User classification method, device, storage medium and electronic device

Abstract

The embodiment of the present invention discloses a user classification method, device, storage medium and electronic device. The method includes: obtaining object operation data of each user; for any user, determining the competitor of the current object based on the object operation data of the current user, and respectively obtaining feature data of the current user's operations on the current object and the competitor; determining the current user's loyalty index to the current object based on the feature data of the current user's operations on the current object and the competitor; and determining the loyalty classification of each user to the current object based on the ranking of the loyalty index of each user. In this embodiment, there is no need to construct a clustering or regression model, the amount of calculation is small, and the efficiency of loyalty classification is improved. At the same time, by determining the competitor, loyalty classification is carried out in a targeted manner, thereby improving the accuracy of loyalty classification.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of data processing technology, and in particular to a user classification method, device, storage medium and electronic device. Background Art

[0002] At present, as traffic is becoming more and more expensive, everyone is thinking about how to break through the status quo. First of all, we need to start with the existing store loyal users and retain loyal users. The loss of loyal users is a huge loss to the store. An effective method for judging store loyal users is needed. The existing methods for identifying user loyalty are: one is to use the classification method in data mining methods; the other is to use the regression method in data mining methods. These two methods are used to reveal the relationship between user feature attributes and conclusions.

[0003] However, in the process of realizing the present invention, the inventors found that there are at least the following technical problems in the prior art: the classification method of the current data mining method has poor accuracy, and the scale of merchants on the current e-commerce platforms is very large, and it is a lot of work to build a clustering or regression model for each store. Summary of the invention

[0004] The embodiments of the present invention provide a user classification method, device, storage medium and electronic device to improve the accuracy and efficiency of user classification.

[0005] In a first aspect, an embodiment of the present invention provides a user classification method, comprising:

[0006] Get object operation data of each user;

[0007] For any user, determine the competitor of the current object based on the object operation data of the current user, and obtain feature data of the current user's operations on the current object and the competitor respectively; determine the loyalty index of the current user to the current object based on the feature data of the current user's operations on the current object and the competitor;

[0008] Based on the loyalty index ranking of each user, the loyalty classification of each user to the current object is determined.

[0009] In a second aspect, an embodiment of the present invention further provides a user classification device, including:

[0010] An object operation data acquisition module is used to acquire the object operation data of each user;

[0011] A competitor determination module, used for determining, for any user, a competitor of a current object based on the object operation data of the current user;

[0012] A feature data acquisition module, used to respectively acquire feature data of the current user's operations on the current object and the competing object;

[0013] A loyalty index determination module, configured to determine the loyalty index of the current user to the current object based on feature data of the current user's operations on the current object and the competing object;

[0014] The loyalty classification module is used to determine the loyalty classification of each user to the current object based on the loyalty index ranking of each user.

[0015] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a user classification method as provided in any embodiment of the present invention is implemented.

[0016] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a user classification method as provided in any embodiment of the present invention.

[0017] The technical solution provided in this embodiment obtains the object operation data of each user, determines the competitor of the current object for each user, determines the loyalty index of each user to the current object based on the feature data of the current object and the competitor, and determines the loyalty classification of each user to the current object based on the ranking of the loyalty index of each user. In this embodiment, there is no need to build a clustering or regression model, the amount of calculation is small, and the efficiency of loyalty classification is improved. At the same time, by determining the competitor, the loyalty classification is carried out in a targeted manner, thereby improving the accuracy of loyalty classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a flow chart of a user classification method provided in Embodiment 1 of the present invention;

[0019] Figure 2 is a structural diagram of a user classification device provided in Embodiment 2 of the present invention;

[0020] Figure 3 A schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0022] Embodiment 1

[0023] Figure 1 This is a flowchart of a user classification method provided in the first embodiment of the present invention. This embodiment can be applied to the situation of identifying the loyalty of users to various objects and classifying the loyalty of users. The method can be executed by the user classification device provided in the embodiment of the present invention. The user classification device can be implemented by software and / or hardware. The user classification device can be configured on an electronic computing device such as a computer or server. The method specifically includes the following steps:

[0024] S110: Obtain object operation data of each user.

[0025] S120. For any user, determine the competing object of the current object based on the object operation data of the current user, and obtain feature data of the current user's operations on the current object and the competing object respectively; determine the current user's loyalty index to the current object based on the feature data of the current user's operations on the current object and the competing object.

