E-commerce platform intelligent guiding method for deep learning of multi-dimensional user portraits

Abstract

The invention relates to the technical field of e-commerce platforms, in particular to an e-commerce platform intelligent guiding method for deep learning of multi-dimensional user portraits, which accurately predicts user demands and performs personalized commodity recommendation by constructing demand tags and interest tags of users and combining behavior data of a target user group. A demand label and an interest label are constructed through a deep learning algorithm, demand budget of a user group is constructed by calculating the association degree of a target user group, different levels of interest guidance are carried out according to the interest intensity of a user for a specific commodity or service, and the system also carries out dynamic adjustment based on market feedback. For example, commodity pricing and promotion strategies are adjusted through forward feedback, the method can effectively improve the purchase conversion rate of the user and enhance the profitability of the platform, has a high intelligent level, and is suitable for application of the e-commerce platform in personalized marketing.

Description

Technical Field

[0001] The present invention relates to the technical field of e-commerce platforms, and in particular to an e-commerce platform intelligent guidance method based on deep learning of multi-dimensional user portraits. Background Art

[0002] An e-commerce platform refers to a third-party organization that provides merchants with a complete set of e-commerce service solutions online, including e-commerce website construction and subsequent operation services. Common apps such as Taobao, JD.com, Douyin, and Xiaohongshu are all e-commerce platforms. With the development of e-commerce companies, the daily basic maintenance of websites has also become an issue that e-commerce companies cannot ignore.

[0003] In the process of website development, data resources, content resources and user resources are becoming increasingly rich, so the situation is becoming increasingly complex. In the daily operation of e-commerce platform websites, the maintenance and optimization of tens of thousands of web pages gradually increase the difficulty of work. In the existing technology, operation and management are usually carried out through the pre-established middle platform, and the middle platform will bundle the relevant sticky data with the user. When the sticky data is offset, the offset value will be recorded, usually calculated using the distance function; Technical routes and existing problems of existing technologies: Traditional user portrait construction method: Existing technologies mostly use traditional user portrait construction methods based on rules and manual annotation, mainly by collecting basic information, historical behavior data and purchase preferences of users to build user portraits. Its technical route mainly relies on feature engineering and simple machine learning algorithms, and divides users through clustering or classification models. Although these methods can reflect user needs and interests to a certain extent, they have the following problems: 1. Data sparsity: Since traditional methods process data in a relatively static way, it is difficult to effectively capture the dynamic changes in user behavior, resulting in data sparsity problems and affecting the accuracy of the portrait.

[0004] 2. Inefficient tag construction: Tag construction usually relies on manual design and cannot deeply explore users' potential complex needs and interests, nor can it make personalized and refined recommendations.

[0005] User portrait construction method based on deep learning: With the development of deep learning, more and more e-commerce platforms have begun to adopt deep learning methods based on neural networks to build user portraits. Its technical route usually includes training through large-scale user data to automatically learn users' potential characteristics and needs. However, this method faces the following problems: 1. High demand for computing resources: Deep learning models require a lot of computing resources and time to train, especially when faced with massive amounts of user data, which places high demands on hardware configuration and storage.

[0006] 2. Complexity of label construction: Although deep learning can automatically learn features, how to design a reasonable label construction mechanism, especially in the construction of multi-dimensional labels, is still a difficult problem, which can easily lead to model overfitting or label failure.

[0007] Personalized recommendation algorithms: Most existing personalized recommendation systems use traditional methods such as collaborative filtering and content-based recommendations, which can provide accurate recommendations in some cases. However, these methods usually only focus on a single dimension of user behavior and fail to fully consider the multi-dimensional characteristics of user interests. Existing recommendation systems have the following problems: 1. Interest cold start problem: For new users or new products, the system cannot obtain sufficient historical data, resulting in poor recommendation results.

