E-commerce transaction system and method based on virtual reality

Abstract

The invention provides an e-commerce transaction system and method based on virtual reality. The method comprises the following steps: extracting an explicit interaction feature and an implicit interaction feature of each user in a virtual reality shopping scene; performing interaction node association on the explicit interaction feature and the implicit interaction feature of each user to obtain interaction attribute information of each user; extracting browsing interest points of other users having the same customer attributes as the target user during virtual reality shopping, and predicting behavior preferences of the target user according to all the browsing interest points to obtain a preference prediction value; determining the difference degree of the interaction behaviors between the target user and other users through all the interaction attribute information and the preference prediction values; and when the target user uses the virtual reality equipment to carry out virtual reality shopping, the virtual reality equipment predicts and displays the recommended commodity according to the difference between the current shopping intention of the target user and the interaction behavior. By adopting the scheme of the invention, personalized recommendation of commodities in the e-commerce transaction system can be realized.

Description

Technical Field

[0001] The present application relates to the field of virtual reality technology, and more specifically, to an e-commerce transaction system and method based on virtual reality. Background Art

[0002] The application of virtual reality in e-commerce transactions is gradually becoming an emerging trend. With the rapid development of Internet technology and virtual reality equipment, traditional online shopping methods are gradually shifting to immersive experiences. Virtual reality provides a new interactive model for e-commerce platforms. In traditional e-commerce, consumers can only understand products through text, pictures, and videos, lacking real perception and interactive experience of products. Virtual reality technology creates an immersive virtual shopping environment, allowing consumers to experience products "personally" in a virtual space. Whether it is clothing, furniture or electronic products, they can use virtual models to perform 3D viewing, try-on, and trial operations, thus breaking through the limitations of traditional e-commerce shopping.

[0003] However, in the existing technology, traditional recommendation systems rely on static data sets and fixed algorithm strategies, resulting in recommendation results that often cannot accurately reflect users' current interests and needs. In addition, the recommendation strategies of traditional e-commerce platforms lack a comprehensive understanding of users' interactive behaviors in virtual reality shopping scenarios, and mainly rely on basic display behavior data (such as click volume, purchase history, etc.). These data cannot deeply explore users' potential needs, resulting in a low degree of personalization in recommendation results. By real-time association and analysis of users' explicit and implicit behavior characteristics in a virtual reality environment, not only can the accuracy of the recommendation system be improved, but the recommendation strategy can also be adjusted according to users' dynamic shopping intentions, so as to more flexibly adapt to changes in user needs, improve shopping experience and increase purchase conversion rate. Therefore, how to achieve personalized product recommendations in e-commerce transaction systems has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides an e-commerce transaction system and method based on virtual reality, which can realize personalized recommendation of commodities in the e-commerce transaction system.

[0005] In a first aspect, the present application provides a commodity recommendation method based on virtual reality, which is used to recommend commodities in an e-commerce transaction system based on virtual reality, and comprises the following steps:

[0006] Collect all users' interactive behavior data in the virtual reality mall from the virtual reality devices, and then extract the explicit and implicit interactive features of each user in the virtual reality shopping scene;

[0007] Associating the explicit interaction features and implicit interaction features of each user with interaction nodes, thereby determining the interaction attribute information of each user in the virtual reality shopping scene;

[0008] Detect the customer attributes of the target user, extract the browsing interest points of other users with the same customer attributes as the target user when they perform virtual reality shopping in the virtual reality device, and then make a confidence prediction of the shopping behavior preference of the target user based on all the browsing interest points to obtain the preference prediction value;

[0009] Determine the difference between the target user and other users' interaction behaviors in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value;

[0010] When the target user uses the virtual reality device to perform virtual reality shopping, the virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors.

[0011] Preferably, extracting explicit interaction features and implicit interaction features of each user in a virtual reality shopping scene specifically includes:

[0012] For each user, obtain the user's interactive behavior data in the virtual reality mall;

[0013] Filter out the user's explicit behaviors in each shopping session from the interactive behavior data, and determine the user's explicit interactive features in the virtual reality shopping scenario through all explicit behaviors;

[0014] The implicit behaviors of users in each shopping are screened out from the interactive behavior data, and the implicit interaction characteristics of users in the virtual reality shopping scene are determined through all implicit behaviors, thereby obtaining the explicit interaction characteristics and implicit interaction characteristics of each user in the virtual reality shopping scene.

