific business services from historical data of Internet services; The model training module adds training labels to the sample users based on information about changes in whether the sample users hold specific business services. The labels include: a first label indicating that the user has changed from never holding to holding specific business services within a specific time period, and a second label indicating that the user has not changed from never holding to holding specific business services within a specific time period. Furthermore, the user screening model is trained based on the sample user feature data and training labels to obtain the trained user screening model. The first processing module is used to determine a candidate user group from users who have used Internet services, input the user feature data of the candidate users in the candidate user group into a trained user screening model to obtain a first user score that represents the probability of a user’s change in a specific business service, determine a specific user in the candidate user group based on the first user score, the specific user having the change characteristic of never having held a specific business service to having it within a specific time period; input the first user score and user feature data of the specific user into a trained user security identification model to obtain a second user score; The second processing module determines the security level of a specific user based on the first user score and the second user score, including: determining the corresponding pending security level interval from the preset security level interval based on the first user score and the second user score of the specific user to determine the pending security level of the specific user, and then adjusting the pending security level according to the adjustment coefficient to determine the security level corresponding to the specific user.

9. The apparatus according to claim 8, characterized in that, It also includes a determination module for determining the adjustment coefficient, wherein, The curve is determined based on the first user rating and security level fitting coefficient of historical users, so as to determine the first calculation coefficient of the security level of the corresponding specific user. Determine the second calculation coefficient corresponding to the second user rating; The product of the first calculated coefficient and the second calculated coefficient is used as the adjustment coefficient for a specific user.

10. The apparatus according to claim 9, characterized in that, The determination module also includes: Based on the first user rating of a specific user and the selection of corresponding coefficients, the curve is determined to obtain the first calculated coefficient 'a' of the specific user. The second calculation coefficient b is determined based on the second user rating of a specific user; The product of the first calculation coefficient a and the second calculation coefficient b is used as the adjustment coefficient for a specific user to adjust the pending security level and obtain the security level of the specific user.

11. The apparatus according to claim 10, characterized in that, The second processing module specifically includes: Based on the adjustment coefficient, determine the corresponding safety level interval from the undetermined safety level interval; The security level of a specific user is determined based on the corresponding security level range.

Images