An AI consumer portrait construction method and system that fuses online and offline behaviors

Abstract

The application relates to the technical field of consumer portrait construction, and relates to an AI-fused online and offline behavior consumer portrait construction method and system, which comprises the following steps: determining a consumer portrait vector set according to historical online index values and historical offline index values, identifying a user portrait graph set corresponding to the consumer portrait vector set, utilizing AI curved surface technology, constructing a consumer population distribution graph according to a portrait vector terminal point and a consumer population, determining a current portrait graph according to current user consumption data, identifying a similar portrait graph corresponding to the current portrait graph in the user portrait graph set according to the consumer population distribution graph, identifying reference shopping data corresponding to the similar portrait graph, and performing consumption prediction on a current user according to the reference shopping data. The application can improve the construction accuracy of a current consumer portrait.

Description

Technical Field

[0001] This invention relates to the field of consumer profiling technology, and in particular to a method and system for constructing consumer profiles that integrates online and offline behaviors using AI. Background Technology

[0002] In today's retail landscape, consumers' decision-making processes are highly fragmented and omnichannel. A consumer's shopping journey may begin with online social media recommendations, followed by in-store experiences and comparisons, and may ultimately lead back to online live streaming channels to place an order.

[0003] The current intertwined online-to-offline (O2O) behavior makes it difficult for traditional, singular consumer profiling methods to capture complete user intent and the full picture. Therefore, current consumer profiling suffers from low accuracy. Summary of the Invention

[0004] This invention provides a method and system for constructing consumer profiles by integrating online and offline behaviors using AI, with the main purpose of improving the accuracy of current consumer profile construction.

[0005] To achieve the above objectives, this invention provides a method for constructing consumer profiles based on AI-integrated online and offline behavior, comprising: acquiring a historical user consumption dataset and a set of online and offline indicators, wherein the set of online and offline indicators includes: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency; sequentially extracting historical user consumption data from the historical user consumption dataset, and extracting historical online indicator values ​​and historical offline indicator values ​​from the historical user consumption data according to the online and offline indicator sets in the set of online and offline indicators, wherein each historical user consumption data corresponds to a historical user; determining the consumer profile angle and consumer profile length based on the historical online indicator values ​​and historical offline indicator values ​​respectively; constructing a consumer profile vector based on the consumer profile angle and consumer profile length to obtain a consumer profile vector set. In this model, the consumer profile angle is represented by the central angle of a unit circle, and the consumer profile vector starts from the point on the circumference of the unit circle corresponding to the consumer profile angle, with the area inside the unit circle as the vector extension area. The model identifies the consumer profile endpoint set corresponding to the consumer profile vector set, determines the user profile graphic based on the endpoint set, and obtains a user profile graphic set, where each user profile graphic corresponds to a historical user. It also identifies the number of consumers corresponding to the endpoints of the consumer profile vectors, and uses pre-built AI surface technology to construct a consumer distribution map based on the endpoints of the profile vectors and the number of consumers. The model receives the current user's current user consumption data and determines the current user profile graphic based on this data. Based on the consumer distribution map, it identifies similar user profile graphics corresponding to the current user profile graphic in the user profile graphic set, and identifies reference shopping data corresponding to these similar user profile graphics. Finally, it predicts the current user's consumption based on the reference shopping data, completing the construction of an AI-integrated online and offline consumer profile.

[0006] Optionally, the step of constructing a consumer profile vector based on the consumer profile angle and the consumer profile diameter to obtain a consumer profile vector set includes: identifying the starting point of the profile vector in a pre-constructed consumer profile scale based on the consumer profile angle, wherein the angle between the starting point of the profile vector and the zero mark is the consumer profile angle; determining the profile vector magnitude based on the consumer profile diameter, identifying the starting point of the profile vector in the inward radial direction of the consumer profile scale, wherein the inward radial direction points from the starting point of the profile vector to the center of the consumer profile scale; using the inward radial direction as the profile vector direction, drawing the consumer profile vector in the consumer profile scale based on the starting point of the profile vector, the profile vector magnitude, and the profile vector direction; and aggregating the consumer profile vectors corresponding to each online and offline indicator group to obtain a consumer profile vector set.

[0007] Optionally, determining the consumer profile angle and consumer profile length based on historical online and offline indicator values ​​respectively includes: identifying the online indicator interval and offline indicator interval corresponding to the online and offline indicator groups in the historical user consumption dataset; calculating the consumer profile angle based on the online indicator interval, historical online indicator values, and a pre-constructed profile angle formula, wherein the profile angle formula is as follows: in, This represents the consumer profile perspective corresponding to the j-th historical user consumption data in the i-th online / offline indicator group. This represents the historical online metric value corresponding to the j-th historical user's consumption data. This represents the upper limit of the online indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the online indicator range corresponding to the i-th online and offline indicator group; the consumer profile length is calculated based on the offline indicator range, historical offline indicator values, and a pre-constructed profile length formula, wherein the profile length formula is as follows: in, This represents the length of the consumer profile corresponding to the j-th historical user's consumption data in the i-th online / offline indicator group. This represents the historical offline indicator value corresponding to the j-th historical user's consumption data. This represents the upper limit of the offline indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the offline indicator interval corresponding to the i-th online and offline indicator group.

[0008] Optionally, the step of determining the user profile graph based on the consumer profile endpoint set to obtain a user profile graph set includes: sequentially extracting online and offline indicator groups from the online and offline indicator group set; determining whether the online and offline indicator group is the last online and offline indicator group in the online and offline indicator group set; if the online and offline indicator group is not the last online and offline indicator group in the online and offline indicator group set, then identifying the associated consumer profile endpoint corresponding to the online and offline indicator group in the consumer profile endpoint set, and returning to the step of sequentially extracting online and offline indicator groups from the online and offline indicator group set; if the online and offline indicator group is the last online and offline indicator group in the online and offline indicator group set, then sequentially connecting the associated consumer profile endpoints to obtain a user profile polyline; identifying the last consumer profile endpoint corresponding to the last online and offline indicator group in the online and offline indicator group set and the first consumer profile endpoint corresponding to the first online and offline indicator group; connecting the last consumer profile endpoint and the first consumer profile endpoint in the user profile polyline to obtain a user profile graph; and collecting the user profile graphs corresponding to each historical user consumption data to obtain a user profile graph set.

[0009] Optionally, the number of consumers corresponding to the endpoint of the consumer profile vector includes: identifying the associated historical online indicator value and the associated historical offline indicator value corresponding to the endpoint of the profile vector; and identifying the number of consumers corresponding to the associated historical online indicator value and the associated historical offline indicator value in the historical user consumption dataset.