[0015] Preferably, associating the explicit interaction features and implicit interaction features of each user with interaction nodes to further determine the interaction attribute information of each user in the virtual reality shopping scene specifically includes:

[0016] For each user, extract all the interaction nodes of the user in the virtual reality shopping scene;

[0017] According to the explicit interaction features and implicit interaction features of the user, the interaction behavior of each interaction node is associated and perceived to obtain an associated attribute label of each interaction node;

[0018] The interactive attribute information of the user in the virtual reality shopping scene is determined through all associated attribute tags, and then the interactive attribute information of each user in the virtual reality shopping scene is obtained.

[0019] Preferably, extracting browsing interest points of other users having the same customer attributes as the target user when performing virtual reality shopping in a virtual reality device specifically includes:

[0020] Other users with the same customer attributes as the target user are designated as the users;

[0021] For each designated user, the interaction behavior data of the designated user in the virtual reality mall is obtained, and then the gaze duration and gaze area of ​​the designated user in the process of browsing products are extracted;

[0022] The browsing interest points of the designated user when performing virtual reality shopping in the virtual reality device are determined by the gaze duration and the gaze area, and then the browsing interest points of each designated user when performing virtual reality shopping in the virtual reality device are obtained.

[0023] Preferably, the shopping behavior preference of the target user is predicted based on all browsing interest points, and the preference prediction value is obtained specifically including:

[0024] Perform confidence association on all browsing interest points to obtain the association factors between browsing interest points;

[0025] Filtering out valid points of interest from all browsing points of interest according to the association factor;

[0026] The shopping behavior preference of the target user is predicted through all valid points of interest to obtain the preference prediction value.

[0027] Preferably, the virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors, specifically including:

[0028] Initialize a recommendation model;

[0029] Adjusting a recommendation strategy for product recommendations according to the differences in the interactive behaviors;

[0030] Updating the recommendation strategy built into the recommendation model through the adjusted recommendation strategy;

[0031] The target user’s current shopping intention is used as the input parameter of the recommendation model, and then the recommendation model is combined with the target user’s interactive behavior data to predict and generate a recommendation list for the target user;

[0032] The commodities in the recommendation list are displayed in sequence through a virtual reality device.

[0033] Preferably, the virtual reality device is a head-mounted display device.

[0034] In a second aspect, the present application provides an e-commerce transaction system based on virtual reality, wherein the e-commerce transaction system based on virtual reality includes a commodity recommendation unit, and the commodity recommendation unit includes:

[0035] The collection module is used to collect the interactive behavior data of all users in the virtual reality mall from the virtual reality device, and then extract the explicit interaction features and implicit interaction features of each user in the virtual reality shopping scene;

[0036] A processing module, used to associate the explicit interaction features and implicit interaction features of each user with interaction nodes, and then determine the interaction attribute information of each user in the virtual reality shopping scene;

[0037] The processing module is further used to detect the customer attributes of the target user, extract the browsing interest points of other users with the same customer attributes as the target user when performing virtual reality shopping in the virtual reality device, and then make a confidence prediction of the shopping behavior preference of the target user based on all the browsing interest points to obtain a preference prediction value;

[0038] The processing module is further used to determine the difference between the interaction behaviors of the target user and other users in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value;

[0039] The execution module is used for predicting and displaying recommended commodities according to the current shopping intention of the target user and the difference degree of the interactive behavior when the target user uses the virtual reality device for virtual reality shopping.

[0040] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned virtual reality-based product recommendation method.

[0041] In a fourth aspect, 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 above-mentioned virtual reality-based product recommendation method is implemented.

[0042] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0043] In an embodiment of the present application, the interactive behavior data of all users in a virtual reality mall are collected from a virtual reality device, and then the explicit interactive features and implicit interactive features of each user in the virtual reality shopping scene are extracted; the explicit interactive features and implicit interactive features of each user are associated with interactive nodes, and then the interactive attribute information of each user in the virtual reality shopping scene is determined; the customer attributes of the target user are detected, and the browsing interest points of other users with the same customer attributes as the target user when performing virtual reality shopping in the virtual reality device are extracted, and then the shopping behavior preference of the target user is confidently predicted based on all the browsing interest points to obtain a preference prediction value; the difference in interactive behavior between the target user and other users in the virtual reality shopping scene is determined through all the interactive attribute information and the preference prediction value; when the target user uses the virtual reality device to perform virtual reality shopping, the virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference in the interactive behavior.

