A method and device for determining the degree of possession of invalid users

Abstract

The present invention discloses a method and device for determining the degree of invalid users, wherein the method comprises: obtaining user data of a reference user group and user data of a user group to be tested; for each of a plurality of data feature dimensions, determining a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested under the data feature dimension; and determining the degree of invalid users of the user group to be tested according to the similarity index of each of the plurality of data feature dimensions. When the above method is applied to financial technology (Fintech), even if the user is not in the invalid user list, the degree of invalid users of the user group to be tested can be determined by comparing the similarity with the user data of the invalid user group or the valid user group, so that the determination of the degree of invalid users is more accurate.

Description

Technical Field

[0001] The present invention relates to the field of information security in the field of financial technology (Fintech), and in particular to a method and device for determining the degree of possession of an invalid user. Background Art

[0002] With the development of computer technology, more and more technologies are being applied in the financial field. Traditional finance is gradually transforming to Fintech. However, due to the security and real-time requirements of the financial industry, higher requirements are also placed on technology. For financial institutions, how to promote corresponding financial tools such as financial mobile software (Application, App) is crucial. The number of effective registered users is an important indicator for evaluating the promotion degree of financial tools.

[0003] However, in the current promotion of financial instruments, there may be a risk that a large number of invalid users are present among the registered users, such as a large number of users who take advantage of the platform. These invalid users occupy a large amount of promotion resources. How to timely detect this risk is of great significance to the promotion of financial instruments. The current method is to match the registered users in the financial instrument with the list of invalid users to determine the proportion of invalid users. If the proportion exceeds the preset ratio, it is determined that there are a large number of invalid users. Obviously, in the current method, the list of invalid users has limited coverage, and it is difficult to accurately determine the proportion of invalid users. Summary of the invention

[0004] The present invention provides a method and device for determining the occupation degree of invalid users, which solves the problem in the prior art that it is difficult to accurately determine the occupation degree of invalid users.

[0005] In a first aspect, the present invention provides a method for determining the degree of invalid users, comprising: obtaining user data of a reference user group and user data of a user group to be tested; the reference user group is an invalid user group or a valid user group; the user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions; for each of the multiple data feature dimensions, determining a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested under the data feature dimension; and determining the degree of invalid users of the user group to be tested according to the similarity index of each of the multiple data feature dimensions.

[0006] Under the above method, after obtaining the user data of the reference user group and the user data of the user group to be tested, the similarity of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension can be determined by comparing the similarity index of each data feature dimension in multiple data feature dimensions. Therefore, even if the users are not in the invalid user list, the proportion of invalid users in the user group to be tested can be determined by comparing the similarity with the user data of the invalid user group or the valid user group, so that the determination of the proportion of invalid users is more accurate.

[0007] Optionally, determining the degree of invalid users of the user group to be tested according to the similarity index of each data feature dimension in the multiple data feature dimensions includes: determining a possession index of the degree of invalid users of the user group to be tested under the data feature dimension according to the similarity index of each data feature dimension in the multiple data feature dimensions; the possession index is used to characterize the degree of invalid users of the user group to be tested under the data feature dimension; determining the possession index of the degree of invalid users of the user group to be tested under the multiple data feature dimensions according to the possession index of the degree of invalid users of the user group to be tested under each data feature dimension in the multiple data feature dimensions and the weight value of the data feature dimension; determining the degree of invalid users of the user group to be tested according to the possession index of the degree of invalid users of the user group to be tested under the multiple data feature dimensions.

[0008] In the above method, firstly, according to the similarity index of each data feature dimension in the multiple data feature dimensions, the occupation index of the invalid users of the user group to be tested under the data feature dimension is determined, so as to first characterize the occupation degree of invalid users of the user group to be tested under this data feature dimension, and then the occupation index of the invalid users of the user group to be tested under the multiple data feature dimensions is used to characterize the occupation degree of invalid users under the multiple data feature dimensions, so as to comprehensively consider the multiple data feature dimensions and the corresponding weight values, and accurately determine the occupation degree of invalid users of the user group to be tested.

[0009] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is negatively correlated with the degree of dispersion of the user data of the reference user group in the data feature dimension.

[0010] In the above method, for the user data of the reference user group and the user data of the user group to be tested, the smaller the discreteness in a data feature dimension, the lower the difference between the data. Therefore, the similarity index in this data feature dimension is more sensitive to the difference in the data, so that the degree of invalid users can be more clearly reflected by setting a negative correlation relationship for the weight value.