[0026] S130: Determine the loyalty classification of each user to the current object based on the loyalty index ranking of each user.

[0027] In this embodiment, there may be multiple objects of the same type, and there is a competitive relationship between the objects of the same type. User loyalty is a key factor for each object to retain users, and the loss of loyal users is a major factor affecting the operation of each object. The objects here may include but are not limited to applications, stores on e-commerce platforms, anchors, online games, etc.

[0028] Users can operate on each object according to their needs to form object operation data. In some embodiments, the object operation data may be a user browsing log. Taking the object as a store on an e-commerce platform as an example, the object operation data may be operation data formed by users clicking, browsing, purchasing, and evaluating each store. Taking the object as an anchor as an example, the object operation data may be operation data generated by users clicking, evaluating, forwarding, sending interactive messages, and other operations on the live broadcast rooms of each anchor. Among them, the above-mentioned object operation data includes a user operation type identifier and a timestamp corresponding to each user operation.

[0029] For any object, the operation data of each user on the current object can be obtained to obtain each user corresponding to the current object, and the loyalty of each user can be classified based on the user operation data of each user to determine the loyal users of the current object, so as to facilitate the current object to retain loyal users. In this embodiment, the object operation data of each user can be the operation data of a preset time period, such as the operation data of a preset time period before the current moment, and the time period can be one year or one month, etc., which can be determined according to the type of the object.

[0030] Obtain each user corresponding to the current object, and obtain the user operation data of each user, determine the user's loyalty index based on the user operation data of each user, and classify the users according to the loyalty index of each user, wherein the larger the loyalty index, the higher the user's loyalty to the current object, and correspondingly, the smaller the loyalty index, the lower the user's loyalty to the current object.

[0031] In this embodiment, each user can operate on multiple objects, determine the competing object of the current object among the multiple objects operated by the user, and determine the user's loyalty index to the current object through the user's operation data on the current object and the competing object. Among them, the competing object is the object with the same purpose as the current object and the highest similarity among the objects operated by the user. It should be noted that for each user, the competing object of the current object is determined among the objects operated by the user, that is, for different users, the competing object of the current object may be different.

[0032] Optionally, the competing object of the current object is determined based on the object operation data of the current user, including: generating at least one item set of objects based on the operation timestamp of each object in the object operation data of the current user, wherein each item set includes at least one object; determining the associated objects of the current object in each item set, and the association coefficient between the current object and each associated object; and determining the competing object of the current object according to the association coefficient between the current object and each associated object.

[0033] Taking the objects as various stores in the e-commerce platform as an example, users can browse through the recommendation page or the search page determined by keywords. Users can view stores of the same type on the recommendation page and the search page to compare the products and determine the target products to be purchased. Therefore, the objects operated in the same time period are often in a competitive relationship. In this embodiment, the objects operated in each time period are divided into item sets, so that objects with an associated relationship with the current object can be determined from the objects operated in the same time period.

[0034] In this embodiment, the object operation data is divided according to the operation timestamp of each operation data, and a corresponding item set is generated based on each segment of the divided object operation data, and the item set includes each object included in the corresponding object operation data. Optionally, based on the operation timestamp of each object in the object operation data of the current user, at least one item set of the object is generated, including: determining the time interval of adjacent operation data according to the operation timestamp of each object in the object operation data; dividing the object operation data into multiple data sessions based on the time interval and the preset time threshold; generating an item set based on at least one object included in each data session. Wherein, the time interval of any adjacent operation data is compared with the preset time threshold. If the time interval is greater than or equal to the preset time threshold, the adjacent operation data is divided into sessions, that is, the operation data of the previous timestamp in the adjacent operation data is used as the last operation data in the previous session, and the operation data of the next timestamp is used as the first operation data in the next session. Wherein, the preset time interval can be determined according to the object type. Taking the object as each store in the e-commerce platform as an example, the preset time interval can be half an hour or one hour. Each session includes operation data formed by users operating on various objects in the same operation time period. Object identifiers in each session are extracted to form item sets corresponding to each session. Each item set may include at least one object. For example, item set 1 may be {store A, store B, store C}, item set 2 may be {store A, store B, store D}, item set 3 may be {store B, store C}, and item set 4 may be {store A, store B}.