[0010] Optionally, the step of constructing a consumer distribution map based on the endpoint of the portrait vector and the number of consumers using pre-built AI surface technology includes: determining two-dimensional disk coordinates based on the endpoint of the portrait vector, and determining three-dimensional vertical axis coordinates based on the number of consumers; determining three-dimensional coordinate points in a pre-built three-dimensional coordinate system based on the two-dimensional disk coordinates and the three-dimensional vertical axis coordinates to obtain a set of three-dimensional coordinate points; wherein the x-axis and y-axis of the three-dimensional coordinate system are located on the plane of the consumer portrait scale, and the z-axis of the three-dimensional coordinate system is the axis passing through the center of the consumer portrait scale and perpendicular to the consumer portrait scale; using the AI ​​surface technology, drawing a three-dimensional surface based on the set of three-dimensional coordinate points, and using the three-dimensional surface as the consumer distribution map, wherein the AI ​​surface technology refers to an AI fitting technology that can fit a three-dimensional spatial surface based on multiple three-dimensional spatial points.

[0011] Optionally, determining the current profile graphic based on the current user consumption data includes: extracting the current online indicator value and the current offline indicator value from the current user consumption data based on the online and offline indicator groups in the online and offline indicator group set; determining the current profile angle and the current profile diameter based on the current online indicator value and the current offline indicator value; determining the current profile vector based on the current profile angle and the current profile diameter to obtain the current profile vector set; identifying the current profile endpoint set corresponding to the current profile vector set; and determining the current profile graphic based on the current profile endpoint set.

[0012] Optionally, the step of identifying similar profile images corresponding to the current profile image in the user profile image set based on the consumer distribution map includes: identifying the current vertical projection area of ​​the current profile image on the consumer distribution map; extracting the current three-dimensional surface image of the current vertical projection area from the consumer distribution map, and determining the current three-dimensional surface volume fraction based on the current three-dimensional surface image and the consumer profile scale; identifying the user vertical projection area of ​​the user profile image on the consumer distribution map; extracting the user three-dimensional surface image of the user vertical projection area from the consumer distribution map to obtain a user three-dimensional surface image set; determining the user three-dimensional surface volume fraction set based on the user three-dimensional surface image set and the consumer profile scale; calculating the three-dimensional volume fraction difference between each user three-dimensional surface volume fraction in the user three-dimensional surface volume fraction set and the current three-dimensional surface volume fraction to obtain a three-dimensional volume fraction difference value set; identifying the minimum three-dimensional volume fraction difference value in the three-dimensional volume fraction difference value set, and identifying the similar profile image image corresponding to the minimum three-dimensional volume fraction difference value.

[0013] Optionally, identifying the reference shopping data corresponding to the similar profile image includes: identifying the historical consumers corresponding to the similar profile image; obtaining the historical shopping data of the historical consumers, and using the historical shopping data as the reference shopping data of the current user.

[0014] To achieve the above objectives, this invention also provides an AI-integrated online and offline behavior-based consumer profiling system, comprising: a historical online and offline indicator value determination module, used to acquire historical user consumption datasets and online and offline indicator sets, wherein the online and offline indicator sets include: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency; extracting historical user consumption data sequentially from the historical user consumption dataset, and extracting historical online indicator values ​​and historical offline indicator values ​​from the historical user consumption data according to the online and offline indicator sets in the online and offline indicator sets, wherein each historical user consumption data corresponds to a historical user; and a user profile graph set construction module, used to determine the consumer profile angle and consumer profile length according to the historical online and offline indicator values ​​respectively; and constructing a consumer profile vector according to the consumer profile angle and consumer profile length. The method involves several steps: First, a consumer profile vector set is obtained, where the consumer profile angle is represented by the central angle of a unit circle. The consumer profile vector originates from the circumference point on the unit circle corresponding to the consumer profile angle, and extends to the area within the unit circle. The method then identifies the consumer profile endpoint set corresponding to the consumer profile vector set, determines the user profile graphic based on the endpoint set, and obtains a user profile graphic set, where each user profile graphic corresponds to a historical user. A consumer number distribution map construction module is used to identify the number of consumers corresponding to the endpoints of the consumer profile vectors, and constructs a consumer number distribution map based on the endpoints of the profile vectors and the number of consumers using pre-built AI surface technology. A current user consumption prediction module receives the current user's current user consumption data, determines the current user profile graphic based on the current user consumption data, identifies similar user profile graphics corresponding to the current user profile graphic in the user profile graphic set based on the consumer number distribution map, identifies reference shopping data corresponding to the similar user profile graphics, and makes consumption predictions for the current user based on the reference shopping data. To address the above problems, this invention also provides an electronic device, comprising: a memory storing at least one instruction; and a processor executing the instructions stored in the memory to implement the aforementioned AI-integrated online and offline behavior consumer profile construction method.

[0015] To address the aforementioned issues, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned AI-integrated online and offline behavior consumer profile construction method.

[0016] To address the problems described in the background art, this invention requires constructing a user profile graphic set and a current user profile graphic. Then, by combining this with a consumer distribution map, similar user profile graphics are identified within the user profile graphic set. Finally, consumption predictions for the current user are made based on reference shopping data corresponding to these similar user profile graphics. Specifically, when constructing the user profile graphic set, it is necessary to first obtain a historical user consumption dataset and an online / offline indicator set. The online / offline indicator set includes: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - actual purchase frequency. Then, historical user consumption data for each historical user is extracted sequentially from the historical user consumption dataset. To graphically represent this historical user consumption data, historical online and offline indicator values ​​can be extracted from the historical user consumption data based on the online / offline indicator sets in the online / offline indicator set. Each historical user consumption data corresponds to one historical user. Finally, a consumer profile is determined based on the historical online and offline indicator values. The process involves determining the angle and length of the consumer profile, then constructing a consumer profile vector based on these parameters, resulting in a consumer profile vector set. To comprehensively represent this vector set graphically, the corresponding consumer profile endpoint set can be identified. Based on this endpoint set, user profile graphics are determined, resulting in a user profile graphic set. Each user profile graphic corresponds to a historical user. To identify similar profile graphics corresponding to the current profile graphic, a three-dimensional representation of the current profile graphic can be created using the number of consumers and the endpoints of the profile vectors, improving the matching accuracy of similar profile graphics. Specifically, the number of consumers corresponding to the endpoints of the consumer profile vectors needs to be identified. Then, using pre-built AI surface technology, a consumer distribution map is constructed based on the endpoints of the profile vectors and the number of consumers. At this point, the current user's current consumption data can be received. The current profile graphic is then determined based on this data. Similar profile graphics corresponding to the current profile graphic are identified within the user profile graphic set based on the consumer distribution map. Finally, reference shopping data corresponding to the similar profile graphics is identified, allowing for consumption prediction of the current user based on this reference shopping data. Therefore, this invention improves the accuracy of constructing the current consumer profile. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a consumer profile construction method that integrates online and offline behaviors according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a consumer profile vector provided in an embodiment of the present invention; Figure 3 A functional module diagram of an AI-integrated online and offline behavior consumer profiling system provided in an embodiment of the present invention; Figure 4This is a schematic diagram of the structure of an electronic device that implements the AI-integrated online and offline behavior consumer profile construction method according to an embodiment of the present invention.

