An internet electronic commerce transaction platform management method and system

Abstract

The application is suitable for the technical field of transaction platform management, and particularly relates to an Internet e-commerce transaction platform management method and system applied to an electronic device. The Internet e-commerce transaction platform management method comprises the following steps: receiving a user's browsing operation on a loading module; wherein the loading module is used to display a first commodity set of different kinds of commodities; in response to the browsing operation, displaying a second commodity set in the loading module; when a marking operation on a commodity in the loading module associated with the browsing operation is received, determining at least one family characteristic based on the commodity characteristic of the corresponding commodity and the second commodity set; determining a family commodity based on the family characteristic, and preloading the corresponding sequence module in the preloading module based on the corresponding order in the marking sequence and the corresponding order of the family commodity in the marking operation. The method realizes real-time response of commodity display, scenario recommendation driven by family characteristics, and intelligent preloading with sequence perception, and improves user browsing efficiency and platform conversion rate.

Description

Technical Field

[0001] The present application belongs to the technical field of transaction platform management, and in particular relates to a method and system for managing an Internet e-commerce transaction platform. Background Art

[0002] With the popularization of Internet technology, e-commerce platforms have become the main channel for commodity transactions. Modern e-commerce platforms usually include product display modules, user interaction systems, etc.

[0003] In existing technologies, product display modules typically recommend related products to users based on their past search and browsing history. For example, the platform will record all products a user searched and browsed during their last login. When the user logs in again, products that are the same or similar to the last searched product will be displayed on the homepage or recommendation page for the user to browse and select. This display method offers limited products and may not meet user product needs in a timely manner. Summary of the Invention

[0004] The embodiments of the present application provide an Internet e-commerce transaction platform management method and system, which can achieve real-time response to product display, scenario-based recommendation and sequence-aware intelligent preloading.

[0005] In a first aspect, an embodiment of the present application provides an Internet e-commerce transaction platform management method, which is applied to an electronic device, and the method includes:

[0006] receiving a user browsing operation on a loading module; wherein the loading module is used to display a first product set of different types of products;

[0007] In response to the browsing operation, displaying a second product set in the loading module; wherein at least one product in the second product set is different from a product in the first product set;

[0008] When receiving a marking operation for a product in the loading module associated with the browsing operation, determining at least one family feature based on the product feature of the product corresponding to the marking operation and the second product set;

[0009] The family products are determined based on the family characteristics, and the family products are preloaded into corresponding sequential modules in the preloading module based on the corresponding order of the marking operation in the marking sequence.

[0010] The above technical solutions in the embodiments of the present application have at least the following technical effects:

[0011] The present application provides an internet e-commerce transaction platform management method. This method receives a user's browsing operation on a loading module that displays a first set of different types of goods. In response to the browsing operation, a second set of goods, including at least one item different from the first set, is displayed at a corresponding position in the loading module, thereby updating the product display on the platform recommendation page in real time. Upon receiving a tagging operation associated with the browsing operation on a product in the loading module, at least one family characteristic is determined based on the product characteristics of the product corresponding to the tagging operation and the second set of goods. Family products are determined based on the family characteristics, and family products are preloaded into a corresponding sequence module in a preloading module based on the corresponding order of the tagging operation in the tag sequence. This method achieves real-time responsiveness to product display, scenario-based recommendations driven by family characteristics, and sequence-aware intelligent preloading, thereby improving user browsing efficiency and platform conversion rate. This method effectively addresses existing issues such as recommendations lagging behind user real-time needs, product association analysis being limited to a single attribute, and preloading strategies lacking scenario awareness. This method not only achieves an upgrade from "single product recommendation" to "demand scenario solution," but also improves page loading efficiency through sequence-aware preloading, ultimately significantly enhancing the user browsing experience and platform cross-selling capabilities.

[0012] In a possible implementation of the first aspect, displaying the second product set in the loading module in response to the browsing operation includes:

[0013] In response to the browsing operation, corresponding current product information is obtained; the current product information is used to reflect the attributes of the browsed product;

[0014] Performing attribute analysis based on the current product information to obtain iterative products;

[0015] Based on the iterated commodities, commodities at corresponding positions of the loading module are replaced, and a second commodity set is displayed in the loading module.

[0016] In a possible implementation of the first aspect, performing attribute analysis based on the current product information to obtain iterative products includes:

[0017] Constructing an attribute graph based on the association relationship between the attributes in the current product information; wherein the attribute graph has each attribute as a node, the association relationship between the attributes as an edge, and the weight of the edge represents the strength of the association between the attributes;

[0018] Calculating attribute feature values ​​based on the attribute graph;

[0019] An iterative commodity is determined from the candidate commodities based on the attribute feature value.

[0020] In a possible implementation of the first aspect, calculating the attribute feature value based on the attribute graph includes:

[0021] Sort the nodes by importance based on the attribute graph to obtain an important sequence;

[0022] Obtain the attention weight of each node based on the important sequence, and calculate the eigenvalue of each node in the important sequence;

[0023] An attribute feature value is calculated based on the attention weight and the feature value.

[0024] In a possible implementation of the first aspect, determining an iterative product from candidate products based on the attribute feature value includes:

[0025] Convert the attribute feature value of each candidate product into a vector representation to construct the candidate product attribute vector;

[0026] Calculating a similarity score between the candidate product attribute vector and the attribute feature value of the current product;

[0027] Screening candidate commodities based on a preset similarity threshold, and determining candidate commodities with similarity scores higher than the similarity threshold as a preliminary iterative commodity set;

[0028] Diversity optimization is performed on the preliminary iterative product set, and the candidate products that meet the diversity requirements are determined as iterative products.

[0029] In a possible implementation of the first aspect, determining at least one family feature based on the product feature of the current product corresponding to the marking operation and the second product set includes:

[0030] determining a first derived feature of each product in the second product set based on a correlation between the product feature of the current product corresponding to the marking operation and the second product set;

[0031] determining a second derived feature for each product in the second product set based on a correlation between the first derived feature, the product feature of the current product corresponding to the marking operation, and the second product set;

[0032] At least one family feature is determined based on the first derived feature and the second derived feature.