[0011] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is calculated in the following manner: for each data feature dimension in the multiple data feature dimensions, determining the standard deviation of the user data of the reference user group in the data feature dimension, so as to characterize the degree of dispersion of the user data of the reference user group in the data feature dimension; determining the weight value of the data feature dimension according to the inverse of the standard deviation of the user data of the reference user group in the multiple data feature dimensions; the weight value of the data feature dimension is positively correlated with the inverse of the standard deviation of the user data of the reference user group in the data feature dimension.

[0012] In the above method, the standard deviation of the user data of the reference user group in the data feature dimension is used to characterize the degree of dispersion of the user data of the reference user group in the data feature dimension, and the weight value of the data feature dimension is determined according to the reciprocal of the standard deviation of the user data of the reference user group in the multiple data feature dimensions, thereby providing a method for determining the weight value of each data feature dimension in the multiple data feature dimensions according to the standard deviation and the reciprocal of the standard deviation.

[0013] Optionally, the possession degree is high possession or low possession; the possession index of the possession degree of invalid users in the user group to be tested under the data feature dimension is the probability that the possession degree of invalid users in the user group to be tested under the data feature dimension is high possession or low possession.

[0014] In the above manner, the occupation degree is represented by using a high occupation probability or a low occupation probability, thereby more intuitively reflecting the occupation degree of the invalid users in the user group to be tested.

[0015] Optionally, determining, for each data feature dimension in multiple data feature dimensions, a similarity index of data distribution between user data of the reference user group and user data of the user group to be tested in the data feature dimension includes: determining, for each data feature dimension in multiple data feature dimensions, multiple categories of subgroups in the data feature dimension; determining, based on the user data of the reference user group and the user data of the user group to be tested, the number of users in each category of the subgroups of the reference user group and the user group to be tested; determining, based on the number of users in each category of the subgroups of the reference user group and the user group to be tested, a similarity index of data distribution between user data of the reference user group and user data of the user group to be tested in the data feature dimension.

[0016] In the above manner, each data feature dimension is subdivided into multiple categories of subgroups, and the number of users in each category of the reference user group and the user group to be tested is determined, so that by considering multiple categories of subgroups, the similarity index of the data distribution of the user data of the reference user group and the user data of the user group to be tested under the data feature dimension is determined.

[0017] Optionally, the similarity index of the data distribution under the data feature dimension is: under the data feature dimension, the Pearson correlation coefficient or cosine similarity between the user data of the reference user group and the user data of the user group to be tested.

[0018] In the above manner, the Pearson correlation coefficient or the cosine similarity can characterize the similarity between the user data of the reference user group and the user data of the user group to be tested.

[0019] Optionally, the multiple data feature dimensions include at least one of the following: the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location.

[0020] Since invalid users are often concentrated in certain age groups or in various numbers’ locations, focusing on the above-mentioned multiple data feature dimensions can more accurately determine the extent of invalid users.

[0021] In a second aspect, the present invention provides a device for determining the degree of invalid user occupation, comprising: an acquisition module, used to acquire user data of a reference user group and user data of a user group to be tested; the reference user group is an invalid user group or a valid user group; the user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions; a processing module, used to determine, for each of the multiple data feature dimensions, a similarity index of data distribution of the user data of the reference user group and the user data of the user group to be tested under the data feature dimension; and used to determine the degree of invalid user occupation of the user group to be tested according to the similarity index of each of the multiple data feature dimensions.

[0022] Optionally, the processing module is specifically used to: determine, according to the similarity index of each data feature dimension in the multiple data feature dimensions, an occupation index of the invalid users of the user group to be tested under the data feature dimension; the occupation index is used to characterize the occupation degree of invalid users of the user group to be tested under the data feature dimension; determine, according to the occupation index of the invalid users of the user group to be tested under each data feature dimension in the multiple data feature dimensions and the weight value of the data feature dimension, the occupation index of the invalid users of the user group to be tested under the multiple data feature dimensions; determine the occupation degree of invalid users of the user group to be tested according to the occupation index of the invalid users of the user group to be tested under the multiple data feature dimensions.

[0023] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is negatively correlated with the degree of dispersion of the user data of the reference user group in the data feature dimension.

