Recognition method and device for point abnormal consumption, and storage medium

By performing multi-grained clustering and screening of the historical behavior data of member users, we can identify abnormal consumption behaviors of member points in the cosmetics industry, and solve the problems of high misjudgment rate and low efficiency in the existing technology, achieving high accuracy and high efficiency detection.

CN118917891APending Publication Date: 2024-11-08GUANGZHOU MARUMI NETWORK TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410944714.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-08

Smart Images

  • Figure CN118917891A_ABST
    Figure CN118917891A_ABST
Patent Text Reader

Abstract

The invention discloses an integral abnormal consumption identification method and device, and a storage medium. The method comprises the steps of querying historical behavior data recorded when a member user generates a consumption behavior; extracting a plurality of pieces of historical behavior feature information, a plurality of purchase indexes and a plurality of periodic consumption indexes from the historical behavior data; the purchase index represents an index for purchasing a product, and the consumption index represents an index for consuming member points exchanged from the product; clustering of multiple granularities is carried out on the member users according to the multiple pieces of historical behavior feature information, and first candidate users which are overlapped in multiple granularities and represent abnormal consumption behaviors are obtained; screening out a plurality of second candidate users with overlapped purchase indexes and abnormal consumption behaviors from the first candidate users; and screening out target users with abnormal consumption behaviors from the second candidate users according to the plurality of consumption indexes. According to the embodiment, the detection comprehensiveness can be improved, so that the accuracy of abnormal consumption behavior detection is effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of computer processing, and in particular to a method, device and storage medium for identifying abnormal consumption of points. Background Art

[0002] In fast-moving consumer goods industries such as cosmetics, operations usually use the method of generating member points based on member users' consumption records and redeeming member points for member benefits to provide value-added services to member users.

[0003] Furthermore, some product packaging includes a QR code for redeeming membership points. After purchasing the product, the user can redeem membership points by scanning the QR code inside the product, and then use the membership points to redeem products or enjoy other membership benefits.

[0004] In fast-moving consumer goods industries such as cosmetics, there is a distribution model in which agents or store staff may redeem membership points in large quantities through unconventional means, and use membership points to redeem membership benefits for sale, thereby damaging the interests of the company and consumers.

[0005] In order to limit the unusual large-scale redemption of member points, a threshold is usually set in the member points program. If the threshold is exceeded, the redemption of member points will be triggered to fail.

[0006] However, there will be a certain misjudgment rate when setting thresholds for member points, especially for some high-value and loyal member users. Therefore, technical personnel are required to manually analyze the behavior of member users and modify the thresholds accordingly. However, in fast-moving consumer goods industries such as cosmetics, products are updated quickly and activities are more frequent, so the thresholds need to be modified frequently, resulting in lower overall efficiency. Summary of the invention

[0007] In view of this, the present invention provides a method, device and storage medium for identifying abnormal consumption of points, so as to improve the accuracy and efficiency of identifying irregular large-scale redemption of member points.

[0008] A first aspect of the present invention provides a method for identifying abnormal point consumption, comprising:

[0009] Query the historical behavior data recorded when the member user generates consumption behavior; the consumption behavior includes purchasing products and consuming member points redeemed from the products;

[0010] Extracting a plurality of historical behavior feature information, a plurality of purchase indicators and a plurality of periodic consumption indicators from the historical behavior data; the purchase indicator represents an indicator for purchasing a product, and the consumption indicator represents an indicator for consuming member points redeemed from the product;

[0011] Clustering the member users at multiple granularities according to the multiple historical behavior feature information to obtain first candidate users with multiple overlapping granularities and indicating abnormal consumption behavior;

[0012] Filter out a plurality of second candidate users whose purchase indicators overlap and whose consumption behaviors are abnormal from the first candidate users;

[0013] The target user with abnormal consumption behavior is screened out from the second candidate users according to the multiple consumption indicators.

[0014] A second aspect of the present invention provides a device for identifying abnormal point consumption, comprising:

[0015] A historical behavior data query module is used to query the historical behavior data recorded when the member user generates consumption behavior; the consumption behavior includes purchasing products and consuming member points redeemed from the products;

[0016] A historical behavior data statistics module, used to extract multiple historical behavior feature information, multiple purchase indicators and multiple periodic consumption indicators from the historical behavior data; the purchase indicator represents an indicator of purchasing a product, and the consumption indicator represents an indicator of consuming member points redeemed from the product;

[0017] A first candidate user clustering module is used to cluster the member users at multiple granularities according to the multiple historical behavior feature information, and obtain first candidate users at multiple overlapping granularities that indicate abnormal consumption behavior;

[0018] A second candidate user screening module, used to screen out a plurality of second candidate users whose purchase indicators overlap and whose consumption behaviors are abnormal from the first candidate users;

[0019] A target user screening module is used to screen out the target user with abnormal consumption behavior from the second candidate users based on the multiple consumption indicators.

[0020] A third aspect of the present invention provides an electronic device, the electronic device comprising:

[0021] at least one processor; and

[0022] a memory communicatively connected to the at least one processor; wherein,

[0023] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for identifying abnormal consumption of points as described in the first aspect above.

[0024] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for identifying abnormal consumption of points as described in the first aspect above.

[0025] A fifth aspect of the present invention provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the method for identifying abnormal consumption of points as described in the first aspect above.