[0008] 2. Interest drift problem: User interests change over time. Existing methods cannot flexibly adapt to the dynamic changes of user interests, resulting in poor long-term effectiveness of recommendations. Summary of the invention

[0009] In view of the above problems, the present invention provides an effective solution. The purpose of the present invention is to provide an e-commerce platform intelligent guidance method based on deep learning multi-dimensional user portraits, comprising the following steps: S1. Label construction: Use deep learning methods to build a label set; S2. Calculate the relevance of target users: for the target user group, construct a target user vector group and a feature label vector, and perform weighted summation to obtain a target user association function as the target user's attention value for the e-commerce platform; S3. Calculate the target user group demand budget: Calculate the group user association and construct the group user association function. Combined with the target user association function, calculate the interest intensity ratio of the target user group. S4. Interest-guided response: Provide guidance to target users based on the interest intensity ratio of the target user group.

[0010] Furthermore, the S1 includes: S101, collecting user data, and dividing the user data into subjective data and objective data; S102. Construct an initial demand label set based on subjective data ; Wherein, the subscript i is the user group number; S103: Initial requirement tag collection Substitute into the matrix to construct the initial demand label matrix ; Wherein, the subscript i is the user group number; S104. Constructing interest tag sets based on objective data ; S105: Collect interest tags Substitute into the matrix to construct the interest tag matrix ; S106. Based on the deep learning method, the initial demand label matrix and interest tag matrix Perform cyclic training to obtain the converged demand label matrix and interest label matrix ; S107, label matrix according to convergence requirements and convergent interest label matrix , construct a generalized composite tag set ; Furthermore, the deep learning method in S106 is implemented by a gradient descent algorithm.

[0011] Furthermore, the target user in S2 is a user who is interested in the same specific product or service. The attention degree of the target user to the target product or service is determined by a target user association function, and the target user association function satisfies:

[0012] in, Indicates the target user group The attention of each user group to the target product. and Respectively represent of user groups The convergent demand label matrix value and the convergent interest label matrix value of each user; Indicates The number of users in a user group; Furthermore, the S2 specifically includes: S201: Construct a target user vector group for the target user group ,in, Indicates Target user vector for user groups; S202: construct a target user tag set based on the target user vector, and construct a target user tag vector for each target user vector. , and assign weight parameters to obtain the target user label set ; S203, calculate the target user association function; group the target user vector and feature label vector Find the angle, calculate the user relevance value, perform weighted summation on the user relevance values ​​of the same user group, and obtain the target user association function .

[0013] Furthermore, the S3 specifically includes: S301, according to the same target time, for the target user group, the target user vector group Convert to target user group vector group ,satisfy:

[0014] in, For The time span over which the smoothing function is performed, is the smoothing function, Indicates that the target user vector Smoothing is performed within the time span T, using a time window sliding algorithm; Indicates that attention will be paid As a multiplier to multiply the smoothed target user vector , get the target user group vector ; S302, calculate the relevance of the target user group, and group the target user group vectors Add and get the target user group vector ; S303, target user group vector and feature label vector Calculate the relevance and obtain the target user group association function ; S304, target user association function And the target user group association function Determine the ratio value and obtain the interest intensity ratio of the target user group, satisfying:

[0015] in, It indicates the interest intensity ratio of the target user group, that is, the interest intensity ratio of the target user group to the interest intensity ratio of the total user group.

[0016] Further, the S4 includes: When the interest intensity ratio of the target user group reaches the preset first threshold, a first-level guidance response is made; When the interest intensity ratio of the target user group reaches the preset second threshold, a secondary guidance response is made; When the interest intensity ratio of the target user group reaches the preset third threshold, a third-level guidance response is made; The third threshold is greater than the second threshold, and the second threshold is greater than the first threshold.

[0017] Furthermore, in the three-level guidance response, targeted guidance is provided to individual users, and target products or services are directly pushed.

[0018] Furthermore, in the secondary guidance response, the guidance response module performs comprehensive guidance for the target user, guides the individual user to compare a variety of similar products, and guides the individual user to complete the purchase.