[0033] In a possible implementation of the first aspect, determining the first derived feature of each product in the second product set based on the correlation between the product feature of the current product corresponding to the marking operation and the second product set includes:

[0034] Calculating a first correlation between the commodities in the second commodity set based on the commodity feature of the current commodity corresponding to the marking operation to obtain a first correlation feature;

[0035] Derive a first derivative feature based on the first correlation feature and the product feature of the current product corresponding to the marking operation.

[0036] In a possible implementation manner of the first aspect, the determining the second derivative feature of each product in the second product set based on the correlation between the first derivative feature, the product feature of the current product corresponding to the marking operation and the second product set comprises:

[0037] calculating a second correlation between products in the second product set based on the first derivative feature, to obtain a second correlation feature;

[0038] deriving a second derivative feature based on the second correlation feature and the product feature of the current product corresponding to the marking operation.

[0039] In a possible implementation manner of the first aspect, the determining the at least one family feature based on the first derivative feature and the second derivative feature comprises:

[0040] constructing a feature fusion space based on the first derivative feature and the second derivative feature, and mapping the first derivative feature and the second derivative feature to a first dimension and a second dimension of the feature fusion space respectively;

[0041] calculating a fusion distance of the first derivative feature and the second derivative feature based on the feature fusion space;

[0042] performing clustering on all features based on the fusion distance, to obtain the at least one family feature.

[0043] In a second aspect, an embodiment of the present application provides an Internet electronic commerce transaction platform management system, applied to an electronic device, the system comprising:

[0044] a receiving module configured to receive a browsing operation of a user on a loading module; wherein the loading module is configured to display a first product set of different kinds of products;

[0045] a display module configured to display a second product set in the loading module in response to the browsing operation; wherein at least one product in the second product set is different from the products in the first product set;

[0046] a determining module configured to, when receiving a marking operation of a product in the loading module associated with the browsing operation, determine at least one family feature based on a product feature of a product corresponding to the marking operation and the second product set;

[0047] The preloading module is configured to determine the family products based on the family characteristics, and preload the family products into the corresponding sequence modules in the preloading module based on the corresponding order of the marking operation in the marking sequence.

[0048] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the methods described in the first aspect when executing the computer program.

[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of the above-mentioned first aspects is implemented.

[0050] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute the Internet e-commerce transaction platform management method described in any one of the first aspects above.

[0051] It can be understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0053] Figure 1 This is a flow chart of the Internet e-commerce transaction platform management method provided by the embodiment of the present application;

[0054] Figure 2 This is a schematic diagram of the implementation process of the Internet e-commerce transaction platform management method provided in the embodiment of the present application;

[0055] Figure 3 This is a schematic diagram of the structure of the Internet e-commerce transaction platform management system provided by the embodiment of the present application;

[0056] Figure 4 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0058] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0059] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0060] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if the described condition or event is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of the described condition or event" or "in response to detecting the described condition or event," depending on the context.

[0061] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0062] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0063] In the prior art, the commodity display module usually recommends relevant commodities to the user according to historical search and browsing records. For example, the platform records all the commodities that the user searched and browsed last time, and when the user logs in the platform again, the same or similar commodities as the last search are displayed on the home page or the recommendation page and the like, so as to provide the user with commodities for browsing and selection. However, due to the real-time and dynamic nature of the user's browsing behavior, the prior art only relies on historical search and browsing records for recommendation, and has the following obvious deficiencies: on the one hand, the recommendation logic lags behind the real-time demand of the user. For example, the user last browsed office supplies, but after the current login, a new demand for purchasing a digital device may be generated, and the platform still recommends office supplies based on the historical record, resulting in a disconnection between the recommended content and the current interest of the user. On the other hand, the commodity correlation analysis stays in a single dimension, and can only match similar commodities according to the basic attributes such as commodity category and brand, and cannot mine deep correlations such as functional complementarity and scene adaptation. For example, after the user browses a mouse, the traditional recommendation cannot be extended to keyboard, mouse pad and the like, limiting the possibility of cross-selling. In addition, the preloading strategy lacks perception of the real-time operation of the user, and cannot dynamically adjust the preloaded content according to the current browsing path, which easily causes invalid loading or delay in loading of key commodities, affecting the smoothness of user browsing.

[0064] More importantly, the prior art ignores the potential intention in the sequence of user behaviors. For example, the user's continuous browsing of mouse, keyboard and earphone may actually imply the demand for setting up an "office peripheral set", but the traditional recommendation only regards it as an independent commodity browsing record, and cannot form an overall scene-based recommendation, resulting in the need for the user to jump multiple times to complete commodity screening, reducing the shopping efficiency. These technical bottlenecks make it difficult for the existing e-commerce platform to meet the development needs of modern e-commerce in terms of user experience refinement and demand mining depth.

[0065] To solve the above problems, the embodiment of the present application provides an Internet e-commerce transaction platform management method and system. In this method, by receiving a user's browsing operation on a loading module for displaying a first product set of different types of products; and in response to the browsing operation, displaying a second product set having at least one product different from the first product set at a corresponding position of the loading module, so as to update the product display of the platform recommendation page in real time; when receiving a marking operation associated with the browsing operation on the product in the loading module, at least one family feature is determined based on the product features of the product corresponding to the marking operation and the second product set; based on the family feature, the family product is determined, and based on the corresponding order of the marking operation in the marking sequence, the family product is preloaded in the corresponding sequence module in the preloading module. Real-time response to product display, scenario-based recommendation driven by family features, and sequence-aware intelligent preloading are achieved, thereby improving user browsing efficiency and platform conversion rate. It effectively solves the problems in existing technologies such as recommendations lagging behind users' real-time needs, product association analysis being limited to a single attribute, and preloading strategies lacking scenario awareness. It not only achieves the upgrade from "single product recommendation" to "demand scenario solution", but also improves page loading efficiency through sequence-aware preloading, ultimately significantly enhancing user browsing experience and platform cross-selling capabilities.