[0035] The associated objects of the current object may be other objects in each item set that belong to the same item set as the current object. In the above example, if the current object is store A, then store B, store C, and store D are associated stores of store A, and the competing objects are determined in the above associated stores. The correlation coefficients between the above associated objects and the current object are determined respectively, and the competing objects are determined based on the correlation coefficients. In some embodiments, the associated object with the largest correlation coefficient may be determined as the competing object of the current object.

[0036] In some embodiments, respectively determining the correlation coefficient between each associated object and the current object may be by obtaining content information of each associated object and content information of the current object, and determining a correlation index based on content similarity between the associated object and the current object. Taking the objects as stores in an e-commerce platform as an example, the content information of the associated object and the current object may be product information in the store.

[0037] In some embodiments, the association coefficient includes an association support and an association confidence, wherein the association support is used to characterize the frequency of the associated object and the current object appearing in the same period, and the association confidence is used to characterize the frequency of the associated object appearing when the current object appears. Accordingly, determining the associated objects of the current object in each set, and the association coefficient between the current object and each of the associated objects, includes: determining an object that appears synchronously with the current object in any set as an associated object of the current object; determining the association support based on the probability of the current object and the associated object appearing synchronously in each set; determining the association confidence based on the probability of the associated object appearing in each set when the current object appears.

[0038] For any associated object, the associated support can be determined based on the number of itemsets that occur simultaneously with the associated object and the current object, as well as the total number of itemsets. For details, see the following formula:

[0039] support = (the number of itemsets where A and B occur simultaneously) / the number of all itemsets, where A is the current object and B is the associated object.

[0040] For example, item set 1 can be {store A, store B, store C}, item set 2 can be {store A, store B, store D}, item set 3 can be {store A, store D}, item set 4 can be {store A, store B}, the number of item sets in which A and B occur simultaneously can be 3, and the number of all item sets is 4. Correspondingly, for the associated object B, the associated support is 3 / 4.

[0041] For any associated object, the association confidence is determined based on the number of itemsets that occur simultaneously with the associated object and the current object, as well as the total number of itemsets that occur with the current object. For details, see the following formula:

[0042] confidence=P(A|B)

[0043] Among them, A is the current object and B is the associated object.

[0044] For example, item set 1 can be {store A, store B, store C}, item set 2 can be {store A, store B, store D}, item set 3 can be {store A, store D}, item set 4 can be {store A, store B}, the number of item sets in which A and B occur simultaneously can be 3, and the total number of item sets occurring in the current object is 4. Correspondingly, for the associated object B, the association confidence is 3 / 4.

[0045] Correspondingly, for the associated object C, the associated support is 1 / 4 and the associated confidence is 1 / 4; for the associated store C, the associated support is 2 / 4 and the associated confidence is 2 / 4.

[0046] In some embodiments, the competitor of the current object is determined based on the correlation coefficient between the current object and each of the associated objects, including: determining the candidate competitor of the current object based on the support threshold and the confidence threshold; determining the competitor of the current object based on the associated support and / or associated confidence of each of the candidate competitor objects. In this embodiment, the support threshold and the confidence threshold are pre-set, and the support threshold and the confidence threshold can be determined according to the object type. Taking the objects as various stores in the e-commerce platform as an example, the support threshold can be 0.5%, and the confidence threshold can be 20%. For any associated object, the associated support of the associated object with the current object is compared with the support threshold, and the associated confidence of the associated object with the current object is compared with the confidence threshold. If the association support between the associated object and the current object is greater than or equal to the support threshold, and the association confidence between the associated object and the current object is greater than or equal to the support threshold, then the association relationship between the associated object and the current object is output, and the associated object is used as a candidate competing object. If the association support between the associated object and the current object is less than the support threshold, and / or the association confidence between the associated object and the current object is less than the support threshold, then it is determined that the associated object is not a candidate competing object.

[0047] Among the candidate competing objects, the competing object of the current object is determined based on the associated support and / or associated confidence of each of the candidate competing objects. Exemplarily, the candidate competing object with the largest associated support can be determined as the competing object, or the candidate competing object with the largest associated confidence can be determined as the competing object, or a weighted sum value based on preset weights is performed based on the associated support and / or associated confidence, and the candidate competing object with the largest weighted sum value is determined as the competing object.