[0018] Explanation of reference numerals in the attached figures: 1. Electronic device; 10. Processor; 11. Memory; 12. Bus.

[0019] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0021] This application provides a method for constructing consumer profiles by integrating online and offline behavior using AI. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for constructing consumer profiles by integrating online and offline behavior can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0022] Reference Figure 1 The diagram shown is a flowchart illustrating a consumer profile construction method that integrates online and offline behavior using AI, according to an embodiment of the present invention. In this embodiment, the consumer profile construction method that integrates online and offline behavior using AI includes: S1, obtaining historical user consumption datasets and sets of online and offline indicators.

[0023] Understandably, the historical user consumption dataset refers to a collection of historical consumption data of users regarding a certain product or category of products. These products can include clothing and footwear, electronic products, beauty and personal care products, food and fresh produce, home furnishings and decoration products, baby and children's products, sports and outdoor equipment, etc. Furthermore, clothing and footwear can be further subdivided into women's wear, men's wear, children's wear, underwear, etc., and electronic products can be further subdivided into mobile phones, computers, photography and video equipment, etc. The online and offline indicator set refers to a collection of consumption indicators comprised of online and offline consumption indicators.

[0024] In detail, the online and offline indicator set includes: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - actual purchase frequency. Product browsing time refers to the length of time a user browses a certain type of product online, and product dwell time refers to the length of time a consumer spends at a store or booth selling that type of product offline. Product click-through rate refers to the number of times a user clicks on a certain type of product within a certain period, for example, the number of times a consumer clicks on that type of product online within a month. Product trial frequency refers to the number of times a user tries that type of product within a certain period. Product keyword search frequency refers to the number of times a consumer searches for a certain type of product online within a certain period, and product inquiry frequency refers to the number of times a consumer inquires about that type of product offline within a certain period. Product detail page viewing frequency refers to the number of times a consumer views the detail page of a certain type of product online within a certain period, and cross-store comparison frequency refers to the number of times a consumer compares prices at different stores offline within a certain period. The frequency of adding products to the shopping cart refers to the number of times online consumers add a certain type of product to their shopping cart, while the actual purchase frequency refers to the number of times consumers purchase that type of product offline within a certain period of time.

[0025] S2. Extract historical user consumption data sequentially from the historical user consumption dataset. Based on the online and offline indicator groups in the online and offline indicator group set, extract historical online indicator values ​​and historical offline indicator values ​​from the historical user consumption data.

[0026] Specifically, the historical user consumption data corresponds to a historical user. The historical user consumption data refers to a user's historical consumption data regarding a specific product or category of products. The online-offline indicator group refers to a combination of online and offline consumption indicators. The historical online indicator value refers to the specific value in the historical user consumption data corresponding to the online indicator in the online-offline indicator group. The historical offline indicator value refers to the specific value in the historical user consumption data corresponding to the offline indicator in the online-offline indicator group. For example, when the online-offline indicator group is product click-through rate minus product trial frequency, the historical online indicator value could be 10 times / month, and the historical offline indicator value could be 5 times / month.

[0027] S3. Determine the consumer profile angle and consumer profile length based on historical online and offline indicator values ​​respectively.

[0028] Understandably, the consumer profile angle refers to the vector angle representing the consumer profile, and the consumer profile diameter refers to the vector magnitude representing the consumer profile.

[0029] In this embodiment of the invention, determining the consumer profile angle and consumer profile length based on historical online and offline indicator values ​​respectively includes: identifying the online indicator interval and offline indicator interval corresponding to the online and offline indicator groups in the historical user consumption dataset; calculating the consumer profile angle based on the online indicator interval, historical online indicator values, and a pre-constructed profile angle formula, wherein the profile angle formula is as follows: in, This represents the consumer profile perspective corresponding to the j-th historical user consumption data in the i-th online / offline indicator group. This represents the historical online metric value corresponding to the j-th historical user's consumption data. This represents the upper limit of the online indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the online indicator range corresponding to the i-th online and offline indicator group; the consumer profile length is calculated based on the offline indicator range, historical offline indicator values, and a pre-constructed profile length formula, wherein the profile length formula is as follows: in, This represents the length of the consumer profile corresponding to the j-th historical user's consumption data in the i-th online / offline indicator group. This represents the historical offline indicator value corresponding to the j-th historical user's consumption data. This represents the upper limit of the offline indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the offline indicator interval corresponding to the i-th online and offline indicator group.

[0030] Furthermore, the online indicator range refers to the data range of the online indicator in the historical user consumption dataset within the online and offline indicator group, and the offline indicator range refers to the data range of the offline indicator in the historical user consumption dataset within the online and offline indicator group. For example, when the online indicator is the product click-through rate, the online indicator range can be [0 clicks / month, 300 clicks / month], and when the offline indicator is the product trial frequency, the offline indicator range can be [0 clicks / month, 100 clicks / month]. The upper limit of the range refers to the maximum value of the online or offline indicator range, and the lower limit of the range refers to the minimum value of the online or offline indicator range.

[0031] S4. Construct consumer profile vectors based on the consumer profile angle and consumer profile length to obtain a consumer profile vector set.

[0032] Understandably, the consumer profile vector refers to the component vector representing a consumer profile. The consumer profile vector set refers to the set of consumer profile vectors for a historical user across various online and offline indicator groups.

[0033] In detail, the consumer profile angle is represented by the central angle of the unit circle, and the consumer profile vector takes the point on the circumference of the unit circle corresponding to the consumer profile angle as the vector starting point, and the area inside the unit circle as the vector extension area.

[0034] Furthermore, the aforementioned consumer profile vector set can be found in [reference needed]. Figure 2 As shown, the consumer profile vector set includes: consumer profile angles Corresponding consumer profile vectors and consumer profile angles Corresponding consumer profile vectors and consumer profile angles Corresponding consumer profile vectors and consumer profile angles Corresponding consumer profile vectors and consumer profile angles The corresponding consumer profile vectors can be represented in sequence as follows: a consumer profile vector representing the product browsing time - product dwell time, a consumer profile vector representing the product click-through rate - product trial frequency, a consumer profile vector representing the product keyword search frequency - product inquiry frequency, a consumer profile vector representing the product detail page viewing frequency - cross-store comparison frequency, and a consumer profile vector representing the product adding to cart frequency - actual purchase frequency.