[0066] The Internet e-commerce transaction platform management method provided in the embodiment of the present application can be applied to electronic devices. In this case, the electronic device is the executor of the Internet e-commerce transaction platform management method provided in the embodiment of the present application. The embodiment of the present application does not impose any restrictions on the specific type of electronic device.

[0067] For example, the electronic device may be a mobile phone, a tablet computer, a wearable device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), a desktop computer, a smart large screen, a smart TV and other terminal devices, an Internet of Things terminal, a computer, a laptop computer, a handheld communication device, a handheld computing device, a customer premises equipment (CPE) and / or other devices for communicating on a wireless system and a next-generation communication system, for example, a mobile terminal in a 5G network or a mobile terminal in a future evolved public land mobile network (PLMN).

[0068] In order to better understand the Internet e-commerce transaction platform management method provided in the embodiment of the present application, the specific implementation process of the Internet e-commerce transaction platform management method provided in the embodiment of the present application is exemplarily introduced below.

[0069] Figure 1 and Figure 2 A schematic flow chart of an Internet e-commerce transaction platform management method provided in an embodiment of the present application is shown. The Internet e-commerce transaction platform management method includes:

[0070] S100, receiving a browsing operation of a loading module by a user; wherein the loading module is used to display a first product set of different types of products.

[0071] It can be understood that the loading module is the core interactive interface used to display products on an e-commerce platform. It can be the homepage, category page, recommendation page, etc. For example, the homepage carousel and "Guess You Like" list that users first see when opening the app are all part of the loading module. The first product set is the combination of products pre-displayed by the platform based on initial strategies (such as popular products and default categories). Its types can include multiple categories (such as electronic devices, clothing, home furnishings, etc.). Users can browse the loading module by clicking, sliding, hovering, etc., and browsing operations can be captured in real time through front-end tracking technology.

[0072] S200 , in response to a browsing operation, displaying a second product set in the loading module; wherein at least one product in the second product set is different from a product in the first product set.

[0073] It can be understood that when a user browses a product in the first product set, for example, the user clicks on the display content of the product and enters the detailed introduction page of the product, when the user returns to the loading module from the detailed introduction page, the position of the product will be replaced by another product, thereby forming a new product display in the loading module, which is the second product set.

[0074] For example, by analyzing the attributes of the browsed item, a product that better meets the user's needs can be matched from the product library, and the product can be targeted and replaced to obtain a second product set. Alternatively, a product of the same category as the product can be found in the product library, and a product with at least one other attribute different from the product can be randomly selected to replace it, and so on, but not limited to these. The product library stores a massive amount of product data, and each product is annotated with complete attribute tags.

[0075] In one possible implementation, in step S200, in response to the browsing operation, displaying the second product set in the loading module includes:

[0076] S210, in response to the browsing operation, obtaining corresponding current commodity information; the current commodity information is used to reflect the attributes of the browsed commodity.

[0077] It can be understood that the current commodity information refers to the attribute data of the commodity focused on by the user in the browsing operation, including basic attributes (such as category, brand, price, color, etc.), functional attributes (such as processor model of mobile phone, camera pixel, etc.), transaction attributes (such as sales, number of evaluations, inventory, etc.), etc. For example, the user clicks on a certain "Logitech wireless mouse" in the loading module, and the system will obtain the complete attribute data of the mouse through the URL parameter of the commodity detail page or the callback function of the click event. These data are usually stored in the distributed database (such as MongoDB) of the platform and are quickly queried through the commodity ID index. At the same time, the browsing context of the user, such as browsing duration, viewing detail page, and other behavior data, can be recorded to more comprehensively reflect the user's interest degree in the commodity or the commodity of the same type.

[0078] S220, performing attribute analysis according to the current commodity information to obtain an iteration commodity.

[0079] It can be understood that attribute analysis is to analyze the correlation between commodity attributes by data mining technology, so as to predict the iteration commodity that the user may be interested in. For example, for "Logitech wireless mouse", its attributes include "wireless connection", "DPI4000", "compatible with Windows / macOS", etc. The system will compare these attributes with the attributes of other commodities in the platform, and analyze which attributes have greater influence on user decision-making. Exemplarily, the attribute relationship graph of the commodity can be constructed, the attribute characteristics of the commodity are analyzed according to the attribute graph, the attribute characteristics that the user is more interested in are analyzed, and the corresponding iteration commodity is matched; or the commodity attributes are analyzed, the importance of the attribute keywords is calculated using the TF-IDF algorithm, the high-frequency and distinguishing attributes such as "wireless" and "high DPI" are identified, and the iteration commodity is screened, etc., but not limited thereto.

[0080] In one possible implementation, in step S220, attribute analysis is performed according to the current commodity information to obtain an iteration commodity, including:

[0081] S221, constructing an attribute graph according to the correlation between attributes in the current commodity information; wherein the attribute graph takes each attribute as a node, the correlation between attributes as an edge, and the weight of the edge represents the correlation strength between attributes.

[0082] It can be understood that the attribute graph is a data model that abstracts commodity attributes into a graph structure, taking attributes (such as "brand" and "price") as nodes and the correlation between attributes as edges.

[0083] S222, calculating attribute characteristic values based on the attribute graph.

[0084] It can be understood that attribute feature values ​​are a quantitative representation of the importance and association strength of nodes in the attribute graph, and are used to measure the contribution of attributes to product recommendations. For example, by sorting the nodes (attributes) by importance, the core attributes can be identified (e.g., "wireless connection" is more important than "color" in mouse recommendations); secondly, the feature values ​​are calculated for each node, such as the mean and standard deviation of numerical attributes (price), and the frequency of occurrence and information entropy of categorical attributes (brand); finally, the attribute feature values ​​are obtained by weighted summation combined with the importance weights of the nodes; the attribute graph can also be input into the learning model, and the learning model outputs the corresponding attribute feature values, etc., but not limited to this. The learning model is trained using multiple sets of training data, each of which includes an attribute graph and corresponding attribute feature values.

[0085] In a possible implementation, in step S222, calculating the attribute feature value based on the attribute graph includes:

[0086] S2221, sort the nodes by importance based on the attribute graph to obtain an important sequence.