[0048] In some embodiments, frequently associated objects with the current object may be screened based on the association support, and further, competing objects may be determined from the frequently associated objects based on the association confidence. Optionally, the association support between each associated object and the current object is first determined, and the frequently associated objects are screened to eliminate the associated objects that appear less frequently in the same period as the current object, that is, the associated objects whose association support is greater than or equal to the support threshold are determined as frequently associated objects, and for the frequently associated objects, the association confidence between the objects and the current object is determined to determine the competing objects, that is, the frequently associated objects whose association confidence is greater than or equal to the support threshold are determined as candidate competing objects, and the competing objects of the current object are determined based on the association support and / or association confidence of each of the candidate competing objects. In this embodiment, the process of calculating the association confidence for non-frequently associated objects is reduced, the process of determining competing objects is simplified, the amount of calculation is reduced, and the efficiency of determining competing objects is improved.

[0049] In some embodiments, if there is no candidate competing object in each associated object, that is, there is no associated object whose associated support is greater than or equal to the support threshold and whose associated confidence is greater than or equal to the support threshold, then the support threshold and the confidence threshold are adjusted. Specifically, the support threshold and the confidence threshold may be adjusted down synchronously. Optionally, the support threshold and the confidence threshold correspond to adjustment values ​​respectively, the support threshold minus the corresponding first adjustment value to obtain a new support threshold, the confidence threshold minus the corresponding second adjustment value to obtain a new confidence threshold, and the competing object is determined in the associated object based on the new support threshold and the new confidence threshold. Among them, the adjustment values ​​corresponding to the support threshold and the confidence threshold may be fixed adjustment values, or may be adjustment values ​​that vary with the number of adjustments. It should be noted that the support threshold and the confidence threshold correspond to minimum values ​​respectively. If the new support threshold or the new confidence threshold is less than the corresponding minimum value, the support threshold and the confidence threshold are stopped from being adjusted, it is determined that there is no competing object for the current object, and the current user is determined as a loyal user.

[0050] In some embodiments, after determining the competitor of the current object, the type of the competitor is determined, and the competitor is verified by the type of the competitor and the type of the current object. If the type of the competitor is the same as the type of the current object, the verification of the competitor is successful; if the type of the competitor is different from the type of the current object, the verification of the competitor fails, the competitor is eliminated, and the competitor of the current object is re-determined. Optionally, the object ranked second in the correlation coefficient ranking can be determined as the competitor of the current object.

[0051] In this embodiment, by determining the competing objects of the current object among the various objects operated by the current user, and determining the current user's loyalty index to the current object through the feature data of the current object and the feature data of the competing objects, the interference of the feature data of other objects in the calculation process is reduced, and the amount of calculation of the feature data of other objects in the calculation process is reduced.

[0052] The characteristic data of the current user's operations on the current object and the competing object may be the operation characteristic data within a preset time period, and different types of objects may correspond to different characteristic data, which may be determined according to the calculation requirements of each type of object. Taking the objects as the various stores in the e-commerce platform as an example, the characteristic data may include but are not limited to the time interval between the user's most recent purchase in the store in the past year and the present, the number of orders, the total order amount, the maximum order amount in the past year, the number of views in the past month, the number of views in the past month, the number of return orders in the past year, the number of rejected orders in the past year, the number of products concerned in the past year, the average customer price, the number of good reviews, and the number of bad reviews. In this embodiment, the characteristic data has a wide data range and a variety of quantity types, which is conducive to improving the calculation accuracy of the loyalty index.

[0053] Each of the above-mentioned feature data is correspondingly provided with a weight, and the loyalty index of the current user to the current object is determined based on the above-mentioned feature data and the corresponding weight. The weight of each feature data can be determined according to the feature data type and the degree of influence on loyalty. The weight of feature data of different types of objects is different. For example, the weights of the above-mentioned features are [0.1, 0.2, 0.35, 0.1, 0.02, 0.02, -0.03, -0.01, 0.07, 0.1, 0.05, -0.01]. The above-mentioned weights are only examples and can be determined according to the object type and user needs, and are not limited to this.

[0054] Optionally, based on the feature data of the current user's operations on the current object and the competing object, the loyalty index of the current user to the current object is determined, including: generating a first index of the current object based on pre-set weights of each feature data and the feature data corresponding to the current object; generating a second index of the competing object based on pre-set weights of each feature data and the feature data corresponding to the competing object; and determining the loyalty index of the current user to the current object based on the weights of the current object and the competing object, the first index and the second index.