[0026] In this embodiment, the historical behavior data recorded by the member user when the consumption behavior is generated is queried; the consumption behavior includes purchasing products and consuming member points exchanged from products; multiple historical behavior feature information, multiple purchase indicators and multiple periodic consumption indicators are extracted from the historical behavior data; the purchase indicator represents the indicator of purchasing products, and the consumption indicator represents the indicator of member points exchanged from products; the member users are clustered at multiple granularities based on multiple historical behavior feature information to obtain the first candidate users with multiple overlapping granularities and indicating abnormal consumption behavior; the second candidate users with multiple overlapping purchase indicators and indicating abnormal consumption behavior are screened out from the first candidate users; the target users with abnormal consumption behavior are screened out from the second candidate users based on multiple consumption indicators. This embodiment runs through the entire life cycle of consumption behavior, and detects abnormal member users in multiple stages from the overall level of consumption behavior, the sub-level of purchased products, and the sub-level of member points exchanged from products, which can improve the comprehensiveness of detection, thereby effectively improving the accuracy of detecting abnormal consumption behavior. The entire detection process has a high degree of automation, and can dynamically adapt to the product update speed and activities of fast-moving consumer goods industries such as cosmetics, reduce the degree of participation of technical personnel, and effectively improve the overall efficiency of detecting abnormal consumption behavior.

[0027] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0029] Figure 1 This is a flow chart of a method for identifying abnormal point consumption provided in Example 1 of the present invention.

[0030] Figure 2This is an example diagram of summary information of a member user provided in the first embodiment of the present invention.

[0031] Figure 3 This is an example diagram of detailed information of a member user provided in the first embodiment of the present invention.

[0032] Figure 4 This is an example diagram of prohibiting the redemption of member points provided in the first embodiment of the present invention.

[0033] Figure 5 It is a structural diagram of a device for identifying abnormal point consumption provided in Embodiment 2 of the present invention.

[0034] Figure 6 It is a structural schematic diagram of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0036] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can cover sequential implementations other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0037] Embodiment 1

[0038] See also Figure 1 , shows a flow chart of a method for identifying abnormal point consumption provided by the first embodiment of the present invention, the method can be executed by a device for identifying abnormal point consumption, the device for identifying abnormal point consumption can be implemented in the form of hardware and / or software, and the device for identifying abnormal point consumption can be configured in an electronic device. Figure 1 As shown, the method includes:

[0039] Step 101: Query the historical behavior data recorded when the member user generates consumption behavior.

[0040] In this embodiment, member users with abnormal consumption behaviors can be mined from all member users or active member users at intervals of a period (such as one month).

[0041] Then, at each interval, you can use psycopg2 (used to interact with the PostgreSQL database) and other functions to query the database for historical behavior data recorded by member users when they generate consumption behaviors, which is sorted out from the server's log files. After that, use pandas (for data processing) and numpy (for numerical calculation) and other functions to process the historical behavior data in order to discover member users with abnormal consumption behaviors.

[0042] Among them, consumption behavior includes purchasing products and consuming membership points redeemed from products.

[0043] Furthermore, the membership points redeemed from consumption of products can be used to redeem designated products, coupons and other membership benefits.

[0044] Step 102: extract multiple historical behavior feature information, multiple purchase indicators, and multiple periodic consumption indicators from the historical behavior data.

[0045] In this embodiment, historical behavior data may be statistically analyzed to generate behavior information in different dimensions for classifying consumption behaviors as normal or abnormal.

[0046] On the one hand, statistical methods can be used to extract multiple features from historical behavior data and record them as historical behavior feature information. Then, the historical behavior feature information can include both the characteristics of purchasing products and the characteristics of member points redeemed from products.

[0047] On the other hand of the behavior information, a statistical method can be used to calculate a plurality of purchase indicators from the historical behavior data, wherein the purchase indicator represents an indicator for purchasing a product.

[0048] In another aspect of the behavior information, a statistical method can be used to calculate a plurality of periodic consumption indicators from the historical behavior data, wherein the consumption indicator represents an indicator of the consumption of member points redeemed from products.

[0049] Step 103: Cluster the member users at multiple granularities based on multiple pieces of historical behavior feature information to obtain first candidate users with multiple overlapping granularities and indicating abnormal consumption behavior.

[0050] In practical applications, different historical behavior feature information is at different levels of semantics from macro to micro when characterizing the characteristics of user consumption behavior. By abstracting these semantics into granularities, member users can be clustered at multiple granularities based on multiple historical behavior feature information.

[0051] For each granularity, member users are clustered using some historical behavior feature information, and the clustering results at each granularity are merged (including deduplication) to obtain the first candidate users who overlap multiple granularities and represent abnormal consumption behavior.

[0052] In one embodiment of the present invention, step 103 may include the following steps:

[0053] Step 1031: Divide the multiple historical behavior characteristic information into first behavior characteristic information and second behavior characteristic information.

[0054] When two granularities are set, multiple historical behavior feature information can be divided into two dimensions, namely, first behavior feature information and second behavior feature information.

[0055] Among them, the semantic granularity of the first behavior characteristic information in consumption behavior is greater than the semantic granularity of the second behavior characteristic information in consumption behavior, that is, the first behavior characteristic information can reflect the semantic macro characteristics of member users' consumption behavior, and the second behavior characteristic information can reflect the semantic micro characteristics of member users' consumption behavior.

[0056] Exemplarily, the first behavior characteristic information includes the historical total consumption amount of purchased products, the historical total number of purchased products, and the historical total redemption amount of member points.