[0035] In this embodiment of the invention, the step of constructing a consumer profile vector based on the consumer profile angle and the consumer profile diameter to obtain a consumer profile vector set includes: identifying the starting point of the profile vector in a pre-constructed consumer profile scale based on the consumer profile angle, wherein the angle between the starting point of the profile vector and the zero mark is the consumer profile angle; determining the profile vector magnitude based on the consumer profile diameter, identifying the starting point of the profile vector in the inward radial direction of the consumer profile scale, wherein the inward radial direction points from the starting point of the profile vector to the center of the consumer profile scale; using the inward radial direction as the profile vector direction, drawing the consumer profile vector in the consumer profile scale based on the starting point of the profile vector, the profile vector magnitude, and the profile vector direction; and collecting the consumer profile vectors corresponding to each online and offline indicator group to obtain a consumer profile vector set.

[0036] Understandably, the starting point of the portrait vector refers to the starting point of the consumer portrait vector. The consumer portrait scale refers to the scale disk representing the consumer portrait vector. The zero mark refers to the mark point representing zero degrees on the consumer portrait scale disk. The portrait vector magnitude refers to the magnitude of the consumer portrait vector. The inward radial direction refers to the vector direction of the consumer portrait vector.

[0037] S5. Identify the consumer profile endpoint set corresponding to the consumer profile vector set, determine the user profile graphic based on the consumer profile endpoint set, and obtain the user profile graphic set.

[0038] Understandably, the consumer profile endpoint set refers to the set of endpoints of each consumer profile vector in the consumer profile vector set. The user profile graph refers to the set of images representing consumer / user profiles. Since each historical user consumption data corresponds to a historical user / consumer, each user profile graph corresponds to a historical user.

[0039] In this embodiment of the invention, determining the user profile graph based on the consumer profile endpoint set to obtain the user profile graph set includes: sequentially extracting online and offline indicator groups from the online and offline indicator group set; determining whether the online and offline indicator group is the last online and offline indicator group in the online and offline indicator group set; if the online and offline indicator group is not the last online and offline indicator group in the online and offline indicator group set, then identifying the associated consumer profile endpoint corresponding to the online and offline indicator group in the consumer profile endpoint set, and returning to the step of sequentially extracting online and offline indicator groups from the online and offline indicator group set; if the online and offline indicator group is the last online and offline indicator group in the online and offline indicator group set, then sequentially connecting the associated consumer profile endpoints to obtain a user profile polyline; identifying the last consumer profile endpoint corresponding to the last online and offline indicator group in the online and offline indicator group set and the first consumer profile endpoint corresponding to the first online and offline indicator group; connecting the last consumer profile endpoint and the first consumer profile endpoint in the user profile polyline to obtain a user profile graph; and collecting the user profile graphs corresponding to each historical user consumption data to obtain a user profile graph set.

[0040] Furthermore, the associated consumer profile endpoint refers to the consumer profile endpoint corresponding to the online and offline indicator group in the consumer profile endpoint set. For example, when the online and offline indicator group is product click-through rate minus product trial frequency, the associated consumer profile endpoint is the vector endpoint of the consumer profile vector corresponding to product click-through rate minus product trial frequency. The user profile line graph refers to a line graph representing the user profile. The last consumer profile endpoint refers to the consumer profile endpoint corresponding to the last online and offline indicator group. The first consumer profile endpoint refers to the consumer profile endpoint corresponding to the first online and offline indicator group.

[0041] S6. Identify the number of consumers corresponding to the endpoint of the consumer profile vector, and use pre-built AI surface technology to construct a consumer distribution map based on the endpoint of the profile vector and the number of consumers.

[0042] In detail, the number of consumers refers to the number of consumers whose historical online and offline metric values ​​correspond to the endpoint of the profile vector. For example, when the historical online metric value corresponding to the endpoint of the profile vector is a product click-through rate of 200 times / month and the historical offline metric value is a product trial frequency of 30 times / month, the number of consumers refers to the number of consumers who meet the criteria of a product click-through rate of 200 times / month and a product trial frequency of 30 times / month. The AI ​​surface technology refers to the technology of fitting a three-dimensional surface based on multiple scattered points in three-dimensional space. The AI ​​surface technology is existing technology and will not be described in detail here. The consumer distribution map refers to a three-dimensional surface map showing the correspondence between the endpoint of the profile vector and the number of consumers.

[0043] In this embodiment of the invention, the identification of the number of consumers corresponding to the endpoint of the consumer profile vector includes: identifying the associated historical online indicator value and the associated historical offline indicator value corresponding to the endpoint of the profile vector; and identifying the number of consumers corresponding to the associated historical online indicator value and the associated historical offline indicator value in the historical user consumption dataset.

[0044] Understandably, the associated historical online indicator value refers to the historical online indicator value corresponding to the endpoint of the profile vector, and the associated historical offline indicator value refers to the historical offline indicator value corresponding to the endpoint of the profile vector. For example, when the associated historical online indicator value is a product click-through rate of 200 times / month and the associated historical offline indicator value is a product trial frequency of 30 times / month, the number of consumers can be 1000.

[0045] In this embodiment of the invention, the step of constructing a consumer distribution map based on the endpoint of the portrait vector and the number of consumers using pre-built AI surface technology includes: determining two-dimensional disk coordinates based on the endpoint of the portrait vector, and determining three-dimensional vertical axis coordinates based on the number of consumers; determining three-dimensional coordinate points in a pre-built three-dimensional coordinate system based on the two-dimensional disk coordinates and the three-dimensional vertical axis coordinates to obtain a set of three-dimensional coordinate points; wherein the x-axis and y-axis of the three-dimensional coordinate system are located on the plane of the consumer portrait scale disk, and the z-axis of the three-dimensional coordinate system is the axis passing through the center of the consumer portrait scale disk and perpendicular to the axis of the consumer portrait scale disk; using the AI ​​surface technology, drawing a three-dimensional surface based on the set of three-dimensional coordinate points, and using the three-dimensional surface as the consumer distribution map, wherein the AI ​​surface technology refers to an AI fitting technology that can fit a three-dimensional spatial surface based on multiple three-dimensional spatial points.