[0019] Furthermore, in the first-level guidance response, category guidance is performed for the target user, and the original other product tags of the individual user are weighted to guide the individual user to turn to other products.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention is based on deep learning of user portraits to construct individual-to-individual correlation and individual-to-group correlation, which are used as basic parameters for e-commerce guidance. According to the changing individual-to-individual correlation and individual-to-group correlation, three-level linkage intelligent guidance can be made, which not only efficiently realizes the accurate capture of demand, but also is conducive to the realization of long-term marketing strategies.

[0021] 2. The deep learning of the present invention constructs a cyclic system with subjective data and objective data for learning, which not only improves the efficiency of deep learning, but also constructs a multi-dimensional labeling system that makes precise guidance possible.

[0022] 3. The present invention conducts statistical analysis based on the real-time behavioral habits of different users to construct the individual-to-group correlation and the individual-to-individual correlation. When the e-commerce platform performs precise push or guidance, the guidance content parameters will be transmitted back to the advertising end, which not only helps to optimize the system itself, but also helps the advertising end to optimize the advertising content. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The method flow chart of a specific embodiment of the present invention. DETAILED DESCRIPTION

[0024] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments: In this embodiment, a method for intelligently guiding an e-commerce platform based on deep learning of multi-dimensional user portraits is disclosed, comprising the following steps: S1. Label construction: Use deep learning methods to construct a label set; specifically including: S101, collecting user data, and dividing the user data into subjective data and objective data; S102. Construct an initial demand label set based on subjective data ; Wherein, the subscript i is the user group number; S103: Initial requirement tag collection Substitute into the matrix to construct the initial demand label matrix ; Wherein, the subscript i is the user group number; S104. Constructing interest tag sets based on objective data ; S105: Collect interest tags Substitute into the matrix to construct the interest tag matrix ; S106. Based on the gradient descent algorithm, the initial demand label matrix and interest tag matrix Perform cyclic training to obtain the converged demand label matrix and interest label matrix ; S107, label matrix according to convergence requirements and convergent interest label matrix , construct a generalized composite tag set .

[0025] S2. Calculate the relevance of target users: for the target user group, construct a target user vector group and a feature label vector, and perform weighted summation to obtain a target user association function as the target user's attention value for the e-commerce platform; The target user is a user who is interested in the same specific product or service. The target user's attention to the target product or service is determined by the target user association function. The target user association function satisfies:

[0026] in, Indicates the target user group The attention of each user group to the target product. and Respectively represent of user groups The convergent demand label matrix value and the convergent interest label matrix value of each user; Indicates The number of users in a user group; S2 specifically includes: S201: Construct a target user vector group for the target user group ,in, Indicates Target user vector for user groups; S202: construct a target user tag set based on the target user vector, and construct a target user tag vector for each target user vector. , and assign weight parameters to obtain the target user label set ; S203, calculate the target user association function; group the target user vector and feature label vector Find the angle, calculate the user relevance value, perform weighted summation on the user relevance values ​​of the same user group, and obtain the target user association function .

[0027] S3. Calculate the target user group demand budget: Calculate the group user association and construct the group user association function. Combined with the target user association function, calculate the interest intensity ratio of the target user group; specifically include: S301, according to the same target time, for the target user group, the target user vector group Convert to target user group vector group ,satisfy:

[0028] in, For The time span over which the smoothing function is performed, is the smoothing function, Indicates that the target user vector Smoothing is performed within the time span T, using a time window sliding algorithm; Indicates that attention will be paid As a multiplier to multiply the smoothed target user vector , get the target user group vector ; S302, calculate the relevance of the target user group, and group the target user group vectors Add and get the target user group vector ; S303, target user group vector and feature label vector Calculate the relevance and obtain the target user group association function ; S304, target user association function And the target user group association function Determine the ratio value and obtain the interest intensity ratio of the target user group, satisfying:

[0029] in, It indicates the interest intensity ratio of the target user group, that is, the interest intensity ratio of the target user group to the interest intensity ratio of the total user group.