[0057] The second behavior characteristic information includes the average consumption amount of products purchased in the current period, the average number of products purchased in the current period, and the joint rate of purchased products.

[0058] Step 1032: cluster the member users according to the first behavior characteristic information to obtain a first user cluster indicating abnormal consumption behavior.

[0059] In this embodiment, a preset clustering algorithm can be called to cluster member users at a macro level based on the first behavior feature information to obtain multiple first user clusters, and a first user cluster representing abnormal consumption behavior is selected from all first user clusters.

[0060] In one clustering method, the first behavior characteristic information includes the historical total consumption amount of purchased products, the historical total number of purchased products and the historical total redemption amount of member points; in this method, two first user clusters are initialized, one of which is used to collect member users with normal consumption behavior, and the other first user cluster is used to collect member users with abnormal consumption behavior.

[0061] Taking the historical total consumption amount of products, the historical total number of purchased products and the historical total redemption amount of member points as points in the multidimensional space, the member users are clustered using K-means under two first user clusters.

[0062] If K-means clustering is completed, the first abnormal indicators representing the historical total consumption amount, the historical total number of orders and the historical total amount of redemption are respectively counted in the two first user clusters, wherein the first abnormal indicators include the center point of the first user cluster, the median value of all points in the first user cluster, and so on.

[0063] The first abnormality indicators in the two first user clusters are compared to determine that the consumption behavior of the member user in the first user cluster with the highest first abnormality indicator is abnormal.

[0064] Step 1033: cluster the member users according to the second behavior characteristic information to obtain a second user cluster indicating abnormal consumption behavior.

[0065] In this embodiment, a preset clustering algorithm can be called to cluster member users at a micro level based on the second behavior feature information to obtain multiple second user clusters, and a second user cluster representing abnormal consumption behavior is selected from all second user clusters.

[0066] In one clustering method, the second behavior characteristic information includes the average consumption amount of products purchased in the current period, the average number of products purchased in the current period, and the joint rate of purchased products; in this method, two second user clusters are initialized, one of which is used to collect member users with normal consumption behavior, and the other second user cluster is used to collect member users with abnormal consumption behavior.

[0067] Taking the average consumption amount, average number of orders and joint rate as points in the multidimensional space, the member users are clustered using the average consumption amount, average number of orders and joint rate under two second user clusters using K-means.

[0068] If K-means clustering is completed, the second abnormal indicators representing the average consumption amount, average number of orders and linkage rate are respectively counted in the two second user clusters, wherein the second abnormal indicators include the center point of the second user cluster, the median value of all points in the second user cluster, and so on.

[0069] The second abnormality indicators in the two second user clusters are compared to determine that the consumption behavior of the member user in the second user cluster with the highest second abnormality indicator is abnormal.

[0070] Step 1034: merge the first user cluster and the second user cluster to obtain a first candidate user representing abnormal consumption behavior.

[0071] The first user cluster indicating abnormal consumption behavior at the macro level and the second user cluster indicating abnormal consumption behavior at the micro level are merged (including deduplication), and the member user after the merger is marked as the first candidate user indicating abnormal consumption behavior.

[0072] Step 104: Filter out second candidate users whose multiple purchase indicators overlap and whose consumption behaviors are abnormal from the first candidate users.

[0073] In practical applications, different purchase indicators characterize the characteristics of users purchasing products in different dimensions. Therefore, first candidate users with abnormal consumption behaviors can be screened out from all first candidate users based on multiple purchase indicators, and the first candidate users with abnormal consumption behaviors under each indicator can be merged (including deduplication) to obtain second candidate users with abnormal consumption behaviors under overlapping multiple purchase indicators.

[0074] In one embodiment of the present invention, the purchase index includes the volatility of the linkage rate (such as the standard deviation), the volatility of the purchase cycle (such as the standard deviation), the average number of units of purchased products (i.e., the number of SKUs), and the average number of series of purchased products; in this embodiment, step 104 may include the following steps:

[0075] Step 1041: Mark the first candidate user with a first abnormal label indicating abnormal consumption behavior according to the fluctuation rate of the linkage rate.

[0076] In this embodiment, anomaly detection may be performed on the first candidate users based solely on the volatility of the linkage rate, and first anomaly labels indicating abnormal consumption behaviors may be marked on some first candidate users.

[0077] In an abnormality detection method, K-means clustering is performed on the first candidate users using the volatility of the linkage rate to obtain multiple (at least two) third user clusters.

[0078] A third abnormality index characterizing the volatility of the joint rate is statistically collected in each of the plurality of third user clusters, wherein the third abnormality index includes a center point of the third user cluster, a median value of all points in the third user cluster, and the like.

[0079] The third abnormality indicators in the plurality of third user clusters are compared to determine a third user cluster having the highest third abnormality indicator.

[0080] For the first candidate user in the third user cluster with the highest third abnormal index, the volatility of its linkage rate is compared with a preset first threshold, and the historical total consumption amount of its purchased products is compared with a preset second threshold.

[0081] In fast-moving consumer goods industries such as cosmetics, member users usually have relatively fixed combinations of products they purchase. For example, they will not make repeated purchases of products in the same category, such as cleansers, water, essences, and creams.

[0082] If the volatility of the linkage rate of a first candidate user is greater than or equal to a preset first threshold, and the total historical consumption amount of the purchased product is greater than or equal to a preset second threshold, the first candidate user is marked with a first abnormal label indicating abnormal consumption behavior.