[0046] Understandably, the two-dimensional disk coordinates refer to the two-dimensional coordinates determined on the consumer profile scale based on the endpoint of the profile vector, see reference. Figure 2As shown, the two-dimensional disk coordinates can be determined in a two-dimensional coordinate system with point O as the origin, the x-axis passing through 0 degrees as the x-axis, and the y-axis passing through 90 degrees as the y-axis. The three-dimensional vertical axis coordinate can be determined by the z-axis passing through the center of the consumer profile scale and perpendicular to the consumer profile scale. The three-dimensional vertical axis coordinate refers to the z-axis coordinate that determines the three-dimensional coordinate point. The three-dimensional coordinate system is constructed using the x-axis, y-axis, and z-axis. The z-axis can be the upward coordinate axis passing through the origin of the two-dimensional coordinate system and perpendicular to the consumer profile scale. The three-dimensional coordinate point refers to the three-dimensional coordinate representing the correspondence between historical online indicator values, historical offline indicator values, and the number of consumers. The three-dimensional coordinate point set refers to the set of coordinate points representing the correspondence between various historical online indicator values, historical offline indicator values, and the number of consumers. The three-dimensional surface refers to the three-dimensional surface representing the correspondence between historical online indicator values, historical offline indicator values, and the number of consumers.

[0047] S7. Receive the current user's current consumption data and determine the current user profile image based on the current user's consumption data.

[0048] Understandably, the "current user" refers to the user for whom consumption forecasting is currently needed. The "current user consumption data" refers to the current user's consumption data. The "current profile graph" refers to a graph representing the current user's consumption data.

[0049] In this embodiment of the invention, determining the current profile graphic based on current user consumption data includes: extracting current online indicator values ​​and current offline indicator values ​​from the current user consumption data based on the online and offline indicator groups in the online and offline indicator group set; determining the current profile angle and current profile diameter based on the current online and offline indicator values; determining the current profile vector based on the current profile angle and current profile diameter to obtain a current profile vector set; identifying the current profile endpoint set corresponding to the current profile vector set; and determining the current profile graphic based on the current profile endpoint set.

[0050] Furthermore, the current online indicator value refers to the value of the online indicator in the online and offline indicator groups corresponding to the current user consumption data, and the current offline indicator value refers to the value of the offline indicator in the online and offline indicator groups corresponding to the current user consumption data. For example, the current online indicator value and the current offline indicator value are respectively a product click-through rate of 200 times / month and a product trial frequency of 30 times / month. The current profile angle refers to the current profile vector angle determined based on the current online indicator value, and the current profile diameter refers to the current profile vector diameter determined based on the current offline indicator value. The current profile vector refers to the profile vector representing the current user under the online and offline indicator groups. The current profile vector set refers to the set of profile vectors of the current user under each online and offline indicator group. The current profile endpoint set refers to the set of profile vector endpoints corresponding to the current profile vector set. The method for determining the current profile graphic is the same as the method for determining the user profile graphic, and will not be repeated here.

[0051] S8. Based on the consumer distribution map, identify similar profile graphics corresponding to the current profile graphic in the user profile graphic set, and identify the reference shopping data corresponding to the similar profile graphics.

[0052] Understandably, the similar profile image refers to the user profile image that is closest to the current profile image.

[0053] In this embodiment of the invention, the step of identifying similar portrait graphics corresponding to the current portrait graphic in the user portrait graphic set based on the consumer distribution map includes: identifying the current vertical projection area of ​​the current portrait graphic on the consumer distribution map; extracting the current three-dimensional surface graphic of the current vertical projection area from the consumer distribution map, and determining the current three-dimensional surface volume fraction based on the current three-dimensional surface graphic and the consumer portrait scale; identifying the user vertical projection area of ​​the user portrait graphic on the consumer distribution map; extracting the user three-dimensional surface graphic of the user vertical projection area from the consumer distribution map to obtain a user three-dimensional surface graphic set; determining the user three-dimensional surface volume fraction set based on the user three-dimensional surface graphic set and the consumer portrait scale; calculating the three-dimensional volume fraction difference between each user three-dimensional surface volume fraction in the user three-dimensional surface volume fraction set and the current three-dimensional surface volume fraction to obtain a three-dimensional volume fraction difference value set; identifying the minimum three-dimensional volume fraction difference value in the three-dimensional volume fraction difference value set, and identifying the similar portrait graphics corresponding to the minimum three-dimensional volume fraction difference value.

[0054] Further, the current vertical projection area refers to the projection area obtained after the current portrait graphic is vertically projected onto the consumer distribution map. The current three-dimensional surface graphic refers to the three-dimensional surface graphic of the consumer distribution map within the current vertical projection area. The current three-dimensional surface volume fraction refers to the volume integral of the projection space when the current three-dimensional surface graphic is vertically projected onto the consumer portrait scale. The user vertical projection area refers to the projection area obtained after the user portrait graphic is vertically projected onto the consumer distribution map. The user three-dimensional surface graphic refers to the three-dimensional surface graphic of the consumer distribution map within the user vertical projection area. The user three-dimensional surface graphic set refers to the set of user three-dimensional surface graphics corresponding to each user portrait graphic. The user three-dimensional surface volume fraction set refers to the set of volume integrals of the projection space when the user three-dimensional surface graphic is vertically projected onto the consumer portrait scale. The three-dimensional volume fraction difference refers to the difference between the user three-dimensional surface volume fraction and the current three-dimensional surface volume fraction. The three-dimensional volume fraction difference set refers to the set of three-dimensional volume fraction difference values ​​between each user three-dimensional surface volume fraction and the current three-dimensional surface volume fraction. The minimum three-dimensional volume difference value refers to the smallest three-dimensional volume difference value in the set of three-dimensional volume difference values.

[0055] In this embodiment of the invention, identifying the reference shopping data corresponding to similar profile images includes: identifying the historical consumers corresponding to the similar profile images; obtaining the historical shopping data of the historical consumers; and using the historical shopping data as the reference shopping data of the current user.

[0056] Furthermore, the "historical consumer" refers to the consumer corresponding to the similar profile image. The "historical shopping data" refers to the historical online and offline shopping data of the historical consumer. The "reference shopping data" refers to shopping data used for predicting the current user's consumption.

[0057] S9. Based on reference shopping data, predict the current user's consumption and complete the construction of a consumer profile that integrates online and offline behaviors using AI.

[0058] Specifically, when the reference shopping data is as follows: 5 clothing and footwear items purchased online per month, 1 electronic product purchased online per month, 2 beauty and personal care products purchased online per month, 30 food and fresh produce items purchased offline per month, 3 home furnishing and decoration products purchased online per month, and 10 baby and toy items purchased offline per month, the current user's consumption situation can be predicted by referring to the above reference shopping data.