[0087] It's understandable that node importance ranking uses a graph algorithm to evaluate the influence of attributes within the entire attribute graph. Each attribute node initially receives the same importance score, which is then iterated based on edge weights (attribute association strength). If attribute A is connected to multiple high-weight attributes, A's importance score will increase. For example, in the mouse attribute graph, the "Wireless Connection" node is strongly associated with nodes like "Battery Life" and "Bluetooth Version." After multiple iterations, its importance score significantly exceeds that of edge attributes like "Appearance Color," resulting in a ranking order of importance.

[0088] S2222: Obtain an attention weight of each node based on the important sequence, and calculate a characteristic value of each node in the important sequence.

[0089] As you can understand, attention weights are weight coefficients assigned to each attribute node based on the importance sequence, and are normalized so that the sum of the weights is 1. For example, the weights of the first three attributes in the importance sequence are 0.4, 0.3, and 0.2, respectively, and the remaining attributes total 0.1. Node feature values ​​are calculated differently depending on the attribute type: for numerical attributes (such as a price of 200 yuan), the absolute difference or relative ratio between the value and the current product attribute is calculated; for categorical attributes (such as the brand "Logitech"), the sales share of that brand's products on the platform or the user preference index is calculated.

[0090] S2223, calculating the attribute feature value based on the attention weight and the feature value.

[0091] It can be understood that attribute eigenvalue = attention weight × eigenvalue.

[0092] With this setup, through attribute graph modeling and eigenvalue calculation, the system can transform the complex associations between product attributes into quantifiable numerical features, breaking through the limitations of traditional recommendations based solely on single attribute matching. For example, when a user browses for "wireless mouse," this method not only focuses on the "wireless" attribute but also discovers, through the attribute graph, the strong correlation between "wireless connection" and attributes such as "battery life" and "cross-device compatibility." This prioritizes mice with these characteristics among candidate products, making the recommendations more tailored to the user's potential needs. Furthermore, dynamically adjusted attention weights ensure the dominant role of core attributes in recommendations, improving the accuracy and efficiency of recommendations.

[0093] S223: Determine an iterative product from the candidate products based on the attribute feature values.

[0094] It can be understood that iterative products refer to candidate products that are screened based on attribute feature values ​​and are highly relevant to the current product. For example, the attribute feature values ​​of all candidate products can be converted into vectors, and then the similarity with the attribute feature values ​​of the current product can be calculated through algorithms such as cosine similarity and Euclidean distance, and products above the threshold are screened as the preliminary iterative set. For example, when the threshold is set to 0.6, candidate products with a similarity of 0.65 are included in the preliminary set. In order to avoid homogenization of recommendation results, the system will also optimize the diversity of the preliminary set, such as using a clustering algorithm to ensure that the recommended products are evenly distributed in terms of price range, brand, function and other dimensions, and finally determine the iterative products; products that are above the threshold and have the maximum value can also be directly used as iterative products, etc., but are not limited to this.

[0095] With this setting, by constructing an attribute graph based on the correlation between the attributes in the current product information, with attributes as nodes, correlation relationships as edges and edge weights representing correlation strength, and then calculating attribute characteristic values ​​based on the attribute graph, and then determining iterative products among the candidate products, the complex correlations between product attributes can be quantified in a graph structure, and the correlation strength between attributes can be accurately captured. The importance and correlation degree of product attributes can be quantitatively analyzed through the calculation of attribute characteristic values, so as to screen out iterative products among the candidate products that are highly matched with the current product attribute characteristics and meet the user's potential needs, effectively improving the accuracy and relevance of product recommendations, making the recommendation results more in line with the user's actual needs, and enhancing the user's browsing experience and platform recommendation efficiency.

[0096] In a possible implementation, in step S223, determining an iterative product from candidate products based on attribute feature values ​​includes:

[0097] S2231, converting the attribute feature value of each candidate product into a vector representation to construct a candidate product attribute vector.

[0098] It can be understood that converting the attribute feature values ​​of candidate products into vector representations maps multidimensional attribute features into points in a mathematical space, facilitating the quantitative calculation of similarity between products. The specific process is as follows: First, the attribute feature values ​​of each candidate product are calculated based on the attribute graph (e.g., "wireless connection" has a value of 0.9, "price" has a value of 0.7, and so on). These feature values ​​are then arranged in a fixed order to form a one-dimensional vector. For example, if the attribute feature values ​​of a candidate mouse include [wireless connection: 0.9, price: 0.7, brand: 0.8, DPI: 0.6], its attribute vector is [0.9, 0.7, 0.8, 0.6]. The vector dimensions typically correspond to the number of core attributes in the attribute graph. Normalization (e.g., Z-score normalization) or normalization (e.g., Min-Max normalization) is used to ensure comparability of feature values ​​of different dimensions. This vector construction process provides a structured data foundation for subsequent similarity calculations, enabling efficient processing of complex attribute relationships between products through vector operations.

[0099] S2232, calculating the similarity score between the candidate product attribute vector and the attribute feature value of the current product.

[0100] As you can understand, the similarity score is used to measure the degree of match between the candidate product attribute vector and the current product attribute feature value. The comprehensive similarity score is generated by calculating the sum of the absolute values ​​of the differences in each dimension of the vector.

[0101] S2233 , screening candidate commodities based on a preset similarity threshold, and determining candidate commodities with similarity scores higher than the similarity threshold as a preliminary iterative commodity set.

[0102] It is understood that the preset similarity threshold can be obtained from the product database or manually entered, etc., but is not limited to this. This threshold can be adjusted dynamically: lowered during promotional events to expand the recommendation range, and raised during daily operations to ensure recommendation accuracy. The screening process batch processes the similarity scores of all candidate products through matrix operations. For example, NumPy vector operations are used to quickly filter out a list of matching product IDs, and then retrieve the corresponding product information from the product database. The initial iterative product set may contain a large number of similar products (such as mice from the same brand and series).

[0103] S2234: Perform diversity optimization on the preliminary iterative product set and determine candidate products that meet the diversity requirements as iterative products.