[0083] Step 1042: Mark the first candidate user with a second abnormal label indicating abnormal consumption behavior based on the fluctuation rate of the purchase cycle.

[0084] In this embodiment, anomaly detection may be performed on the first candidate users based solely on the volatility of the purchase cycle, and a second anomaly label indicating abnormal consumption behavior may be marked on some first candidate users.

[0085] In an abnormality detection method, K-means clustering is performed on the first candidate users using the volatility of the purchase cycle to obtain multiple (at least two) fourth user clusters.

[0086] Fourth abnormal indicators representing volatility of the purchase cycle are statistically analyzed in each of the plurality of fourth user clusters, wherein the fourth abnormal indicators include a center point of the fourth user cluster, a median value of all points in the fourth user cluster, and the like.

[0087] The fourth abnormality indicators in the plurality of fourth user clusters are compared to determine a fourth user cluster having the highest fourth abnormality indicator.

[0088] For the first candidate user in the fourth user cluster having the highest fourth abnormality index, the volatility of the purchase cycle is compared with a preset third threshold.

[0089] In fast-moving consumer goods industries such as cosmetics, member users consume products at a relatively fixed rate, and therefore have a relatively stable repurchase cycle for purchasing products.

[0090] If the volatility of the purchase cycle of a first candidate user is greater than or equal to a preset third threshold, a second abnormal label indicating abnormal consumption behavior is marked on the first candidate user.

[0091] Step 1043: Mark the first candidate user with a third abnormal label indicating abnormal consumption behavior based on the average number of items purchased.

[0092] In this embodiment, anomaly detection may be performed on the first candidate users based solely on the average number of items purchased, and a third anomaly label indicating abnormal consumption behavior may be marked on some first candidate users.

[0093] In one anomaly detection method, K-means clustering is performed on the first candidate users using the average number of items purchased to obtain multiple (at least two) fifth user clusters.

[0094] A fifth abnormality index representing the average number of items purchased is respectively counted in multiple fifth user clusters, wherein the fifth abnormality index includes a center point of the fifth user cluster, a median value of all points in the fifth user cluster, and the like.

[0095] The fifth abnormality indicators in the plurality of fifth user clusters are compared to determine a fifth user cluster having the highest fifth abnormality indicator.

[0096] For the first candidate user in the fifth user cluster with the highest fifth abnormality index, the average number of products purchased by the first candidate user is compared with a preset fourth threshold.

[0097] In fast-moving consumer goods industries such as cosmetics, member users tend to purchase relatively fixed items.

[0098] If the average number of items purchased by a first candidate user is greater than or equal to a preset fourth threshold, a third abnormal label indicating abnormal consumption behavior is marked on the first candidate user.

[0099] Step 1044: Mark the first candidate user with a fourth abnormal label indicating abnormal consumption behavior based on the average number of series of purchased products.

[0100] In this embodiment, anomaly detection may be performed on the first candidate users based solely on the average number of series of purchased products, and a fourth anomaly label indicating abnormal consumption behavior may be marked on some first candidate users.

[0101] In an abnormality detection method, K-means clustering is performed on the first candidate user using the average series number of purchased products to obtain multiple (at least two) sixth user clusters.

[0102] A sixth abnormality index representing an average series number of purchased products is statistically analyzed in each of the plurality of sixth user clusters, wherein the sixth abnormality index includes a center point of the sixth user cluster, a median value of all points in the sixth user cluster, and the like.

[0103] The sixth abnormality indicators in the plurality of sixth user clusters are compared to determine a sixth user cluster having the highest sixth abnormality indicator.

[0104] For the first candidate user in the sixth user cluster having the highest sixth abnormality index, the average number of series of products purchased by the first candidate user is compared with a preset fifth threshold.

[0105] In fast-moving consumer goods industries such as cosmetics, member users have a relatively fixed series of products they purchase.

[0106] If the average number of series of purchased products of a first candidate user is greater than or equal to a preset fifth threshold, a fourth abnormal label indicating abnormal consumption behavior is marked on the first candidate user.

[0107] Step 1045: merge the first candidate users marked with the first abnormal label, the second abnormal label, the third abnormal label, and the fourth abnormal label as the second candidate users indicating abnormal consumption behavior.

[0108] The first candidate user marked with the first abnormal label, the first candidate user marked with the second abnormal label, the first candidate user marked with the third abnormal label, and the first candidate user marked with the fourth abnormal label are merged (including deduplication) as the second candidate user indicating abnormal consumption behavior.

[0109] Step 105: Filter out target users with abnormal consumption behaviors from the second candidate users based on multiple consumption indicators.

[0110] In this embodiment, different consumption indicators characterize the characteristics of users' consumption of membership points in different dimensions. Therefore, second candidate users with abnormal consumption behaviors can be screened out from all second candidate users based on multiple consumption indicators and recorded as target users.

[0111] In the specific implementation, the consumption indicators include the changing speed of the member points balance, the usage ratio of the member points and the redemption frequency of the member points; among them, the changing speed of the member points balance is the difference between the member points balance at the end of the current cycle and the member points balance at the beginning of the current cycle divided by the cycle, the usage ratio of the member points is the ratio between the total redemption amount of the member points and the total consumption amount, and the redemption frequency of the member points is the ratio between the number of times the member points are consumed to redeem member benefits and the cycle.

[0112] Then, two seventh user clusters may be initialized, wherein one seventh user cluster is used to collect member users with normal consumption behaviors, and the other seventh user cluster is used to collect member users with abnormal consumption behaviors.