[0030] S4. Interest guidance response: Based on the interest intensity ratio of the target user group, guide the target users; specifically include: When the interest intensity ratio of the target user group reaches a preset first threshold, a first-level guidance response is made; in the first-level guidance response, category guidance is performed for the target user, and the original other product tags of the individual user are weighted to guide the individual user to turn to other products; When the interest intensity ratio of the target user group reaches a preset second threshold, a secondary guidance response is made; in the secondary guidance response, the guidance response module performs comprehensive guidance on the target users, guides individual users to compare multiple similar products, and guides individual users to complete the purchase; When the interest intensity ratio of the target user group reaches a preset third threshold, a third-level guidance response is made; in the third-level guidance response, targeted guidance is provided to individual users, and target products or services are directly pushed; The third threshold is greater than the second threshold, and the second threshold is greater than the first threshold.

[0031] The market monitoring system will make different degrees of reverse operations based on the feedback from users in different markets. When the monitoring system detects positive feedback such as "99+ good reviews, 0 bad reviews, and 99+ inventory clearance", the system manager will recommend raising the price of the product; Target users: registered users of e-commerce platforms.

[0032] User groups with the same target: multiple users who have browsed, consulted, evaluated, forwarded or purchased the same product or service.

[0033] The personal information and all behavior records of each user in the same target user group are used as elements to construct a user vector and form a target user group vector group. Specific embodiment 2 An intelligent guidance method for e-commerce platforms based on deep learning of multi-dimensional user portraits: Background: An e-commerce platform hopes to improve the user purchase conversion rate through intelligent guidance. The platform provides personalized product recommendations for users through deep learning of multi-dimensional user portraits. In order to achieve accurate guidance, the platform adopts the following intelligent guidance methods.

[0035] Example: The actual operation steps of an e-commerce platform applying this method S1. Label construction: Collecting user data: E-commerce platforms obtain each user's behavioral data through the user's browsing history, search history, purchase behavior, evaluation content and other data. These data are divided into two categories: subjective data and objective data. Subjective data includes users' evaluation, likes or comments on products; objective data includes users' basic information (such as age, gender, geographic location, etc.) and behavioral data (such as browsing time, number of clicks, etc.).

[0036] Constructing an initial set of demand labels: Based on the user's subjective evaluation data, the system constructs demand labels for each user. For example, user A may mention in a comment that he "likes cost-effective mobile phones" or "pays more attention to the camera performance of mobile phones", which constructs labels such as "cost-effectiveness" and "photography" for user A.

[0037] Constructing interest tag collections: Based on the user's objective data (such as browsing history and purchase behavior), the system constructs interest tags. For example, if user A frequently browses mobile phones and buys a cost-effective smartphone, the system will tag user A with interest tags such as "smartphone" or "cost-effective".

[0038] Gradient descent training: Through deep learning algorithms, the system trains these tags, identifies each user's potential needs and interests, and generates a converged tag matrix. For example, user A's demand tag matrix may show that he prefers to choose a mobile phone with a higher cost performance, while the interest tag matrix may show that he is particularly interested in the camera function.

[0039] Constructing a generalized composite tag set: On this basis, the system combines demand tags and interest tags to generate a comprehensive tag set. For example, user A's generalized composite tags may include: "cost-effective mobile phone", "strong camera performance", "smartphone", etc.

[0040] S2. Calculate the relevance of target users: Constructing a target user vector group: The platform constructs a target user vector for each user who has interacted with a specific product. For example, if user A, user B, and user C have all browsed and purchased "cost-effective smartphones", the platform constructs a target user group vector that includes user A, B, and C.

[0041] Calculate the target user's attention: Based on the user's tag set and interest tags, the platform calculates each user's attention. For example, user A and user B have high attention, indicating that they are very interested in the target product, while user C has low attention and may just browse the product by chance.