[0104] Diversity optimization involves introducing differentiated attributes into similar products to ensure that recommendations cover diverse dimensions. This can be achieved by calculating the information entropy of attribute distribution and then using an optimization algorithm to maximize this entropy (i.e., achieve the most even distribution of attributes). For example, the distribution entropy of attributes such as "wireless / wired," "price range," and "brand" in the recommendation set can reach a preset target. After optimization, the iterative product set can maintain high similarity with the current product while also covering different user preference scenarios. For example, when recommending a high-end wireless mouse, it can also include a mid-range wireless mouse and a keyboard from the same brand, expanding user choice.

[0105] With this setup, a closed loop from attribute features to precise recommendations is achieved through the process of "vector representation-similarity calculation-threshold screening-diversity optimization". The attribute vector and similarity calculation ensure that the iterated product is highly relevant to the user's current browsing product. For example, when a user browses "Logitech wireless mouse", wireless mice of the same brand with high similarity are recommended first; diversity optimization introduces differentiated attributes, such as adding options of different price ranges or functional types to similar mice, guiding users to explore more possibilities. Recommendation results that take into account both accuracy and diversity can not only meet users' deep needs for specific products, but also stimulate potential purchasing intentions. Through data structuring and algorithm optimization, abstract user needs are converted into computable quantitative indicators, realizing intelligent and personalized e-commerce recommendations.

[0106] S230: Based on the iterated products, replace the products at the corresponding positions of the loading module, and display a second product set in the loading module.

[0107] As you can understand, the determined iterated products are replaced with the original products in the loading module through front-end rendering technology, forming a second product set. In specific implementation, the backend generates JSON data containing the iterated product ID, attributes, price, and other information, and returns it to the front-end via an API. Upon receiving this data, the front-end framework (such as React) compares the new and old product lists using a virtual DOM diff algorithm, updating only the changed parts to reduce page redrawing overhead. For example, if the loading module originally had 10 products and 3 of them were replaced with the iterated products, the system would only re-render the components of these 3 products, leaving the other products unchanged to ensure smooth page updates.

[0108] With this setting, this real-time interaction reduces the user's search cost and improves browsing efficiency. At the same time, it reduces the client's performance consumption through local update technology, ensuring a smooth user experience.

[0109] S300: When a marking operation on a product in a loaded module is received in association with a browsing operation, at least one family feature is determined based on product features of the product corresponding to the marking operation and a second product set.

[0110] It's understood that tagging includes actions such as adding to favorites, adding to carts, and giving likes, which indicate a user's explicit preference for a product. When a user tags a product in the loaded module, the system extracts the product's characteristics (e.g., the attributes of a "Logitech wireless mouse") and the current set of second products, analyzes their correlation, and generates "family characteristics." Family characteristics are composite features shared by a group of products that reflect potential user needs, such as "wireless peripherals" and "office suite."

[0111] In a possible implementation, in step S300, determining at least one family feature based on the product feature of the product corresponding to the marking operation and the second product set includes:

[0112] S310 : Determine a first derived feature of each product in the second product set based on the correlation between the product feature of the current product corresponding to the marking operation and the second product set.

[0113] It can be understood that the first derived feature is an intermediate feature generated based on the direct correlation between the current product feature and the second product set. For example, principal component analysis (PCA) dimensionality reduction or a neural network can be used to automatically extract high-order features to calculate the attribute similarity between each product in the second product set and the current product. Then, the correlation feature can be derived based on the current product feature to generate a derived feature vector. Alternatively, a graph neural network (GNN) can be used to calculate the semantic distance between each product in the second product set and the current product in the knowledge graph. For example, the semantic path between keyboard and mouse is "mouse → functional complement → keyboard," with the accumulated path weight being 0.7 + 0.6 = 1.3, which translates to a similarity score of 1.3 / 2 = 0.65 (normalized), etc., but the present invention is not limited to these.

[0114] In one possible implementation, in step S310, based on the correlation between the product feature of the current product corresponding to the marking operation and the second product set, determining the first derived feature of each product in the second product set includes:

[0115] S311 , calculating a first correlation between commodities in the second commodity set based on a commodity feature of the current commodity corresponding to the marking operation, and obtaining a first correlation feature.

[0116] It can be understood that the similarity between the feature matrix of each product in the second product set and the currently marked product can be calculated using cosine similarity or Jaccard coefficient. For example, after vectorizing structural features such as product title, category, and price range, a similarity algorithm is used to generate an N×N correlation coefficient matrix, where the diagonal elements are the product features of the current product.

[0117] S312: derive the first correlation feature based on the product feature of the current product corresponding to the marking operation to obtain a first derived feature.

[0118] It can be understood that the current product features and the first correlation features can be concatenated into a high-dimensional vector, input into a 3-layer MLP network for nonlinear mapping, and the complex interaction relationship between features can be learned through the ReLU activation function and the Adam optimizer, and finally the low-dimensional first derivative features can be output. For example,

[0119] With this setting, the first correlation feature of each product in the second product set is calculated based on the current product feature corresponding to the marking operation, and the first derived feature is further derived from it, so that the similarity of the explicit attributes of the products is converted into a derived feature that includes implicit needs such as user temporal preferences and semantic associations, thereby accurately capturing the user's potential needs, improving the semantic understanding ability and personalization of product recommendations, and enhancing the recommendation system's adaptability to complex user behaviors. At the same time, through feature dimensionality reduction and dynamic updating, the computing efficiency is optimized, so that the recommendation results are more in line with the actual usage scenarios, and the accuracy of platform recommendations and user experience are effectively improved.

[0120] S320 : Determine a second derived feature for each product in the second product set based on the first derived feature, the product feature of the current product corresponding to the marking operation, and the correlation between the second product set.

[0121] It can be understood that the second derived feature is a higher-order feature generated by further integrating the deep semantic association between the current product features and the second product set based on the first derived feature. For example, a correlation coefficient matrix can be generated using the first derived feature, the product features of the current product corresponding to the tagging operation, and the second product set, and the eigenvectors of the matrix can be processed to determine the second derived feature. Alternatively, the first derived feature, the product features of the current product corresponding to the tagging operation, and the second product set can be input into a learning model, with the learning model outputting the corresponding second derived feature, and so on, but the present invention is not limited thereto.