[0055] For the current object and the competing object, the first index and the second index are determined based on the same calculation method, that is, the feature data is weighted by weight. It should be noted that due to the inconsistency of the dimensions of each feature data, all feature data are normalized to avoid the impact of inconsistent dimensions. It is necessary to consider the feature distribution of users and normalize the feature data of users in each object; it is also necessary to consider the distribution of user data of each object and normalize the feature data of all users in the same object.

[0056] The current object and the competing object are respectively provided with weights. Exemplarily, the weights of the current object and the competing object are 1 and -20% respectively. The current user's loyalty index to the current object is obtained by weighted calculation through the weight of the current object, the weight of the competing object, the first index and the second index. In some embodiments, the weights of the current object and the competing object may be fixed values. In other embodiments, the competing object may be determined based on the correlation coefficient between the competing object and the current object. The greater the correlation coefficient between the competing object and the current object, the greater the absolute weight of the competing object. It should be noted that the weight of the competing object is a negative value, indicating that the competing object has a side effect on the current user's loyalty to the current object.

[0057] On the basis of the above embodiment, before determining the loyalty index of the current user to the current object based on the characteristic data of the current user's operation on the current object and the competing object, it also includes: verifying whether each characteristic data is an abnormal value, and eliminating the abnormal value. In this embodiment, the characteristic data is judged to be abnormal based on the following formula, and the characteristic data that does not satisfy the following formula is determined as abnormal data, and the abnormal data is set to a default value, for example, 0, to reduce the interference of abnormal data on the loyalty index.

[0058]

[0059] Among them, data_mean ij is the mean of the characteristic data, is the feature data variance, data ij is the characteristic data.

[0060] For each user who operates the current object, the loyalty index of each user to the current object is calculated based on the above method, and each user is classified based on the loyalty index of each user to the current object. In some embodiments, the users may be classified according to the ratio of each loyalty classification and the loyalty index of each user to the current object. Optionally, based on the ranking of the loyalty index of each user, the loyalty classification of each user to the current object is determined, including: based on the ratio of each classification and the ranking of the loyalty index of each user, the loyalty classification of each user to the current object is determined. According to the Pareto principle, that is, the 80 / 20 rule, the first 20% of the users in the loyalty index ranking are determined as loyal users, and the last 80% of the users in the loyalty index ranking are determined as non-loyal users. Optionally, for loyal users, loyal users and non-loyal users may be further classified based on the golden section point, in the ranking of the loyalty index of loyal users, loyal users before the golden section point are determined as very high loyalty, loyal users after the golden section point are determined as high loyalty, and in the ranking of the loyalty index of non-loyal users, non-loyal users before the golden section point are determined as low loyalty, and non-loyal users after the golden section point are determined as very low loyalty.

[0061] In some embodiments, each user may be classified according to the value of their loyalty index to the current object, that is, the loyalty index range corresponding to each loyalty level is determined, and users are classified based on the loyalty index range and the loyalty index of each user to the current object.

[0062] The technical solution provided in this embodiment obtains the object operation data of each user, determines the competitor of the current object for each user, determines the loyalty index of each user to the current object based on the feature data of the current object and the competitor, and determines the loyalty classification of each user to the current object based on the ranking of the loyalty index of each user. In this embodiment, there is no need to build a clustering or regression model, the amount of calculation is small, and the efficiency of loyalty classification is improved. At the same time, by determining the competitor, the loyalty classification is carried out in a targeted manner, thereby improving the accuracy of loyalty classification.

[0063] Embodiment 2

[0064] Figure 2 : is a schematic diagram of the structure of a user classification device provided in Embodiment 2 of the present invention, the device comprising:

[0065] The object operation data acquisition module 210 is used to acquire the object operation data of each user;

[0066] A competitor determination module 220, for determining, for any user, a competitor of a current object based on the object operation data of the current user;

[0067] A feature data acquisition module 230, used to respectively acquire feature data of the current user's operations on the current object and the competing object;

[0068] A loyalty index determination module 240, configured to determine the loyalty index of the current user to the current object based on feature data of the current user's operations on the current object and the competing object;

[0069] The loyalty classification module 250 is used to determine the loyalty classification of each user to the current object based on the loyalty index ranking of each user.

[0070] Based on the above embodiment, the competitor determination module 220 includes:

[0071] an item set determining unit, configured to generate at least one item set of objects based on the operation timestamp of each object in the object operation data of the current user, wherein each item set includes at least one object;

[0072] An association coefficient determination unit, used to determine the associated objects of the current object in each item set, and the association coefficient between the current object and each of the associated objects;

[0073] The competitor object determination unit is used to determine the competitor object of the current object according to the correlation coefficient between the current object and each of the associated objects.