[0042] Calculate the target user association function: The platform compares the target user vector with the product feature label and calculates the correlation. For example, the correlation between user A's vector and the label "cost-effective mobile phone" is high, so the platform determines that he is more interested in this type of product.

[0043] S3. Calculate the target user group demand budget: Convert to target user group vector group: The platform will perform time smoothing on the target user group vector. For example, if user A has been paying more attention to a certain mobile phone in the past week, his target user vector will be given a higher weight.

[0044] Calculate the relevance of the target user group: The platform adds up the demands of the target user group and calculates the overall demand intensity of the target user group. For example, if users A, B, and C all have high attention, their combined demand intensity will affect the promotion strategy of the product.

[0045] S4. Interest-guided response: First-level guided response: When the interest intensity ratio of the target user group reaches the preset first threshold, the system will recommend other similar products to the users. For example, if users A and B are more interested in "cost-effective mobile phones", the platform will recommend other cost-effective products of the same type to them, rather than just recommending the same mobile phone.

[0046] Secondary guided response: When the interest intensity ratio of the target user group reaches the second threshold, the system will provide users with more choices and guide them to compare different products. For example, users A and B start to compare mobile phones of different brands. The platform provides some detailed comparison data (such as cost performance, user reviews, photo effects, etc.) to help users make decisions.

[0047] Level 3 guided response: When the interest intensity ratio of the target user group reaches the third threshold, the platform will directly push the target product or service. For example, if user A has browsed a certain mobile phone several times and added it to the shopping cart, the platform will directly push the limited-time discount information of the mobile phone to encourage the user to place an order.

[0048] Market monitoring feedback and dynamic adjustment: When the system detects that a product has sustained high positive feedback (such as favorable comments, forwarding volume, etc.), the platform will dynamically adjust the product's pricing. For example, if the sales volume of a certain mobile phone continues to rise, the platform may raise the price to increase profits, or conduct promotional activities when there is excess inventory.

[0049] Examples of actual application scenarios: User A: 30 years old, male, has recently browsed many smartphones and is particularly interested in camera performance. The system has created labels for him such as "strong camera", "smartphone", and "cost-effective". Through deep learning algorithms, the platform recognizes that user A is highly interested in mobile phones and recommends a number of cost-effective smartphones with strong camera performance.

[0050] User B: 25 years old, female, recently browsed a variety of smartphones and noticed discount information for low-priced smartphones. The platform constructed "price sensitive" and "smartphone" tags based on her behavior data and pushed promotional information suitable for her needs, such as discounted smartphones.

[0051] Through such intelligent guidance, e-commerce platforms can more accurately predict and guide user behavior, increase product exposure and sales conversion rates, and thus enhance the platform's profitability.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement of the technical solutions that does not depart from the essence and scope of the technical solutions of the present invention should also be included in the scope of the claims of the present invention.

Claims

1. An e-commerce platform intelligent guidance method based on deep learning multi-dimensional user portraits, characterized in that: The following steps are involved: S1. Label construction: Use deep learning methods to build a label set; S2. Calculate the relevance of target users: for the target user group, construct a target user vector group and a feature label vector, and perform weighted summation to obtain a target user association function as the target user's attention value for the e-commerce platform; S3. Calculate the target user group demand budget: Calculate the group user association and construct the group user association function. Combined with the target user association function, calculate the interest intensity ratio of the target user group. S4. Interest-guided response: Provide guidance to target users based on the interest intensity ratio of the target user group.

2. According to claim 1, a method for intelligent guidance of an e-commerce platform based on deep learning of multi-dimensional user portraits is characterized by: The S1 includes: S101, collecting user data, and dividing the user data into subjective data and objective data; S102. Construct an initial demand label set based on subjective data ; Wherein, the subscript i is the user group number; S103: Initial requirement tag set Substitute into the matrix to construct the initial demand label matrix ; Wherein, the subscript i is the user group number; S104. Constructing interest tag sets based on objective data ; S105: Collect interest tags Substitute into the matrix to construct the interest tag matrix ; S106. Based on the deep learning method, the initial demand label matrix and interest tag matrix Perform cyclic training to obtain the converged demand label matrix and interest label matrix ; S107, label matrix according to convergence requirements and convergent interest label matrix , construct a generalized composite tag set .