[0122] In one possible implementation, in step S320, based on the first derived feature, the product feature of the current product corresponding to the marking operation, and the correlation between the second product set, determining the second derived feature of each product in the second product set includes:

[0123] S321 , calculating a second correlation between commodities in a second commodity set based on the first derived feature to obtain a second correlation feature.

[0124] It can be understood that the knowledge graph relationship is introduced on the basis of the first derived feature. For example, the weight scores of complex relationships such as "functional complementarity-0.7 / brand association-0.3" between products are calculated through the TransE algorithm to generate a second correlation coefficient matrix containing semantic associations. The vector of the matrix is ​​the second correlation feature.

[0125] S322: Derive the second correlation feature based on the current product feature corresponding to the marking operation to obtain a second derived feature.

[0126] It's understandable that the original features of the current product (such as brand, price, and functional parameters) can be concatenated with secondary correlation features (such as the "functional complementarity score of 0.7" and "brand association score of 0.3" calculated by the TransE algorithm) to form a high-dimensional vector. The product features and their relationships (such as "brand homology" and "functional complementarity") can then be mapped into low-dimensional vectors using the TransE algorithm. For example, if the current product is the "Huawei P60 mobile phone," the first derived feature has calculated the "attribute matching degree of 0.7" for a certain Huawei headset. The second derived feature further incorporates the brand homology relationship (weight 0.8) between "Huawei mobile phone and Huawei headset" and the scenario complementarity relationship (weight 0.6) between "mobile phone photography and headset noise reduction" in the knowledge graph. Through the TransE algorithm and MLP network processing, a second derived feature vector containing dimensions such as "ecological adaptability 0.85," "scenario fit 0.72," and "brand loyalty 0.9" is generated.

[0127] With this setting, the explicit attribute associations of products are transformed into high-level features that contain implicit semantics such as users' deep needs, ecological adaptation, and scenario preferences, thereby accurately capturing users' potential ecological chain needs and scenario preferences, improving the semantic understanding and personalization of product recommendations, and enhancing the recommendation system's adaptability to complex user behaviors. At the same time, through feature dimensionality reduction and dynamic updates, the computing efficiency is optimized, making the recommendation results more in line with actual usage scenarios, effectively improving the accuracy of platform recommendations and user experience.

[0128] S330: Determine at least one family feature based on the first derived feature and the second derived feature.

[0129] For example, the first derived feature (such as attribute similarity, temporal preference) and the second derived feature (such as ecological adaptability, scene fit) can be mapped to a unified feature fusion space to form a composite feature vector containing "explicit attribute matching" and "implicit semantic association"; then, the distance of the feature vectors in the fusion space is calculated by Euclidean distance, cosine similarity and other methods to measure the comprehensive similarity between products; finally, hierarchical clustering or K-means algorithm is used to determine the feature vectors whose distance meets the preset conditions as family features; the first derived feature and the second derived feature can also be input into a learning model, and the learning model outputs the corresponding family features, etc., but not limited to this.

[0130] With this setting, the first derived feature is determined based on the correlation between the current product feature and the second product set, and the second derived feature is determined based on the correlation between the first derived feature, the current product feature and the second product set. Finally, the family feature is determined based on the two types of derived features. The explicit and implicit associations between products are mined from multiple dimensions such as product attribute matching, semantic association, and user temporal preference. The single product feature is transformed into a derived feature that includes deep needs such as ecological adaptation and scenario preference. Semantic families are generated through feature fusion and clustering, so as to accurately capture the user's potential needs, improve the semantic understanding ability and personalization of the recommendation system, and enhance the adaptability to complex user behaviors and market dynamics. At the same time, feature dimensionality reduction and dynamic clustering are used to optimize computing efficiency, so that the recommendation results are more in line with the actual usage scenarios, effectively improving the platform's recommendation accuracy, user experience and operational efficiency.

[0131] In a possible implementation, in step S330, determining at least one family feature based on the first derived feature and the second derived feature includes:

[0132] S331 , constructing a feature fusion space based on the first derived feature and the second derived feature, and mapping the first derived feature and the second derived feature to a first dimension and a second dimension of the feature fusion space, respectively.

[0133] It can be understood that the K1 dimension of the first derived feature (such as K1=5, representing brand matching, price matching, etc.) is mapped to the first K1 dimension of the fusion space, and the K2 dimension of the second derived feature (such as K2=3, representing ecological adaptation, scene fit, etc.) is mapped to the latter K2 dimension to form a K1+K2-dimensional fusion feature vector.

[0134] S332: Calculate the fusion distance between the first derived feature and the second derived feature based on the feature fusion space.

[0135] It can be understood that the fusion distance can be obtained by calculating the Euclidean distance between feature vectors.

[0136] S333: Cluster all features based on the fusion distance to obtain at least one family feature.

[0137] As you can understand, the distances between all products are calculated based on the fusion distance to form an N×N distance matrix. For example, the matrix size for 100 products is 100×100. Starting with each product as an independent cluster, the closest clusters are gradually merged. For example, "digital products" and "electronic accessories" have the smallest distance, so they are first merged into the "electronics cluster." The process stops when the inter-cluster distance exceeds the threshold or the number of clusters reaches the preset value K (such as K=5). For example, five clusters are ultimately formed, such as the "electronics cluster" and the "audio equipment cluster." The product features within each cluster are then averaged to obtain the cluster feature vector. The common attributes of the products within the cluster are then analyzed to generate semantic labels for the cluster, which are the cluster features.

[0138] With this setup, the first and second derivative features are mapped to different dimensions by constructing a feature fusion space, and the fusion distance is calculated and clustered to determine the family features, thus realizing the transformation from single product features to structured semantic families. This solution uses multi-dimensional feature mapping to retain comprehensive information such as product attribute similarity, temporal preference, and ecological adaptation. It accurately measures the deep associations between products through fusion distance, and then uses clustering algorithms to automatically identify natural groupings of products such as functional complementary families and brand ecological families. It not only improves the diversity and accuracy of product recommendations and shortens the user decision-making path, but also dynamically adapts to market changes such as new product launches. At the same time, it greatly optimizes storage and computing efficiency through dimensionality reduction clustering, significantly enhancing the intelligent operation capabilities and user experience of e-commerce platforms.