[0113] Taking the changing speed of the member points balance, the usage ratio of the member points and the redemption frequency of the member points as points in the multidimensional space, the changing speed of the member points balance, the usage ratio of the member points and the redemption frequency of the member points are used to perform K-means clustering on the second candidate user under the two seventh user clusters.

[0114] If K-means clustering is completed, the seventh abnormality indicators representing the changing speed of the member points balance, the usage ratio of the member points and the redemption frequency of the member points are respectively counted in the two seventh user clusters, where the seventh abnormality indicators include the center point of the seventh user cluster, the median value of all points in the seventh user cluster, and so on.

[0115] The seventh abnormality indicators of the two seventh user clusters are compared to determine the seventh user cluster with the highest seventh abnormality indicator.

[0116] The second candidate user in the seventh user cluster with the highest seventh abnormal index has a relatively high change rate of the member points balance, a usage ratio of the member points, and a redemption frequency of the member points, and can be marked as a target user with abnormal consumption behavior.

[0117] like Figure 2 and Figure 3 As shown, the target users can be marked with designated labels, such as "group buying member", etc., pending in-depth investigation to ensure platform order and member rights.

[0118] In this embodiment, multiple K-means clustering is used in the process of screening target users. The metrics.calinski_harabasz_score function can be used to evaluate the clustering effect of each K-means clustering. The metrics.calinski_harabasz_score function includes the following concepts:

[0119] 1. Calculate the between-cluster dispersion: It indicates the distance between different clusters. If the distance between clusters is larger, it means that the different clusters are more dispersed, that is, the sample points within the cluster are more similar, and the differences between clusters are higher.

[0120] 2. Calculate within-cluster dispersion: It indicates the closeness of sample points in the same cluster. If the sample points in a cluster are closer, it means that the sample points in the cluster are more similar and the difference within the cluster is lower.

[0121] 3. Calinski-Harabasz score: The score is calculated based on the ratio of the distance between classes to the distance within classes. The higher the score, the greater the distance between classes relative to the distance within classes, that is, the greater the difference between clusters, the greater the similarity and closeness within clusters, and the better the clustering effect. The clustering results are evaluated by the score.

[0122] When screening out target users, the user identification of the target users (such as member ID, etc.) can be output to the intermediate library. Based on the user identification in the intermediate library, an analysis report is generated for the target users using data such as historical behavior characteristic information, purchase indicators and periodic consumption indicators.

[0123] The analysis report and historical behavior data are pushed to the designated management user (such as user identification representation) for reference, so as to further verify the consumption behavior of the target user.

[0124] Management users can browse the analysis reports and historical behavior data of target users, maintain target users' consumption behavior abnormalities, and also correct target users' consumption behavior to normalize.

[0125] The message of abnormal consumption behavior of the target user is published to the member business system in a read-only manner; the member business system is used to display the abnormal consumption behavior in the client logged in by the target user, and, Figure 4 As shown, it is prohibited to allocate membership points generated by purchasing products to target users. That is, when target users generate membership points by purchasing products, the default newly added membership points are 0, which is convenient for membership operation personnel, agents and shopping guides to manage membership operations.

[0126] In this embodiment, the historical behavior data recorded by the member user when the consumption behavior is generated is queried; the consumption behavior includes purchasing products and consuming member points exchanged from products; multiple historical behavior feature information, multiple purchase indicators and multiple periodic consumption indicators are extracted from the historical behavior data; the purchase indicator represents the indicator of purchasing products, and the consumption indicator represents the indicator of member points exchanged from products; the member users are clustered at multiple granularities based on multiple historical behavior feature information to obtain the first candidate users with multiple overlapping granularities and indicating abnormal consumption behavior; the second candidate users with multiple overlapping purchase indicators and indicating abnormal consumption behavior are screened out from the first candidate users; the target users with abnormal consumption behavior are screened out from the second candidate users based on multiple consumption indicators. This embodiment runs through the entire life cycle of consumption behavior, and detects abnormal member users in multiple stages from the overall level of consumption behavior, the sub-level of purchased products, and the sub-level of member points exchanged from products, which can improve the comprehensiveness of detection, thereby effectively improving the accuracy of detecting abnormal consumption behavior. The entire detection process has a high degree of automation, and can dynamically adapt to the product update speed and activities of fast-moving consumer goods industries such as cosmetics, reduce the degree of participation of technical personnel, and effectively improve the overall efficiency of detecting abnormal consumption behavior.

[0127] Embodiment 2

[0128] See also Figure 5 , shows a schematic diagram of the structure of a device for identifying abnormal point consumption provided by Embodiment 2 of the present invention. Figure 5 As shown, the device comprises:

[0129] The historical behavior data query module 501 is used to query the historical behavior data recorded when the member user generates consumption behavior; the consumption behavior includes purchasing products and consuming member points redeemed from the products;

[0130] The historical behavior data statistics module 502 is used to extract multiple historical behavior feature information, multiple purchase indicators and multiple periodic consumption indicators from the historical behavior data; the purchase indicator represents an indicator of purchasing a product, and the consumption indicator represents an indicator of consuming member points redeemed from the product;

[0131] A first candidate user clustering module 503 is used to cluster the member users at multiple granularities according to the multiple historical behavior feature information, and obtain first candidate users with multiple overlapping granularities indicating abnormal consumption behavior;

[0132] A second candidate user screening module 504 is used to screen out a plurality of second candidate users whose purchase indicators overlap and whose consumption behaviors are abnormal from the first candidate users;

[0133] The target user screening module 505 is used to screen out the target user with abnormal consumption behavior from the second candidate users according to the multiple consumption indicators.