[0024] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is calculated in the following manner: for each data feature dimension in the multiple data feature dimensions, determining the standard deviation of the user data of the reference user group in the data feature dimension, so as to characterize the degree of dispersion of the user data of the reference user group in the data feature dimension; determining the weight value of the data feature dimension according to the inverse of the standard deviation of the user data of the reference user group in the multiple data feature dimensions; the weight value of the data feature dimension is positively correlated with the inverse of the standard deviation of the user data of the reference user group in the data feature dimension.

[0025] Optionally, the possession degree is high possession or low possession; the possession index of the possession degree of invalid users in the user group to be tested under the data feature dimension is the probability that the possession degree of invalid users in the user group to be tested under the data feature dimension is high possession or low possession.

[0026] Optionally, the processing module is specifically used to: determine, for each data feature dimension in multiple data feature dimensions, multiple categories of subgroups under the data feature dimension; determine, based on user data of the reference user group and user data of the user group to be tested, the number of users in each category of subgroups of the reference user group and the user group to be tested in the multiple categories of subgroups; determine, based on the number of users in each category of subgroups of the reference user group and the user group to be tested in the multiple categories of subgroups, a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension.

[0027] Optionally, the similarity index of the data distribution under the data feature dimension is: under the data feature dimension, the Pearson correlation coefficient or cosine similarity between the user data of the reference user group and the user data of the user group to be tested.

[0028] Optionally, the multiple data feature dimensions include at least one of the following: the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location.

[0029] The beneficial effects of the above-mentioned second aspect and each optional device of the second aspect can refer to the beneficial effects of the above-mentioned first aspect and each optional method of the first aspect, and will not be repeated here.

[0030] In a third aspect, the present invention provides a computer device, including a program or an instruction, which, when executed, is used to execute the above-mentioned first aspect and each optional method of the first aspect.

[0031] In a fourth aspect, the present invention provides a storage medium, including a program or an instruction, which, when executed, is used to execute the above-mentioned first aspect and each optional method of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic diagram of the steps of a method for determining the degree of possession of an invalid user provided in an embodiment of the present application;

[0033] Figure 2 A schematic diagram of the structure of a device for determining the degree of possession of invalid users provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to better understand the above-mentioned technical scheme, the above-mentioned technical scheme will be described in detail below in combination with the accompanying drawings and specific implementation methods of the specification. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments may be combined with each other.

[0035] The following first lists the nouns that appear in the embodiments of the present application.

[0036] Valid users: Users who register the App for normal usage purposes.

[0037] Invalid users: users who register the App for non-normal purposes, such as wool-picking users. The main purpose of these users registering the App is to obtain certain promotional rewards, and they rarely log in to the App again.

[0038] Real-name registration: refers to the user information submitted when registering the App.

[0039] Pearson correlation test: By calculating the correlation coefficient of two groups of data X and Y with a sample size of n, we can test whether the two groups of data are significantly correlated. The correlation coefficient range is [-1,1]. The closer the value is to 1, the stronger the positive linear correlation between the two groups of data is. The closer the value is to -1, the stronger the negative linear correlation between the two groups of data is. The closer the value is to 0, the less linear correlation between the two groups of data is. The specific definition of the correlation coefficient is as follows

[0040]

[0041] In the process of financial institutions (banking institutions, insurance institutions or securities institutions) conducting business (such as bank loan business, deposit business, etc.), how to promote the corresponding financial tool App to valid users is of great significance to the financial institutions. However, in the current promotion process of financial tools, there may be a large number of invalid users registered, and invalid users will waste a lot of promotion resources. Therefore, how to determine the proportion of invalid users is an issue worthy of attention. In the current method, the proportion of invalid users is determined by matching the registered users in the financial tool with the list of invalid users. However, the coverage of the list of invalid users is limited, and it is difficult to accurately determine the proportion of invalid users. This situation does not meet the needs of banks and other financial institutions, and cannot guarantee the efficient operation of various businesses of financial institutions.

[0042] For this reason, Figure 1 As shown, an embodiment of the present application provides a method for determining the degree of possession of invalid users.

[0043] Step 101: Obtain user data of a reference user group and user data of a user group to be tested.

[0044] Step 102: for each data feature dimension among the multiple data feature dimensions, determine a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension.

[0045] Step 103: Determine the proportion of invalid users in the user group to be tested according to the similarity index of each data feature dimension in the multiple data feature dimensions.