[0074] Based on the above embodiment, the item set determination unit is used to:

[0075] Determine the time interval of adjacent operation data according to the operation timestamp of each object in the object operation data;

[0076] Based on the time interval and a preset time threshold, dividing the object operation data into a plurality of data sessions;

[0077] An itemset is generated based on at least one object included in each data session.

[0078] Based on the above embodiment, the association coefficient includes association support and association confidence;

[0079] Accordingly, the correlation coefficient determination unit is used to:

[0080] Determine an object in any item set that appears synchronously with the current object as an associated object of the current object;

[0081] Determining the association support based on the probability of the current object and the associated object appearing simultaneously in each set;

[0082] The association confidence is determined based on the probability that the associated object appears in each item set when the current object appears.

[0083] Based on the above embodiment, the competitor determination unit is used to:

[0084] Determine a candidate competing object with the current object according to a support threshold and a confidence threshold;

[0085] The competitor of the current object is determined based on the associated support and / or associated confidence of each of the candidate competitor objects.

[0086] Based on the above embodiment, the competitor determination unit is further configured to:

[0087] If there is no candidate competitor, the support threshold and / or the confidence threshold are adjusted, and the competitor is re-determined based on the adjusted support threshold and confidence threshold.

[0088] Based on the above embodiment, the loyalty index determination module 240 is used to:

[0089] Generate a first index of the current object based on the preset weights of the feature data and the feature data corresponding to the current object;

[0090] Generate a second index of the competitor based on the preset weights of the feature data and the feature data corresponding to the competitor;

[0091] A loyalty index of the current user to the current object is determined based on the weights of the current object and the competing object, the first index, and the second index.

[0092] Based on the above embodiment, the device further includes:

[0093] The abnormal data elimination module is used to verify whether each feature data is an abnormal value and eliminate the abnormal value before determining the loyalty index of the current user to the current object based on the feature data of the current user's operation on the current object and the competing object.

[0094] Based on the above embodiment, the loyalty classification module 250 is used to:

[0095] Based on the proportion of each category and the loyalty index ranking of each user, the loyalty category of each user to the current object is determined.

[0096] The user classification device provided in the embodiment of the present invention can execute the user classification method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0097] Embodiment 3

[0098] Figure 3 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. Figure 3 A block diagram of an electronic device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 3 The electronic device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. The device 12 is typically an electronic device that undertakes the image classification function.

[0099] like Figure 3 As shown, the electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors 16, a storage device 28, and a bus 18 connecting various system components (including the storage device 28 and the processor 16).

[0100] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. For example, these architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0101] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0102] The storage device 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 3 not shown, usually called a "hard drive"). Although Figure 3Not shown in the figure, a disk drive for reading and writing a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a read-only optical disk (Compact Disc-Read Only Memory, CD-ROM), a digital video disk (Digital Video Disc-Read Only Memory, DVD-ROM) or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The storage device 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of various embodiments of the present invention.

[0103] A program 36 having a set (at least one) of program modules 26 may be stored, for example, in a storage device 28, such program modules 26 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof may include an implementation of a gateway environment. The program modules 26 generally perform the functions and / or methods of the embodiments described herein.

[0104] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, cameras, displays 24, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 12, and / or may communicate with any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via an input / output (I / O) interface 22. In addition, the electronic device 12 may also communicate with one or more gateways (e.g., a local area network (LAN), a wide area network (WAN), and / or a public gateway, such as the Internet) via a gateway adapter 20. As shown, the gateway adapter 20 communicates with other modules of the electronic device 12 via a bus 18. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, disk arrays (Redundant Arrays of Independent Disks, RAID) systems, tape drives, and data backup storage systems.

[0105] The processor 16 executes various functional applications and data processing by running the programs stored in the storage device 28, such as implementing the user classification method provided in the above embodiment of the present invention.

[0106] Embodiment 4

[0107] A fourth embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the user classification method provided in the embodiment of the present invention is implemented.

[0108] Of course, the computer-readable storage medium provided by the embodiment of the present invention stores a computer program which is not limited to the method operation described above, and can also execute the user classification method provided by any embodiment of the present invention.