[0059] To address the problems described in the background art, this invention requires constructing a user profile graphic set and a current user profile graphic. Then, by combining this with a consumer distribution map, similar user profile graphics are identified within the user profile graphic set. Finally, consumption predictions for the current user are made based on reference shopping data corresponding to these similar user profile graphics. Specifically, when constructing the user profile graphic set, it is necessary to first obtain a historical user consumption dataset and an online / offline indicator set. The online / offline indicator set includes: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - actual purchase frequency. Then, historical user consumption data for each historical user is extracted sequentially from the historical user consumption dataset. To graphically represent this historical user consumption data, historical online and offline indicator values ​​can be extracted from the historical user consumption data based on the online / offline indicator sets in the online / offline indicator set. Each historical user consumption data corresponds to one historical user. Finally, a consumer profile is determined based on the historical online and offline indicator values. The process involves determining the angle and length of the consumer profile, then constructing a consumer profile vector based on these parameters, resulting in a consumer profile vector set. To comprehensively represent this vector set graphically, the corresponding consumer profile endpoint set can be identified. Based on this endpoint set, user profile graphics are determined, resulting in a user profile graphic set. Each user profile graphic corresponds to a historical user. To identify similar profile graphics corresponding to the current profile graphic, a three-dimensional representation of the current profile graphic can be created using the number of consumers and the endpoints of the profile vectors, improving the matching accuracy of similar profile graphics. Specifically, the number of consumers corresponding to the endpoints of the consumer profile vectors needs to be identified. Then, using pre-built AI surface technology, a consumer distribution map is constructed based on the endpoints of the profile vectors and the number of consumers. At this point, the current user's current consumption data can be received. The current profile graphic is then determined based on this data. Similar profile graphics corresponding to the current profile graphic are identified within the user profile graphic set based on the consumer distribution map. Finally, reference shopping data corresponding to the similar profile graphics is identified, allowing for consumption prediction of the current user based on this reference shopping data. Therefore, this invention improves the accuracy of constructing the current consumer profile.

[0060] like Figure 3 The diagram shown is a functional block diagram of a consumer profile construction system that integrates online and offline behaviors based on AI, provided in an embodiment of the present invention.

[0061] The AI-integrated online and offline behavior consumer profiling system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the AI-integrated online and offline behavior consumer profiling system 100 may include a historical online and offline indicator value determination module 101, a user profile graph set construction module 102, a consumer number distribution map construction module 103, and a current user consumption prediction module 104. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0062] The historical online and offline indicator value determination module 101 is used to acquire historical user consumption datasets and online and offline indicator sets. The online and offline indicator sets include: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency. Historical user consumption data is extracted sequentially from the historical user consumption dataset. Based on the online and offline indicator sets in the online and offline indicator sets, historical online indicator values ​​and historical offline indicator values ​​are extracted from the historical user consumption data. Each historical user consumption data corresponds to a historical user. The user profile graphic set construction module 102 is used to determine the consumption profile angle and consumption profile length based on the historical online and offline indicator values, respectively. A consumption profile vector is constructed based on the consumption profile angle and consumption profile length to obtain a consumption profile vector set. Represented by the central angle of a unit circle, the consumer profile vector starts at the circumference point on the unit circle corresponding to the consumer profile angle, and extends to the area inside the unit circle. The consumer profile endpoint set corresponding to the consumer profile vector set is identified, and user profile graphics are determined based on the consumer profile endpoint set, resulting in a user profile graphic set, where each user profile graphic corresponds to a historical user. The consumer number distribution map construction module 103 is used to identify the number of consumers corresponding to the endpoint of the consumer profile vector, and construct a consumer number distribution map based on the endpoint of the profile vector and the number of consumers using pre-built AI surface technology. The current user consumption prediction module 104 is used to receive the current user's current user consumption data, determine the current profile graphic based on the current user consumption data, identify similar profile graphics corresponding to the current profile graphic in the user profile graphic set based on the consumer number distribution map, identify reference shopping data corresponding to the similar profile graphics, and perform consumption prediction for the current user based on the reference shopping data.

[0063] In detail, the modules in the AI-integrated online and offline behavior consumer profile construction system 100 described in this embodiment of the invention employ the same methods as described above. Figure 1The method used is the same as the AI-based consumer profiling method that integrates online and offline behaviors described above, and it can produce the same technical effects, so it will not be elaborated here.

[0064] like Figure 4 The diagram shown is a structural schematic of an electronic device that implements a method for constructing consumer profiles by integrating online and offline behaviors using AI, according to an embodiment of the present invention.

[0065] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and capable of running on the processor 10, such as a consumer profile construction method program that integrates online and offline behaviors using AI.

[0066] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as the portable hard drive of the electronic device 1. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a consumer profile construction method program that integrates online and offline behavior using AI, but also to temporarily store data that has been output or will be output.

[0067] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a method for constructing consumer profiles by integrating online and offline behavior using AI), and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0068] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0069] Figure 4 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 4 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0070] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0071] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0072] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0073] The AI-integrated online and offline behavior consumer profile construction method program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When run in the processor 10, it can achieve the following: acquiring historical user consumption datasets and online and offline indicator sets, wherein the online and offline indicator sets include: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency; sequentially extracting historical user consumption data from the historical user consumption dataset, extracting historical online indicator values ​​and historical offline indicator values ​​from the historical user consumption data according to the online and offline indicator sets in the online and offline indicator sets, wherein each historical user consumption data corresponds to a historical user; determining the consumer profile angle and consumer profile length according to the historical online indicator values ​​and historical offline indicator values ​​respectively; and constructing a consumer profile according to the consumer profile angle and consumer profile length. The process involves several steps: First, a consumer profile vector set is generated. The consumer profile angle is represented by the central angle of a unit circle. The consumer profile vector originates at the point on the circumference of the unit circle corresponding to the consumer profile angle, and extends to the area within the unit circle. Next, the consumer profile endpoint set is identified. Based on this endpoint set, user profile graphics are determined, resulting in a user profile graphic set. Each user profile graphic corresponds to a historical user. Then, the number of consumers corresponding to the endpoints of the consumer profile vectors is identified. Using pre-built AI surface technology, a consumer distribution map is constructed based on the endpoints of the profile vectors and the number of consumers. Finally, the current user's current consumption data is received, and the current user profile graphic is determined based on this data. Based on the consumer distribution map, similar user profile graphics corresponding to the current user profile graphic are identified within the user profile graphic set. Reference shopping data corresponding to these similar user profile graphics is also identified. Finally, consumption predictions are made for the current user based on the reference shopping data, completing the construction of an AI-integrated online-offline consumer profile.