[0139] S400 , determining family products based on family features, and preloading family products into corresponding sequence modules in a preloading module based on corresponding sequences of marking operations in a marking sequence.

[0140] It can be understood that the preloaded module refers to the product display interface after the platform's homepage and recommendation pages. Products can be screened by calculating the cosine similarity between the candidate product feature vector and the family feature vector, and then setting a preset similarity. For example, for the "mobile phone accessories family," the similarity threshold is set to 0.85. Its family features are [0.75, 0.65, 0.7, 0.8, 0.55, 0.65, 0.5, 0.55]. A charger has features of [0.8, 0.7, 0.75, 0.85, 0.6, 0.7, 0.55, 0.6], and a similarity of 0.98, making it a product within the family. A Bluetooth headset has a similarity of 0.75 and is excluded from the "mobile phone accessories family." And so on. For example, let's assume the tag sequence is [phone → charger → phone case], with a corresponding order of 1 → 2 → 3. When a user tags "phone" (order 1), module 1 is triggered to preload, extracting the top three highly similar products from the "Mobile Accessories" category (such as fast chargers, magnetic chargers, and wireless chargers). When a user tags "charger" (order 2), module 2 is triggered to preload, extracting the top two highly relevant products from the "Audio Devices" category (such as noise-canceling headphones and sports headphones), and so on. Furthermore, preloading priority can be adjusted based on popularity metrics such as product sales and ratings. For example, if sales of a fast charger increase by 30% week-over-week, the preloading weight is increased to 1.2. The order module's product pool is dynamically adjusted based on actual user click behavior. For example, if a user frequently skips products in the "Protective Accessories" category, its preloading weight can be reduced.

[0141] This setting, through multi-dimensional feature derivation and clustering technology, improves the personalization and response speed of product recommendations, and transforms traditional category-based static recommendations into dynamic semantic recommendations based on user behavior. It not only enhances the depth of interaction between users and the platform, but also optimizes system performance through the preloading mechanism, making the recommendation results more in line with users' actual needs, effectively improving product conversion rate and user satisfaction.

[0142] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0143] Corresponding to the Internet e-commerce transaction platform management method described in the above embodiment, the embodiment of the present application also provides an Internet e-commerce transaction platform management system, and each module of the system can implement each step of the Internet e-commerce transaction platform management method. Figure 3 A structural block diagram of the Internet e-commerce transaction platform management system provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0144] Reference Figure 3 , the Internet e-commerce transaction platform management system includes:

[0145] The receiving module is used to receive a user's browsing operation on the loading module; wherein the loading module is used to display a first product set of different types of products.

[0146] The display module is configured to display the second product set in the loading module in response to the browsing operation; wherein at least one product in the second product set is different from the products in the first product set.

[0147] The determination module is configured to, upon receiving a marking operation on a product in the loading module associated with the browsing operation, determine at least one family feature based on the product feature of the product corresponding to the marking operation and the second product set.

[0148] The preloading module is used to determine the family products based on the family characteristics, and preload the family products into the corresponding sequence modules in the preloading module based on the corresponding order of the marking operation in the marking sequence.

[0149] It should be noted that the information interaction, execution process and other contents between the above modules are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0150] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above modules is used as an example for illustration. In actual applications, the above functions can be distributed and completed by different modules as needed, that is, the internal structure of the system can be divided into different modules to complete all or part of the functions described above. The modules in the embodiment can be integrated into one processing unit, or each module can exist physically alone, or two or more modules can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of the modules are only for the convenience of distinguishing each other and are not used to limit the scope of protection of this application. The specific working process of the modules in the above system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0151] The embodiment of the present application also provides an electronic device, Figure 4 This is a schematic diagram of the structure of an electronic device 6 provided in one embodiment of the present application. Figure 4 As shown, the electronic device 6 of this embodiment includes: at least one processor 60 ( Figure 4 Only one is shown), at least one memory 61 ( Figure 4 Only one is shown in the figure) and a computer program 62 stored in the at least one memory 61 and executable on the at least one processor 60. When the processor 60 executes the computer program 62, the electronic device 6 implements the steps of any of the above-mentioned embodiments of the Internet e-commerce transaction platform management method, or implements the functions of the modules in the above-mentioned system embodiments.

[0152] For example, the computer program 62 may be divided into one or more modules / units, which are stored in the memory 61 and executed by the processor 60 to implement the present application. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program 62 in the electronic device 6.

[0153] The electronic device 6 can be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The electronic device can include, but is not limited to, a processor 60 and a memory 61. Those skilled in the art will understand that Figure 4 It is only an example of the electronic device 6 and does not constitute a limitation on the electronic device 6. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, buses, etc.

[0154] The processor 60 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0155] In some embodiments, the memory 61 may be an internal storage unit of the electronic device 6, such as a hard drive or memory of the electronic device 6. In other embodiments, the memory 61 may also be an external storage device of the electronic device 6, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped on the electronic device 6. Furthermore, the memory 61 may include both an internal storage unit of the electronic device 6 and an external storage device. The memory 61 is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory 61 may also be used to temporarily store data that has been output or is about to be output.

[0156] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0157] An embodiment of the present application provides a computer program product. When the computer program product is run on an electronic device, the electronic device implements the steps of any of the above method embodiments.

[0158] If the integrated unit is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the process steps in the above-mentioned method embodiments by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to an electronic device, a recording medium, computer memory, read-only memory (ROM), random access memory (RAM), an electrical carrier signal, a telecommunications signal, and a software distribution medium. Examples include a USB flash drive, a removable hard drive, a magnetic disk, or an optical disk.