[0046] In step 101 to step 103, the meaning of user may be a registered user, and the reference user group may be an invalid user group or a valid user group. The invalid user group may be a user group including a large number of invalid users (such as exceeding an invalid user ratio threshold) among the users, and the valid user group may be a user group including a large number of valid users (such as exceeding an invalid user ratio threshold) among the users.

[0047] Specifically, the invalid user group or valid user group can be determined according to the invalid users in the user group within the preset historical time range. The reference user group (such as marked as user_group0) can be an invalid user group or a valid user group, and whether there is a large number of suspected invalid users to be detected is used as the user group to be detected (such as marked as user_group1), for example, all users registered on the Nth day are used as user_group1, and all users registered from the Nth to the 7th day are used as user_group0.

[0048] The user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions. Specifically, a data feature dimension may refer to a characteristic used to describe a certain aspect of a group of user data. For example, a data feature dimension may be the age of the user, the online duration of the user, etc. Under each of the multiple data feature dimensions, the similarity index of the user data of the reference user group and the user data of the user group to be tested represents the data similarity between the user data of the reference user group and the user data of the user group to be tested. The more similar the similarity index represents, the higher the degree of invalid users in the user group to be tested (such as the invalid user group), or the lower the degree of valid users in the user group to be tested. The same is true for the case where the user group to be tested is a valid user group. In addition, the degree of possession may be defined as a specific value, such as a high degree of possession or a low degree of possession. Further, more values ​​(such as multiple degrees of possession) may be defined to represent the degree of possession of valid users or invalid users, and a mapping relationship between the value of the similarity index and the value of the degree of possession may be established.

[0049] It should be noted that invalid users usually show clustering characteristics in certain data feature dimensions when registering. For example, the registered users are all older, the ID card numbers used for registration are concentrated in certain places, and the IP addresses reported during registration are concentrated in certain places.

[0050] Therefore, several specific data feature dimensions can be selected as multiple data feature dimensions. In an optional implementation of selecting multiple data feature dimensions, the multiple data feature dimensions may include at least one of the following: the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location.

[0051] An optional implementation of step 102 is as follows:

[0052] For each data feature dimension in a plurality of data feature dimensions, a plurality of categories of subgroups under the data feature dimension are determined; based on the user data of the reference user group and the user data of the user group to be tested, the number of users in each category of the subgroups of the reference user group and the user group to be tested in the plurality of categories is determined; based on the number of users in each category of the subgroups of the reference user group and the user group to be tested in the plurality of categories, a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension is determined.

[0053] Specifically, taking an optional implementation method of selecting multiple data feature dimensions as an example, the multiple data feature dimensions can be obtained by processing in the following manner:

[0054] a) Statistics on the number of users in multiple subgroups of the user's age group: The user ages at the time of real-name registration are grouped by age group, such as 20 years old and below, 21-30 years old, 31-40 years old, 41-50 years old, 51-60 years old, 61-70 years old, 71-80 years old, 81-90 years old, 91 years old and above, and the number of users in each subgroup of each age group and category is counted, which can be recorded as age_cnt.

[0055] b) Statistics on the number of users in multiple subgroups of the place where the user's ID card number belongs: Take the first 6 digits of the ID card number and associate them with the city to which the ID card number belongs, and count the number of users in each subgroup of the place where the ID card number belongs, recorded as idno_attr_cnt.

[0056] c) Statistics on the number of users in multiple subgroups of the user's mobile phone number's location: Take the first k digits (such as 7 digits) of the mobile phone number and associate them to get the city where the mobile phone number belongs. If the mobile phone number is a virtual operator mobile phone number, such as those starting with 147, 170, or 171, the location is divided into a special virtual operator category subgroup (also regarded as a subgroup of the city category). Count the number of users in each subgroup of the mobile phone number's location, and record it as phone_attr_cnt.

[0057] d) Statistics on the number of users in multiple categories of subgroups of the user's bank card number location: associate the bank card number with the bank card number location, and count the number of users in each category of subgroups of the bank card number location, recorded as card_attr_cnt.

[0058] e) Statistics on the number of users in multiple categories of subgroups of the user's Internet protocol (IP) address attribution: associate the IP address to obtain the IP address attribution, and count the number of users in each category of subgroups of the IP address, recorded as ip_attr_cnt.