[0074] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 4 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0075] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0076] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor of an electronic device, the computer program can: acquire a historical user consumption dataset and a set of online and offline indicators, wherein the set of online and offline indicators includes: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency; sequentially extract historical user consumption data from the historical user consumption dataset; extract historical online indicator values ​​and historical offline indicator values ​​from the historical user consumption data according to the online and offline indicator sets in the set of online and offline indicators, wherein each historical user consumption data corresponds to a historical user; determine the consumer profile angle and consumer profile length according to the historical online indicator values ​​and historical offline indicator values ​​respectively; and construct a consumer profile vector according to the consumer profile angle and consumer profile length. A consumer profile vector set is obtained, where the consumer profile angle is represented by the central angle of a unit circle, and the consumer profile vector starts from the circumference point on the unit circle corresponding to the consumer profile angle, with the area inside the unit circle as the vector extension area. The consumer profile endpoint set corresponding to the consumer profile vector set is identified, and user profile graphics are determined based on the consumer profile endpoint set, resulting in a user profile graphic set, where each user profile graphic corresponds to a historical user. The number of consumers corresponding to the endpoints of the consumer profile vectors is identified, and a consumer distribution map is constructed based on the endpoints of the profile vectors and the number of consumers using pre-built AI surface technology. The current user's current user consumption data is received, and the current user profile graphic is determined based on this data. Similar profile graphics corresponding to the current user profile graphic are identified in the user profile graphic set based on the consumer distribution map, and reference shopping data corresponding to these similar profile graphics is also identified. Based on the reference shopping data, consumption predictions are made for the current user, completing the construction of a consumer profile that integrates online and offline behavior using AI.

[0077] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0078] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0079] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for constructing consumer profiles by integrating online and offline behavior using AI, characterized in that, The method includes: Obtain historical user consumption datasets and online and offline indicator sets, including: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency. Historical user consumption data is extracted sequentially from the historical user consumption dataset. Based on the online and offline indicator groups in the online and offline indicator group set, historical online indicator values ​​and historical offline indicator values ​​are extracted from the historical user consumption data. Each historical user consumption data corresponds to a historical user. The angle and length of the consumer profile are determined based on historical online and offline indicator values, respectively. A consumer profile vector is constructed based on the consumer profile angle and the consumer profile diameter, resulting in a consumer profile vector set. The consumer profile angle is represented by the central angle of the unit circle, and the consumer profile vector starts from the circumference point on the unit circle corresponding to the consumer profile angle, and extends from the area inside the unit circle. Identify the consumer profile endpoint set corresponding to the consumer profile vector set, determine the user profile graphic based on the consumer profile endpoint set, and obtain the user profile graphic set, where each user profile graphic corresponds to a historical user. Identify the number of consumers corresponding to the endpoint of the consumer profile vector, and use pre-built AI surface technology to construct a consumer distribution map based on the endpoint of the profile vector and the number of consumers. Receive the current user's current user consumption data, and determine the current user profile image based on the current user consumption data; Based on the consumer distribution map, identify similar profile images corresponding to the current profile image in the user profile image set, and identify the reference shopping data corresponding to the similar profile images. Based on reference shopping data, current user consumption is predicted, and a consumer profile is constructed by integrating AI with online and offline behavior; The process of constructing a consumer profile vector based on the consumer profile angle and consumer profile length yields a consumer profile vector set, including: The starting point of the profile vector is identified in the pre-constructed profile scale according to the profile angle, wherein the angle between the starting point of the profile vector and the zero mark is the profile angle. The image vector magnitude is determined based on the diameter of the consumer image, and the starting point of the image vector is identified in the inward radial direction of the consumer image scale, wherein the inward radial direction points from the starting point of the image vector to the center of the consumer image scale. Using the inward radial direction as the image vector direction, the consumer image vector is drawn in the consumer image dial according to the image vector starting point, the image vector magnitude and the image vector direction. By aggregating the consumer profile vectors corresponding to various online and offline indicator groups, a consumer profile vector set is obtained. The step of identifying similar profile images corresponding to the current profile image in the user profile image set based on the consumer distribution map includes: Identify the current vertical projection area of ​​the current portrait image on the consumer distribution map; The current three-dimensional surface graphic of the current vertical projection area is extracted from the consumer distribution map, and the current three-dimensional surface volume fraction is determined based on the current three-dimensional surface graphic and the consumer profile scale. Identify the vertical projection area of ​​the user profile graphic onto the consumer distribution map; By extracting the user's three-dimensional surface graphic from the vertical projection area of ​​the user in the consumer distribution map, a user three-dimensional surface graphic set is obtained. The user's three-dimensional surface volume diversity is determined based on the user's three-dimensional surface graphics set and the consumer profile scale. Calculate the difference between the three-dimensional volume fraction of each user's three-dimensional surface volume fraction in the user's three-dimensional surface volume fraction set and the current three-dimensional surface volume fraction to obtain a set of three-dimensional volume fraction difference values; Identify the smallest three-dimensional volume difference value in the set of three-dimensional volume difference values, and identify the similar image graphic corresponding to the smallest three-dimensional volume difference value.

2. The consumer profile construction method based on AI-integrated online and offline behavior as described in claim 1, characterized in that, The process of determining the consumer profile angle and consumer profile length based on historical online and offline indicator values ​​includes: Identify the online and offline indicator intervals corresponding to the online and offline indicator groups in the historical user consumption dataset. The consumer profile angle is calculated based on the online indicator range, historical online indicator values, and a pre-constructed profile angle formula, wherein the profile angle formula is as follows: in, This represents the consumer profile perspective corresponding to the j-th historical user consumption data in the i-th online / offline indicator group. This represents the historical online metric value corresponding to the j-th historical user's consumption data. This represents the upper limit of the online indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the online indicator interval corresponding to the i-th online and offline indicator group; The consumer profile length is calculated based on the offline indicator range, historical offline indicator values, and a pre-constructed profile length formula, wherein the profile length formula is as follows: in, This represents the length of the consumer profile corresponding to the j-th historical user's consumption data in the i-th online / offline indicator group. This represents the historical offline indicator value corresponding to the j-th historical user's consumption data. This represents the upper limit of the offline indicator interval corresponding to the i-th online-offline indicator group. This represents the lower limit of the offline indicator interval corresponding to the i-th online and offline indicator group.

3. The consumer profile construction method integrating online and offline behavior using AI as described in claim 2, characterized in that, The step of determining user profile graphics based on the consumer profile endpoint set to obtain a user profile graphic set includes: Extract the online and offline indicator groups sequentially from the set of online and offline indicator groups; Determine whether the online-offline indicator group is the last online-offline indicator group in the online-offline indicator group set; If the online and offline indicator group is not the last online and offline indicator group in the online and offline indicator group set, then identify the associated consumer profile endpoint corresponding to the online and offline indicator group in the consumer profile endpoint set, and return to the above steps of sequentially extracting online and offline indicator groups in the online and offline indicator group set. If the online and offline indicator group is the last online and offline indicator group in the set of online and offline indicator groups, then the endpoints of the associated consumer profiles are connected sequentially to obtain the user profile polyline. Identify the endpoint of the last consumer profile corresponding to the last online and offline indicator group in the online and offline indicator group set, and the endpoint of the first consumer profile corresponding to the first online and offline indicator group. Connect the endpoint of the last consumer profile and the endpoint of the first consumer profile in the user profile polyline to obtain the user profile graphic. By compiling user profile graphics corresponding to historical user consumption data, a user profile graphic set is obtained.