[0134] In one embodiment of the present invention, the first candidate user clustering module 503 includes:

[0135] A historical behavior characteristic information division module, used to divide the plurality of historical behavior characteristic information into first behavior characteristic information and second behavior characteristic information; the semantic granularity of the first behavior characteristic information in the consumption behavior is greater than the semantic granularity of the second behavior characteristic information in the consumption behavior;

[0136] A first behavior characteristic information clustering module, used to cluster the member users according to the first behavior characteristic information to obtain a first user cluster indicating abnormal consumption behavior;

[0137] A second behavior characteristic information clustering module, used for clustering the member users according to the second behavior characteristic information to obtain a second user cluster indicating abnormal consumption behavior;

[0138] The user cluster merging module is used to merge the first user cluster and the second user cluster to obtain a first candidate user representing the abnormal consumption behavior.

[0139] In one embodiment of the present invention, the first behavior characteristic information includes the historical total consumption amount of purchased products, the historical total number of purchased products, and the historical total amount of member points redeemed;

[0140] The first behavior feature information clustering module includes:

[0141] A first user cluster initialization module, used for initializing two first user clusters;

[0142] A first user clustering module, configured to perform K-means clustering on the member users under two first user clusters using the historical total consumption amount, the historical total number of orders and the historical total amount of redemption;

[0143] A first abnormality indicator statistics module is used for respectively counting the first abnormality indicators representing the historical total consumption amount, the historical total number of orders and the historical total amount of exchange in the two first user clusters if K-means clustering is completed;

[0144] The first user cluster anomaly determination module is used to determine that the consumption behavior of the member user in the first user cluster with the highest first anomaly index is abnormal.

[0145] In one embodiment of the present invention, the second behavior characteristic information includes the average consumption amount of the product purchased in the current period, the average number of units of the product purchased in the current period, and the joint rate of the purchased product;

[0146] The second behavior feature information clustering module includes:

[0147] A second user cluster initialization module, used for initializing two second user clusters;

[0148] A second user cluster clustering module, configured to perform K-means clustering on the member users under two second user clusters using the average consumption amount of the purchased products, the average number of units of the purchased products and the joint rate of the purchased products;

[0149] A second abnormality indicator statistics module, for respectively counting the second abnormality indicators representing the average consumption amount, the average number of orders and the linkage rate in the two second user clusters if K-means clustering is completed;

[0150] The second user cluster anomaly determination module is used to determine that the consumption behavior of the member user in the second user cluster with the highest second anomaly index is abnormal.

[0151] In one embodiment of the present invention, the purchase index includes the volatility of the linkage rate, the volatility of the purchase cycle, the average number of units of purchased products, and the average number of series of purchased products;

[0152] The second candidate user screening module 504 includes:

[0153] A first abnormal label marking module, configured to mark the first candidate user with a first abnormal label indicating that the consumption behavior is abnormal according to the fluctuation rate of the joint rate;

[0154] A second abnormal label marking module, used for marking the first candidate user with a second abnormal label indicating that the consumption behavior is abnormal according to the volatility of the purchase cycle;

[0155] A third abnormal label marking module, configured to mark the first candidate user with a third abnormal label indicating that the consumption behavior is abnormal according to the average number of purchased products;

[0156] A fourth abnormal label marking module, used for marking the first candidate user with a fourth abnormal label indicating that the consumption behavior is abnormal according to the average series number of the purchased products;

[0157] The abnormal label merging module is used to merge the first candidate user marked with the first abnormal label, the second abnormal label, the third abnormal label and the fourth abnormal label as the second candidate user indicating that the consumption behavior is abnormal.

[0158] In one embodiment of the present invention, the first abnormal label marking module includes:

[0159] A third user clustering module, configured to perform K-means clustering on the first candidate users using the volatility of the linkage rate to obtain a plurality of third user clusters;

[0160] A third abnormality indicator statistics module, used for respectively counting a third abnormality indicator representing the volatility of the joint rate in the plurality of third user clusters;

[0161] The first abnormal label setting module is used to mark the first candidate user in the third user cluster with the highest third abnormal index with a first abnormal label indicating that the consumption behavior is abnormal if the volatility of the linkage rate is greater than or equal to a preset first threshold and the historical total consumption amount of purchased products is greater than or equal to a preset second threshold.

[0162] In one embodiment of the present invention, the second abnormal label marking module includes:

[0163] a fourth user cluster clustering module, configured to perform K-means clustering on the first candidate users using the volatility of the purchase cycle to obtain a plurality of fourth user clusters;

[0164] A fourth abnormal index statistical module, used for respectively counting fourth abnormal indexes representing the volatility of the purchase cycle in a plurality of the fourth user clusters;

[0165] The second abnormal label setting module is used to mark the first candidate user in the fourth user cluster with the highest fourth abnormal index with a second abnormal label indicating that the consumption behavior is abnormal if the volatility of the purchase cycle is greater than or equal to a preset third threshold.

[0166] In one embodiment of the present invention, the third abnormal label marking module includes:

[0167] a fifth user cluster clustering module, configured to perform K-means clustering on the first candidate users using the average number of purchased products to obtain a plurality of fifth user clusters;

[0168] A fifth abnormality indicator statistics module, used to respectively count the fifth abnormality indicator representing the average number of purchased products in the plurality of fifth user clusters;

[0169] The third abnormal label setting module is used to mark the first candidate user in the fifth user cluster with the highest fifth abnormal index with a third abnormal label indicating that the consumption behavior is abnormal if the average number of purchased products is greater than or equal to a preset fourth threshold.