[0059] At the same time, for the same data feature of the user data of the user group to be tested and the user data of the reference user group, if one group of users has a subgroup but the other group of users does not have a corresponding subgroup, then the corresponding number of users in this group on the subgroup is filled with 0, such as the two 0s in Table 1:

[0060]

[0061] Table 1

[0062] It should be noted that in an optional implementation of steps 101 to 103, the similarity index of data distribution under the data feature dimension can be: the Pearson correlation coefficient or cosine similarity between the user data of the reference user group and the user data of the user group to be tested under the data feature dimension. The following takes the Pearson correlation coefficient and the reference user group as an example to illustrate the similarity index ρ between the user data of the reference user group and the user data of the user group to be tested under the data feature dimension of the user's age group in step 102. age :

[0063]

[0064] where age_cnt0 i is the number of users in the i-th age group of the reference user group, is the average number of users in the n age groups of the reference user group, age_cnt1 iis the number of users in the i-th age group of the user group to be tested, is the average number of users in the n age groups of the user group to be tested.

[0065] An optional implementation of step 103 is as follows:

[0066] Step (103-1): according to the similarity index of each data feature dimension in the multiple data feature dimensions, determine the occupation index of the invalid users in the user group to be tested under the data feature dimension.

[0067] Step (103-2): Determine the occupation index of the invalid users of the user group to be tested in each of the multiple data feature dimensions according to the occupation index of the invalid users of the user group to be tested in each of the multiple data feature dimensions and the weight value of the data feature dimension.

[0068] Step (103-3): Determine the occupation degree of invalid users in the user group to be tested according to the occupation index of the invalid users in the user group to be tested under the multiple data feature dimensions.

[0069] In step (103-1), the occupancy index is used to characterize the occupancy degree of invalid users in the user group to be tested under the data feature dimension. In an optional implementation, the occupancy degree is high occupancy or low occupancy; the occupancy index of the occupancy degree of invalid users in the user group to be tested under the data feature dimension is the probability that the occupancy degree of invalid users in the user group to be tested under the data feature dimension is high occupancy or low occupancy.

[0070] In the above optional implementation method where the degree of possession is probability, based on the similarity index ρ age In an example of an optional implementation mode, step (103-1) obtains the probability P that the proportion of invalid users in the user group to be tested is high under the data feature dimension of the user's age group. age An example can be as follows, P age The value range is [0,1]. Specifically:

[0071]

[0072] According to the definition of Pearson correlation coefficient, under a certain data feature dimension (such as the user's age group), the stronger the positive linear correlation between the user data of the reference user group and the user data of the user group to be tested, the more similar the distribution of the user data of the reference user group (effective user group) and the user data of the user group to be tested is under this data feature dimension (such as the user's age group). Then P ageThe smaller it is, that is, the possibility that the proportion of invalid users in the user group to be tested is high under the data feature dimension of the user's age group is low, and the corresponding probabilities of other data feature dimensions are similar.

[0073] Further, based on the example of the above step (103-1), the user's age group; the user's ID card number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location can be obtained in sequence. The probability that the proportion of invalid users in the user group to be tested is high is: P age , P idno , P phone , P card , P ip , and according to the weight value of each data feature dimension (corresponding to w age 、w idno 、w phone 、w card 、w ip ), to specifically execute step (103-2): S = w age *P age +w idno *P idno +w phone *P phone +w card *P card +w ip *P ip S is the occupation index of the invalid users in the user group to be tested in the multiple data feature dimensions.

[0074] In an optional implementation, the weight value of each data feature dimension may be set in the following manner: the weight value of each data feature dimension in the multiple data feature dimensions is negatively correlated with the degree of dispersion of the user data of the reference user group in the data feature dimension.

[0075] Since the user data of the reference user group and the user data of the user group to be tested have a smaller degree of discreteness in a data feature dimension, the lower the difference between the data, the similarity index in this data feature dimension is more sensitive to the difference in the data. Therefore, by setting a negative correlation between the weight values, the degree of invalid users can be more clearly reflected.

[0076] It should be noted that the weight value of each data feature dimension in the multiple data feature dimensions is specifically calculated in the following manner:

[0077] For each data feature dimension of the multiple data feature dimensions, determine the standard deviation of the user data of the reference user group in the data feature dimension, so as to characterize the degree of dispersion of the user data of the reference user group in the data feature dimension; determine the weight value of the data feature dimension according to the inverse of the standard deviation of the user data of the reference user group in the multiple data feature dimensions; the weight value of the data feature dimension is positively correlated with the inverse of the standard deviation of the user data of the reference user group in the data feature dimension.