4. The method for constructing consumer profiles by integrating online and offline behavior using AI as described in claim 3, characterized in that, The number of consumers corresponding to the endpoint of the consumer profile vector includes: Identify the associated historical online indicator values ​​and associated historical offline indicator values ​​corresponding to the endpoint of the portrait vector; Identify the number of consumers corresponding to the associated historical online indicator values ​​and the associated historical offline indicator values ​​in the historical user consumption dataset.

5. The consumer profile construction method based on AI-integrated online and offline behavior as described in claim 4, characterized in that, The method of using pre-constructed AI surface technology to construct a consumer distribution map based on the endpoints of the profile vectors and the number of consumers includes: The two-dimensional disk coordinates are determined based on the endpoint of the image vector, and the three-dimensional vertical axis coordinates are determined based on the number of consumers. Based on the two-dimensional disk coordinates and the three-dimensional vertical axis coordinates, three-dimensional coordinate points are determined in a pre-constructed three-dimensional coordinate system to obtain a set of three-dimensional coordinate points; wherein, the x-axis and y-axis of the three-dimensional coordinate system are located in the plane of the consumer profile scale, and the z-axis of the three-dimensional coordinate system is the axis passing through the center of the consumer profile scale and perpendicular to the consumer profile scale; Using the AI ​​surface technology, a three-dimensional surface is drawn based on the set of three-dimensional coordinate points, and the three-dimensional surface is used as a distribution map of the number of consumers. The AI ​​surface technology refers to an AI fitting technology that can fit a three-dimensional spatial surface based on multiple three-dimensional spatial points.

6. The method for constructing consumer profiles by integrating online and offline behavior using AI as described in claim 5, characterized in that, The step of determining the current user profile image based on current user consumption data includes: Based on the online and offline indicator groups, extract the current online indicator values ​​and the current offline indicator values ​​from the current user consumption data. Determine the current profile angle and current profile diameter based on the current online and offline indicator values. The current image vector is determined based on the current image angle and the current image diameter, resulting in the current image vector set; Identify the current image endpoint set corresponding to the current image vector set, and determine the current image graphic based on the current image endpoint set.

7. The consumer profile construction method integrating online and offline behavior using AI as described in claim 6, characterized in that, The reference shopping data corresponding to the similar portrait images includes: Identify the historical consumers corresponding to the similar profile images; Obtain the historical shopping data of the historical consumers and use the historical shopping data as reference shopping data for the current user.

8. A consumer profiling system that integrates online and offline behavior using AI, characterized in that, The system includes: The historical online and offline indicator value determination module is used to obtain historical user consumption datasets and online and offline indicator sets. The online and offline indicator sets include: product browsing time - product dwell time, product click-through rate - product trial frequency, product keyword search frequency - product inquiry frequency, product detail page viewing frequency - cross-store comparison frequency, and product add-to-cart frequency - product actual purchase frequency. Historical user consumption data is extracted sequentially from the historical user consumption dataset. Based on the online and offline indicator sets in the online and offline indicator sets, historical online indicator values ​​and historical offline indicator values ​​are extracted from the historical user consumption data. Each historical user consumption data corresponds to one historical user. The user profile graph set construction module is used to determine the consumption profile angle and consumption profile diameter based on historical online and offline indicator values, respectively; construct consumption profile vectors based on the consumption profile angle and consumption profile diameter to obtain a consumption profile vector set, where the consumption profile angle is represented by the central angle of a unit circle, and the consumption profile vector starts at the circumference point on the unit circle corresponding to the consumption profile angle, with the area inside the unit circle as the vector extension area; identify the consumption profile endpoint set corresponding to the consumption profile vector set, and determine the user profile graph based on the consumption profile endpoint set to obtain a user profile graph set, where each user profile graph corresponds to a historical user; the construction of consumption profile vectors based on the consumption profile angle and consumption profile diameter to obtain... The consumer profile vector set includes: identifying the starting point of the profile vector in a pre-constructed consumer profile scale based on the consumer profile angle, wherein the angle between the starting point of the profile vector and the zero mark is the consumer profile angle; determining the profile vector magnitude based on the consumer profile diameter, identifying the starting point of the profile vector in the inward radial direction of the consumer profile scale, wherein the inward radial direction points from the starting point of the profile vector to the center of the consumer profile scale; using the inward radial direction as the profile vector direction, drawing the consumer profile vector in the consumer profile scale based on the starting point of the profile vector, the profile vector magnitude, and the profile vector direction; and aggregating the consumer profile vectors corresponding to various online and offline indicator groups to obtain the consumer profile vector set. The consumer distribution map construction module is used to identify the number of consumers corresponding to the endpoint of the consumer profile vector. It uses pre-built AI surface technology to construct a consumer distribution map based on the endpoint of the profile vector and the number of consumers. The current user consumption prediction module is used to receive the current user's current user consumption data, determine the current user profile image based on the current user consumption data; identify similar user profile images corresponding to the current user profile image in the user profile image set based on the consumer distribution map, and identify reference shopping data corresponding to the similar user profile images; and make consumption predictions for the current user based on the reference shopping data. The step of identifying similar user profile images corresponding to the current user profile image in the user profile image set based on the consumer distribution map includes: identifying the current vertical projection area of ​​the current user profile image on the consumer distribution map; extracting the current three-dimensional surface image of the current vertical projection area from the consumer distribution map, and predicting the current user's consumption based on the current three-dimensional surface image. The system determines the current 3D surface volume fraction using the consumer profile scale; identifies the user vertical projection area of ​​the user profile graphic on the consumer distribution map; extracts the user 3D surface graphic of the user vertical projection area from the consumer distribution map to obtain a user 3D surface graphic set; determines the user 3D surface volume fraction set based on the user 3D surface graphic set and the consumer profile scale; calculates the 3D volume fraction difference between each user 3D surface volume fraction in the user 3D surface volume fraction set and the current 3D surface volume fraction to obtain a 3D volume fraction difference value set; identifies the minimum 3D volume fraction difference value in the 3D volume fraction difference value set, and identifies the similar profile graphic corresponding to the minimum 3D volume fraction difference value.

Images