[0170] In one embodiment of the present invention, the fourth abnormal label marking module includes:

[0171] a sixth user cluster clustering module, configured to perform K-means clustering on the first candidate users using the average series number of purchased products to obtain a plurality of sixth user clusters;

[0172] A sixth abnormality indicator statistical module, used for respectively counting the sixth abnormality indicator representing the average series number of the purchased products in the plurality of the sixth user clusters;

[0173] The fourth abnormal label setting module is used to mark the first candidate user in the sixth user cluster with the highest sixth abnormal index with a fourth abnormal label indicating that the consumption behavior is abnormal if the average number of series of purchased products is greater than or equal to a preset fifth threshold.

[0174] In one embodiment of the present invention, the consumption index includes the change rate of the member points balance, the usage ratio of the member points and the redemption frequency of the member points;

[0175] The target user screening module 505 includes:

[0176] A seventh user cluster initialization module, used for initializing two seventh user clusters;

[0177] a seventh user clustering module, configured to perform K-means clustering on the second candidate user in two of the seventh user clusters using the change rate of the member points balance, the usage ratio of the member points and the redemption frequency of the member points;

[0178] a seventh abnormality indicator statistics module, configured to, if K-means clustering is completed, respectively count the seventh abnormality indicators representing the change rate of the member points balance, the usage ratio of the member points, and the redemption frequency of the member points in the two seventh user clusters;

[0179] The target user marking module is used to mark the second candidate user in the seventh user cluster with the highest seventh abnormality index as a target user with abnormal consumption behavior.

[0180] In one embodiment of the present invention, the device further comprises:

[0181] A target user output module, used for outputting the user identification of the target user to the intermediate library;

[0182] An analysis report generating module, used to generate an analysis report for the target user according to the user identifier in the intermediate library;

[0183] A user data push module, used to push the analysis report and the historical behavior data to a designated management user and verify the consumption behavior of the target user;

[0184] The abnormal message publishing module is used to publish the message of the abnormal consumption behavior of the target user to the member business system in a read-only manner; the member business system is used to display the abnormal consumption behavior in the client logged in by the target user, and to prohibit the member points generated by purchasing products from being allocated to the target user.

[0185] The device for identifying abnormal point consumption provided in the embodiment of the present invention can execute the method for identifying abnormal point consumption provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method for identifying abnormal point consumption.

[0186] Embodiment 3

[0187] See also Figure 6 , shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0188] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0189] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0190] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the method for identifying abnormal consumption of points.

[0191] In some embodiments, the method for identifying abnormal consumption of points may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for identifying abnormal consumption of points described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the method for identifying abnormal consumption of points in any other appropriate manner (e.g., by means of firmware).

[0192] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0193] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0194] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0195] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0196] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0197] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0198] Embodiment 4

[0199] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the method for identifying abnormal consumption of points as provided in any embodiment of the present invention.

[0200] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0201] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0202] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying abnormal point consumption, characterized in that: include: Query the historical behavior data recorded when the member user generates consumption behavior; the consumption behavior includes purchasing products and consuming member points redeemed from the products; Extracting a plurality of historical behavior feature information, a plurality of purchase indicators and a plurality of periodic consumption indicators from the historical behavior data; the purchase indicator represents an indicator for purchasing a product, and the consumption indicator represents an indicator for consuming member points redeemed from the product; Clustering the member users at multiple granularities according to the multiple historical behavior feature information to obtain first candidate users with multiple overlapping granularities and indicating abnormal consumption behavior; Filter out a plurality of second candidate users whose purchase indicators overlap and whose consumption behaviors are abnormal from the first candidate users; The target user with abnormal consumption behavior is screened out from the second candidate users according to the multiple consumption indicators.

2. The method according to claim 1, characterized in that The clustering of the member users at multiple granularities according to the multiple historical behavior feature information to obtain the first candidate users at multiple overlapping granularities and indicating abnormal consumption behavior includes: Dividing the plurality of historical behavior characteristic information into first behavior characteristic information and second behavior characteristic information; the semantic granularity of the first behavior characteristic information in the consumption behavior is greater than the semantic granularity of the second behavior characteristic information in the consumption behavior; Clustering the member users according to the first behavior characteristic information to obtain a first user cluster indicating abnormal consumption behavior; Clustering the member users according to the second behavior characteristic information to obtain a second user cluster indicating that the consumption behavior is abnormal; The first user cluster and the second user cluster are merged to obtain a first candidate user representing the abnormal consumption behavior.

3. The method according to claim 2, characterized in that The first behavior characteristic information includes the historical total consumption amount of purchased products, the historical total number of purchased products, and the historical total amount of member points redeemed; The clustering of the member users according to the first behavior characteristic information to obtain a first user cluster indicating abnormal consumption behavior includes: Initialize two first user clusters; Perform K-means clustering on the member users under the two first user clusters using the historical total consumption amount, the historical total number of orders and the historical total amount of redemption; If the K-means clustering is completed, the first abnormal indicators representing the historical total consumption amount, the historical total number of orders and the historical total amount of exchange are respectively counted in the two first user clusters; It is determined that the consumption behavior of the member user in the first user cluster having the highest first abnormality index is abnormal.