[0078] Taking the data feature dimension as the user's age group as an example, the weight value w corresponding to the user's age group is age It can be determined as follows:

[0079]

[0080] where σ age , σ idno , σ phone , σ card , σ ip The data of the reference user group are the standard deviations in the data feature dimensions of the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; and the user's Internet Protocol IP address location.

[0081] Therefore w age +w idno +w phone +w card +w ip =1, the value range of S is [0,1]. You can set the preset alarm threshold corresponding to S. When S is greater than the preset alarm threshold, an alarm is triggered, indicating that there are a high proportion of invalid users in the user group to be tested. At the same time, the closer the value of S is to 1, the more serious the phenomenon of invalid user registration is.

[0082] like Figure 2 The present invention provides a device for determining the degree of invalid user occupation, comprising: an acquisition module 201, used to acquire user data of a reference user group and user data of a user group to be tested; the reference user group is an invalid user group or a valid user group; the user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions; a processing module 202, used to determine, for each of the multiple data feature dimensions, a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested under the data feature dimension; and used to determine the degree of invalid user occupation of the user group to be tested according to the similarity index of each of the multiple data feature dimensions.

[0083] Optionally, the processing module 202 is specifically used to: determine, according to the similarity index of each data feature dimension in the multiple data feature dimensions, a possession index of the invalid users of the user group to be tested under the data feature dimension; the possession index is used to characterize the possession degree of invalid users of the user group to be tested under the data feature dimension; determine, according to the possession index of the invalid users of the user group to be tested under each data feature dimension in the multiple data feature dimensions and the weight value of the data feature dimension, the possession index of the invalid users of the user group to be tested under the multiple data feature dimensions; determine the possession degree of invalid users of the user group to be tested according to the possession index of the invalid users of the user group to be tested under the multiple data feature dimensions.

[0084] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is negatively correlated with the degree of dispersion of the user data of the reference user group in the data feature dimension.

[0085] Optionally, the weight value of each data feature dimension in the multiple data feature dimensions is calculated in the following manner: for each data feature dimension in the multiple data feature dimensions, determining the standard deviation of the user data of the reference user group in the data feature dimension, so as to characterize the degree of dispersion of the user data of the reference user group in the data feature dimension; determining the weight value of the data feature dimension according to the inverse of the standard deviation of the user data of the reference user group in the multiple data feature dimensions; the weight value of the data feature dimension is positively correlated with the inverse of the standard deviation of the user data of the reference user group in the data feature dimension.

[0086] Optionally, the possession degree is high possession or low possession; the possession index of the possession degree of invalid users in the user group to be tested under the data feature dimension is the probability that the possession degree of invalid users in the user group to be tested under the data feature dimension is high possession or low possession.

[0087] Optionally, the processing module 202 is specifically used to: determine, for each data feature dimension in multiple data feature dimensions, multiple categories of subgroups under the data feature dimension; determine, based on the user data of the reference user group and the user data of the user group to be tested, the number of users in each category of the subgroups of the reference user group and the user group to be tested in the multiple categories of subgroups; determine, based on the number of users in each category of the subgroups of the reference user group and the user group to be tested in the multiple categories of subgroups, a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension.

[0088] Optionally, the similarity index of the data distribution under the data feature dimension is: under the data feature dimension, the Pearson correlation coefficient or cosine similarity between the user data of the reference user group and the user data of the user group to be tested.

[0089] Optionally, the multiple data feature dimensions include at least one of the following: the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location.

[0090] An embodiment of the present application provides a computer device, including a program or instruction, which, when executed, is used to execute a method for determining the degree of possession of an invalid user and any optional method provided in an embodiment of the present application.

[0091] An embodiment of the present application provides a storage medium, including a program or an instruction, which, when executed, is used to execute a method for determining the degree of possession of an invalid user and any optional method provided in an embodiment of the present application.