[0159] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0160] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0161] In the embodiments provided in this application, it should be understood that the disclosed electronic device and Internet e-commerce transaction platform management system can be implemented in other ways. For example, the Internet e-commerce transaction platform management system embodiment described above is only illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0162] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0163] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for managing an Internet e-commerce transaction platform, applied to electronic equipment, characterized in that: The method comprises: receiving a user browsing operation on a loading module; wherein the loading module is used to display a first product set of different types of products; In response to the browsing operation, displaying a second product set in the loading module; wherein at least one product in the second product set is different from a product in the first product set; When receiving a mark operation associated with the browsing operation for a product in the loaded module, determining at least one family feature based on the product features of the product corresponding to the mark operation and the second product set; wherein the family feature is used to indicate a composite feature shared by products and capable of reflecting potential needs of the user; Determining family products based on the family characteristics, and preloading the family products in corresponding sequence modules in the preloading module based on the corresponding order of the marking operation in the marking sequence; In response to the browsing operation, displaying the second product set in the loading module includes: In response to the browsing operation, corresponding current product information is obtained; the current product information is used to reflect the attributes of the browsed product; Performing attribute analysis based on the current product information to obtain iterative products; Based on the iterated products, the products at the corresponding positions of the loading module are replaced, and a second set of products is displayed in the loading module; The performing attribute analysis based on the current product information to obtain iterative products includes: Constructing an attribute graph based on the association relationship between the attributes in the current product information; wherein the attribute graph has each attribute as a node, the association relationship between the attributes as an edge, and the weight of the edge represents the strength of the association between the attributes; Calculating attribute feature values ​​based on the attribute graph; Determining an iterative product among candidate products based on the attribute feature value; The calculating the attribute feature value based on the attribute graph includes: Sort the nodes by importance based on the attribute graph to obtain an important sequence; Obtain the attention weight of each node based on the important sequence, and calculate the eigenvalue of each node in the important sequence; An attribute feature value is calculated based on the attention weight and the feature value.

2. The Internet e-commerce transaction platform management method according to claim 1, characterized in that: The step of determining an iterative commodity from candidate commodities based on the attribute feature value includes: Convert the attribute feature value of each candidate product into a vector representation to construct the candidate product attribute vector; Calculating a similarity score between the candidate product attribute vector and the attribute feature value of the current product; Screening candidate commodities based on a preset similarity threshold, and determining candidate commodities with similarity scores higher than the similarity threshold as a preliminary iterative commodity set; Diversity optimization is performed on the preliminary iterative product set, and the candidate products that meet the diversity requirements are determined as iterative products.

3. The Internet e-commerce transaction platform management method according to claim 1, characterized in that: The determining at least one family feature based on the product feature of the current product corresponding to the marking operation and the second product set includes: determining a first derived feature of each product in the second product set based on a correlation between the product feature of the current product corresponding to the marking operation and the second product set; determining a second derived feature for each product in the second product set based on a correlation between the first derived feature, the product feature of the current product corresponding to the marking operation, and the second product set; At least one family feature is determined based on the first derived feature and the second derived feature.

4. The Internet e-commerce transaction platform management method according to claim 3, characterized in that: The determining, based on the correlation between the product feature of the current product corresponding to the marking operation and the second product set, a first derived feature of each product in the second product set includes: Calculating a first correlation between the commodities in the second commodity set based on the commodity feature of the current commodity corresponding to the marking operation to obtain a first correlation feature; The first correlation feature is derived based on the product feature of the current product corresponding to the marking operation to obtain a first derived feature.

5. The Internet e-commerce transaction platform management method according to claim 3, characterized in that: The determining, based on the first derived feature, the correlation between the feature of the current product corresponding to the marking operation and the second product set, a second derived feature of each product in the second product set includes: Calculating a second correlation between the commodities in the second commodity set based on the first derived feature to obtain a second correlation feature; The second correlation feature is derived based on the current product feature corresponding to the marking operation to obtain a second derived feature.

6. The Internet e-commerce transaction platform management method according to claim 3, characterized in that: The determining of at least one family feature based on the first derived feature and the second derived feature comprises: constructing a feature fusion space based on the first derived feature and the second derived feature, and mapping the first derived feature and the second derived feature to a first dimension and a second dimension of the feature fusion space respectively; Calculating a fusion distance between the first derived feature and the second derived feature based on the feature fusion space; Based on the fusion distance, all features are clustered to obtain at least one family feature.

7. An Internet e-commerce transaction platform management system, applied to electronic equipment, characterized in that: The system comprises: A receiving module, configured to receive a user's browsing operation on a loading module; wherein the loading module is configured to display a first set of commodities of different categories; a display module configured to display a second product set in the loading module in response to the browsing operation; wherein at least one product in the second product set is different from a product in the first product set; a determination module configured to, upon receiving a mark operation associated with the browsing operation for a product in the loading module, determine at least one family characteristic based on product characteristics of the product corresponding to the mark operation and the second product set; wherein the family characteristic is used to indicate a composite characteristic shared by products and capable of reflecting potential needs of the user; A preloading module, configured to determine family products based on the family characteristics, and preload the family products in corresponding sequence modules in the preloading module based on the corresponding order of the marking operation in the marking sequence; Wherein, the display module is further used for: In response to the browsing operation, corresponding current product information is obtained; the current product information is used to reflect the attributes of the browsed product; Performing attribute analysis based on the current product information to obtain iterative products; Based on the iterated products, the products at the corresponding positions of the loading module are replaced, and a second set of products is displayed in the loading module; The performing attribute analysis based on the current product information to obtain iterative products includes: Constructing an attribute graph based on the association relationship between the attributes in the current product information; wherein the attribute graph has each attribute as a node, the association relationship between the attributes as an edge, and the weight of the edge represents the strength of the association between the attributes; Calculating attribute feature values ​​based on the attribute graph; Determining an iterative product among candidate products based on the attribute feature value; The calculating the attribute feature value based on the attribute graph includes: Sort the nodes by importance based on the attribute graph to obtain an important sequence; Obtain the attention weight of each node based on the important sequence, and calculate the eigenvalue of each node in the important sequence; An attribute feature value is calculated based on the attention weight and the feature value.

Images