[0109] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, 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 above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0110] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable source code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0111] Source code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0112] Computer source code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The source code may 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 may be connected to the user's computer through any type of gateway, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0113] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A user classification method, It is characterized in that include: Get object operation data of each user; For any user, determining a competitor of the current object based on the object operation data of the current user, and obtaining feature data of the current user's operations on the current object and the competitor respectively; Determine the loyalty index of the current user to the current object based on the characteristic data of the current user's operations on the current object and the competing object; wherein the competing object is an object of the same purpose as the current object and with the highest similarity among the objects operated by the current user; Determining the loyalty classification of each user to the current object based on the loyalty index ranking of each user; Wherein, determining the loyalty index of the current user to the current object based on the characteristic data of the current user's operations on the current object and the competing object includes: Generate a first index of the current object based on the preset weights of the feature data and the feature data corresponding to the current object; Generate a second index of the competitor based on the preset weights of the feature data and the feature data corresponding to the competitor; The loyalty index of the current user to the current object is determined based on the weights of the current object and the competing object, the first index, and the second index; wherein the weight of the competing object is a negative value.

2. The method according to claim 1, It is characterized in that The determining the competing object of the current object based on the object operation data of the current user includes: generating at least one item set of objects based on the operation timestamp of each object in the object operation data of the current user, wherein each item set includes at least one object; Determine the associated objects of the current object in each item set, and the correlation coefficient between the current object and each of the associated objects; The competitor of the current object is determined according to the correlation coefficient between the current object and each of the associated objects.

3. The method according to claim 2, It is characterized in that The generating at least one item set of objects based on the operation timestamp of each object in the object operation data of the current user comprises: Determine the time interval of adjacent operation data according to the operation timestamp of each object in the object operation data; Based on the time interval and a preset time threshold, dividing the object operation data into a plurality of data sessions; An itemset is generated based on at least one object included in each data session.

4. The method according to claim 2, It is characterized in that The association coefficient includes association support and association confidence; The determining of the associated objects of the current object in each item set and the correlation coefficient between the current object and each of the associated objects includes: Determine an object in any item set that appears synchronously with the current object as an associated object of the current object; Determining the association support based on the probability of the current object and the associated object appearing simultaneously in each set; The association confidence is determined based on the probability that the associated object appears in each item set when the current object appears.

5. The method according to claim 4, wherein, determining the competing object of the current object according to the correlation coefficients between the current object and each of the associated objects includes: determining candidate competing objects for the current object according to a support threshold and a confidence threshold; determining the competing object of the current object based on the association support and / or association confidence of each of the candidate competing objects.

6. The method according to claim 5, wherein, the method further includes: if there are no candidate competing objects, adjusting the support threshold and / or the confidence threshold, and re-determining the competing objects with the adjusted support threshold and confidence threshold.

7. The method according to claim 1, wherein, before determining the loyalty index of the current user for the current object based on the feature data of the operations of the current user on the current object and the competing object, the method further includes: verifying whether each of the feature data is an outlier, and removing the outlier.

8. The method according to claim 1, wherein, determining the loyalty classification of each user for the current object based on the sorting of the loyalty indices of each user includes: determining the loyalty classification of each user for the current object based on the proportion of each classification and the sorting of the loyalty indices of each user.

9. A user classification device, wherein, it includes: an object operation data acquisition module for acquiring the object operation data of each user; a competing object determination module for, for any user, determining the competing object of the current object based on the object operation data of the current user; wherein, the competing object is the object with the same category as the current object and the highest similarity among the objects operated by the current user; a feature data acquisition module for respectively acquiring the feature data of the operations of the current user on the current object and on the competing object; a loyalty index determination module for determining the loyalty index of the current user for the current object based on the feature data of the operations of the current user on the current object and the competing object; a loyalty classification module for determining the loyalty classification of each user for the current object based on the sorting of the loyalty indices of each user; wherein, the loyalty index determination module is used for: generating a first index of the current object based on the weights of each pre-set feature data and the feature data corresponding to the current object; generating a second index of the competing object based on the weights of each pre-set feature data and the feature data corresponding to the competing object; determining the loyalty index of the current user for the current object based on the weights of the current object and the competing object, the first index and the second index; wherein, the weight of the competing object is negative.

10. An electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, when the processor executes the program, it implements the user classification method according to any one of claims 1-8.

11. A computer-readable storage medium, on which a computer program is stored, It is characterized in that When the program is executed by a processor, the user classification method as described in any one of claims 1 to 8 is implemented.

Images