[0092] Finally, it should be noted that: those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0093] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0094] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for determining the degree of possession of invalid users, characterized in that: include: Obtain user data of a reference user group and user data of a user group to be tested; The reference user group is an invalid user group or a valid user group; the user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions; For each data feature dimension of the plurality of data feature dimensions, determining a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension; Determine, according to the similarity index of each data feature dimension in the multiple data feature dimensions, a occupancy index of the occupancy degree of invalid users of the user group to be tested under the data feature dimension; the occupancy index is used to characterize the occupancy degree of invalid users of the user group to be tested under the data feature dimension; Determine the occupation index of the invalid users of the user group to be tested in the multiple data feature dimensions according to the occupation index of the invalid users of the user group to be tested in each of the multiple data feature dimensions and the weight value of the data feature dimension; the weight value of each of the multiple data feature dimensions is negatively correlated with the discrete degree of the user data of the reference user group in the data feature dimension; The degree of invalid users in the user group to be tested is determined according to the occupation index of the invalid users in the user group to be tested under the multiple data feature dimensions.

2. The method according to claim 1, characterized in that The weight value of each data feature dimension in the multiple data feature dimensions is specifically calculated in the following manner: For each data feature dimension of the multiple data feature dimensions, determine a standard deviation of the user data of the reference user group in the data feature dimension, so as to characterize a degree of dispersion of the user data of the reference user group in the data feature dimension; The weight value of the data feature dimension is determined according to the inverse of the standard deviation of the user data of the reference user group in the multiple data feature dimensions; the weight value of the data feature dimension is positively correlated with the inverse of the standard deviation of the user data of the reference user group in the data feature dimension.

3. The method according to claim 1, characterized in that The occupancy degree is high occupancy or low occupancy; the occupancy index of the occupancy degree of invalid users in the user group to be tested under the data feature dimension is the probability that the occupancy degree of invalid users in the user group to be tested under the data feature dimension is high occupancy or low occupancy.

4. The method according to claim 1, characterized in that The step of determining, for each data feature dimension of the plurality of data feature dimensions, a similarity index of data distribution between the user data of the reference user group and the user data of the user group to be tested in the data feature dimension includes: For each data feature dimension of the plurality of data feature dimensions, determining a plurality of subgroups of categories under the data feature dimension; Determine the number of users of each of the subgroups of the reference user group and the subgroups of the user group to be tested according to the user data of the reference user group and the user data of the user group to be tested; According to the number of users in each of the subgroups of the multiple categories of the reference user group and the user group to be tested, a similarity index of data distribution of user data of the reference user group and user data of the user group to be tested in the data feature dimension is determined.

5. The method according to any one of claims 1 to 4, characterized in that: The similarity index of the data distribution under the data feature dimension is: under the data feature dimension, the Pearson correlation coefficient or the cosine similarity between the user data of the reference user group and the user data of the user group to be tested.

6. The method according to any one of claims 1 to 4, characterized in that: The multiple data feature dimensions include at least one of the following: the user's age group; the user's ID number location; the user's mobile phone number location; the user's bank card number location; the user's Internet Protocol IP address location.

7. A device for determining the degree of possession of invalid users, characterized in that: include: An acquisition module, used to acquire user data of a reference user group and user data of a user group to be tested; The reference user group is an invalid user group or a valid user group; the user data of the reference user group and the user data of the user group to be tested both include data of multiple data feature dimensions; A processing module is used to determine, for each data feature dimension in a plurality of data feature dimensions, a similarity index of data distribution between user data of the reference user group and user data of the user group to be tested under the data feature dimension; determine, according to the similarity index of each data feature dimension in the plurality of data feature dimensions, an occupation index of the degree of occupation of invalid users of the user group to be tested under the data feature dimension; the occupation index is used to characterize the degree of occupation of invalid users of the user group to be tested under the data feature dimension; determine, according to the occupation index of the degree of occupation of invalid users of the user group to be tested under each data feature dimension in the plurality of data feature dimensions and the weight value of the data feature dimension; the weight value of each data feature dimension in the plurality of data feature dimensions is negatively correlated with the degree of dispersion of user data of the reference user group under the data feature dimension; determine, according to the occupation index of the degree of occupation of invalid users of the user group to be tested under the plurality of data feature dimensions, the degree of occupation of invalid users of the user group to be tested.

8. A computer device, characterized in that: The method comprises a program or an instruction, and when the program or the instruction is executed, the method according to any one of claims 1 to 6 is performed.

9. A storage medium, characterized in that: The method comprises a program or an instruction, and when the program or the instruction is executed, the method according to any one of claims 1 to 6 is performed.

Images