3. According to claim 2, a method for intelligent guidance of an e-commerce platform based on deep learning of multi-dimensional user portraits is characterized by: The deep learning method in S106 is implemented by a gradient descent algorithm.

4. According to claim 3, the e-commerce platform intelligent guidance method based on deep learning multi-dimensional user portraits is characterized by: The target user in S2 is a user who is interested in the same specific product or service. The target user's attention to the target product or service is determined by a target user association function. The target user association function satisfies: ； in, Indicates the target user group The attention of each user group to the target product. and Respectively represent of user groups The convergent demand label matrix value and the convergent interest label matrix value of each user; Indicates The number of users in user groups; where Indicates that the convergence requirement label matrix value and the convergence interest label matrix value are multiplied as the value of the continuous addition operation.

5. According to claim 4, the method for intelligent guidance of an e-commerce platform based on deep learning of multi-dimensional user portraits is characterized by: The S2 specifically includes: S201: Construct a target user vector group for the target user group ,in, Indicates Target user vector for user groups; S202: construct a target user tag set based on the target user vector, and construct a target user tag vector for each target user vector. , and assign weight parameters to obtain the target user tag set ; S203, calculate the target user association function; group the target user vector and feature label vector Find the angle, calculate the user relevance value, perform weighted summation on the user relevance values ​​of the same user group, and obtain the target user association function .

6. The e-commerce platform intelligent guidance method of deep learning multi-dimensional user portrait according to claim 5 is characterized by: The S3 specifically includes: S301, according to the same target time, for the target user group, the target user vector group Convert to target user group vector group ,satisfy: ； in, For The time span over which the smoothing function is performed, is the smoothing function, Indicates that the target user vector Smoothing is performed within the time span T, using a time window sliding algorithm; Indicates that attention will be paid As a multiplier to multiply the smoothed target user vector , get the target user group vector ; S302, calculate the relevance of the target user group, and group the target user group vectors Add and get the target user group vector ; S303, target user group vector and feature label vector Calculate the relevance and obtain the target user group association function ; S304, target user association function And the target user group association function Determine the ratio value and obtain the interest intensity ratio of the target user group, satisfying: ； in, It indicates the interest intensity ratio of the target user group, that is, the interest intensity ratio of the target user group to the interest intensity ratio of the total user group.

7. The e-commerce platform intelligent guidance method of deep learning multi-dimensional user portrait according to claim 6 is characterized by: The S4 includes: When the interest intensity ratio of the target user group reaches the preset first threshold, a first-level guidance response is made; When the interest intensity ratio of the target user group reaches the preset second threshold, a secondary guidance response is made; When the interest intensity ratio of the target user group reaches the preset third threshold, a third-level guidance response is made; The third threshold is greater than the second threshold, and the second threshold is greater than the first threshold.

8. The e-commerce platform intelligent guidance method of deep learning multi-dimensional user portrait according to claim 7 is characterized by: In the three-level guidance response, targeted guidance is provided to individual users, and target products or services are directly pushed.

9. The e-commerce platform intelligent guidance method of deep learning multi-dimensional user portrait according to claim 8 is characterized in that: In the secondary guidance response, the guidance response module performs comprehensive guidance on the target user, guides the individual user to compare a variety of similar products, and guides the individual user to complete the purchase.

10. The e-commerce platform intelligent guidance method of deep learning multi-dimensional user portrait according to claim 9, characterized in that: In the first-level guidance response, category guidance is performed for target users, and the original other product tags of individual users are weighted to guide individual users to switch to other products.

Images