4. The method according to claim 2, characterized in that: The second behavior characteristic information includes the average consumption amount of the product purchased in the current period, the average number of units of the product purchased in the current period, and the joint rate of the purchased product; The clustering of the member users according to the second behavior characteristic information to obtain a second user cluster indicating abnormal consumption behavior includes: Initialize two second user clusters; Perform K-means clustering on the member users under two second user clusters using the average consumption amount of the purchased products, the average number of units of the purchased products and the joint rate of the purchased products; If the K-means clustering is completed, the second abnormality index representing the average consumption amount, the average number of orders and the linkage rate is respectively counted in the two second user clusters; It is determined that the consumption behavior of the member user in the second user cluster having the highest second abnormality index is abnormal.

5. The method according to claim 1, characterized in that The purchase index includes the volatility of the joint rate, the volatility of the purchase cycle, the average number of units of purchased products, and the average number of series of purchased products; The step of selecting a plurality of second candidate users whose purchase indicators overlap and indicate abnormal consumption behaviors from the first candidate users includes: Marking the first candidate user with a first abnormal label indicating that the consumption behavior is abnormal according to the volatility of the association rate; Marking the first candidate user with a second abnormal label indicating that the consumption behavior is abnormal according to the volatility of the purchase cycle; Marking the first candidate user with a third abnormal label indicating that the consumption behavior is abnormal according to the average number of purchased products; Marking the first candidate user with a fourth abnormal label indicating that the consumption behavior is abnormal according to the average series number of the purchased products; The first candidate users marked with the first abnormal label, the second abnormal label, the third abnormal label and the fourth abnormal label are merged as the second candidate users indicating that the consumption behavior is abnormal.

6. The method according to claim 5, characterized in that The step of marking the first candidate user with a first abnormal label indicating that the consumption behavior is abnormal according to the fluctuation rate of the joint rate includes: Perform K-means clustering on the first candidate users using the volatility of the linkage rate to obtain a plurality of third user clusters; Counting a third abnormality indicator representing the volatility of the linkage rate in each of the plurality of third user clusters; For the first candidate user in the third user cluster with the highest third abnormal index, if the volatility of the linkage rate is greater than or equal to a preset first threshold, and the total historical consumption amount of purchased products is greater than or equal to a preset second threshold, then a first abnormal label indicating that the consumption behavior is abnormal is marked on the first candidate user; The marking of the first candidate user with a second abnormal label indicating that the consumption behavior is abnormal according to the fluctuation rate of the purchase cycle includes: Performing K-means clustering on the first candidate users using the volatility of the purchase cycle to obtain a plurality of fourth user clusters; Counting fourth abnormality indicators representing the volatility of the purchase cycle in each of the plurality of fourth user clusters; For the first candidate user in the fourth user cluster with the highest fourth abnormality index, if the volatility of the purchase cycle is greater than or equal to a preset third threshold, marking the first candidate user with a second abnormality label indicating that the consumption behavior is abnormal; The marking of the first candidate user with a third abnormal label indicating that the consumption behavior is abnormal according to the average number of purchased products includes: Perform K-means clustering on the first candidate users using the average number of purchased products to obtain multiple fifth user clusters; Counting a fifth abnormality indicator representing the average number of purchased products in each of the plurality of fifth user clusters; For the first candidate user in the fifth user cluster having the highest fifth abnormality index, if the average number of purchased products is greater than or equal to a preset fourth threshold, a third abnormality label indicating that the consumption behavior is abnormal is marked on the first candidate user; The marking of the first candidate user with a fourth abnormal label indicating that the consumption behavior is abnormal according to the average series number of the purchased products includes: Performing K-means clustering on the first candidate users using the average series number of the purchased products to obtain a plurality of sixth user clusters; Counting a sixth abnormality indicator representing an average number of series of purchased products in each of the plurality of the sixth user clusters; For the first candidate user in the sixth user cluster having the highest sixth abnormality index, if the average number of series of purchased products is greater than or equal to a preset fifth threshold, the first candidate user is marked with a fourth abnormality label indicating that the consumption behavior is abnormal.

7. The method according to claim 1, characterized in that The consumption indicators include the rate of change of the member points balance, the usage ratio of the member points and the frequency of redemption of the member points; The step of screening out the target user with abnormal consumption behavior from the second candidate users based on the plurality of consumption indicators includes: Initialize two seventh user clusters; Performing K-means clustering on the second candidate user under the two seventh user clusters using the change rate of the member points balance, the usage ratio of the member points and the redemption frequency of the member points; If the K-means clustering is completed, the seventh abnormality index representing the change speed of the member points balance, the usage ratio of the member points and the redemption frequency of the member points is statistically analyzed in the two seventh user clusters respectively; The second candidate user in the seventh user cluster having the highest seventh abnormality index is marked as a target user with abnormal consumption behavior.

8. The method according to any one of claims 1 to 7, characterized in that Also includes: Outputting the user ID of the target user to the intermediate library; Generating an analysis report for the target user according to the user identifier in the intermediate database; Pushing the analysis report and the historical behavior data to a designated management user and verifying the consumption behavior of the target user; The message about the abnormal consumption behavior of the target user is published to the member service system in a read-only manner; the member service system is used to display the abnormal consumption behavior in the client where the target user logs in, and prohibit the member points generated by purchasing products from being allocated to the target user.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for identifying abnormal consumption of points as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for identifying abnormal consumption of points as described in any one of claims 1 to 8 is implemented.