[0044] It can be seen that the present application determines the difference in interactive behavior between the target user and other users in the virtual reality shopping scene through interactive attribute information and preference prediction values, and then predicts the recommended products for the target user based on the difference in interactive behavior; firstly, by associating the user's explicit interactive features and implicit interactive features with interactive nodes, the user's behavior pattern and preference can be accurately described, and then the user's interactive attribute information in the virtual reality shopping scene can be determined, wherein the extraction and association of explicit and implicit interactive features help the system to fully understand the user's shopping habits and interest directions; secondly, the shopping behavior preferences of the target user are confidently predicted based on the browsing interest points of all users, thereby obtaining a preference prediction value, which can reflect the user's potential interest in a specific product or product category, and can be used to predict the target user's shopping behavior based on the browsing interest points of all users. When a user clearly expresses interest, the system can predict the products that may interest him / her. The preference prediction value can not only improve the foresight of product recommendations, but also adjust the recommendation strategy in real time to cope with the dynamic changes in user needs. Then, by combining the interaction attribute information and the preference prediction value, the system can analyze the differences in interaction behaviors between the target user and other users in the virtual reality shopping scene, thereby identifying user groups with similar shopping needs as the target user, which makes the recommendation strategy more flexible and can more accurately match user needs. Finally, based on the analysis of the differences in interaction behaviors, personalized recommendations can accurately push products that meet user needs, significantly improve the recommendation effect, enhance the user shopping experience, and promote the growth of transaction volume on the e-commerce platform. In summary, the present application scheme can realize personalized recommendations for products in e-commerce transaction systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is an exemplary flow chart of a commodity recommendation method based on virtual reality according to some embodiments of the present application;

[0046] Figure 2 is a schematic diagram of the process of e-commerce transactions according to some embodiments of the present application;

[0047] Figure 3 is a schematic diagram of a process for determining a preference prediction value according to some embodiments of the present application;

[0048] Figure 4 is a structural schematic diagram of a commodity recommendation unit according to some embodiments of the present application;

[0049] Figure 5 It is a structural diagram of a computer device for implementing a commodity recommendation method based on virtual reality according to some embodiments of the present application. DETAILED DESCRIPTION

[0050] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0051] refer to Figure 1 , which is an exemplary flow chart of a commodity recommendation method based on virtual reality according to some embodiments of the present application. The commodity recommendation method 100 based on virtual reality mainly includes the following steps:

[0052] In step 101, the interaction behavior data of all users in the virtual reality mall are collected from the virtual reality devices, and then the explicit interaction features and implicit interaction features of each user in the virtual reality shopping scene are extracted.

[0053] It should be noted that the reference Figure 2 As shown, the figure is a schematic diagram of the process of e-commerce transactions in some embodiments of the present application. First, the dealer submits an application for dealer qualification to the e-commerce transaction platform. The e-commerce transaction platform is responsible for reviewing and authorizing the store. Secondly, the operating customer and the operating store place an order through the virtual reality device or the client, and the order information is transmitted to the store. After the store processes the order, the e-commerce transaction platform settles the bill and delivers the goods to the customer. The customer signs for the goods after receiving them, completing the entire transaction process. The figure shows the interactive relationship between each link, including the qualification application and authorization of the dealer and the e-commerce transaction platform, the operating customer and the store place an order through the virtual reality device (ie, the client), and the settlement and goods delivery between the e-commerce transaction platform and the store.

[0054] It should also be noted that the interactive behavior data in this application refers to the human-computer interaction data generated by users when using virtual reality devices for e-commerce transactions; it should also be noted that the virtual reality device in this application is a head-mounted display device, and in other embodiments, the virtual reality device can also be other devices, which is not specifically limited here.

[0055] In specific implementation, the collection of interactive behavior data of all users in the virtual reality mall from the virtual reality device can be achieved in the following manner, namely: first, the user's operation behavior (such as clicks, gestures, and eye tracking) in the virtual reality mall is perceived in real time by the virtual reality device to obtain the user's interactive behavior data, and then, the collected interactive behavior data is transmitted to the cloud using a high-speed network protocol (such as WebSocket), and a distributed storage system (such as HDFS or a cloud database) is used to store and manage large-scale interactive behavior data. Finally, the interactive behavior data of all users in the virtual reality mall within a specified historical time period can be obtained from the distributed storage system; it should be noted that the specified historical time period in this application is the time period corresponding to the current moment to the past six months. In other embodiments, the specified historical time period can also be other time periods, which are not specifically limited here.

[0056] In some embodiments, extracting explicit interaction features and implicit interaction features of each user in a virtual reality shopping scene may be achieved by using the following steps:

[0057] For each user, obtain the user's interactive behavior data in the virtual reality mall;

[0058] Filter out the user's explicit behaviors in each shopping session from the interactive behavior data, and determine the user's explicit interactive features in the virtual reality shopping scenario through all explicit behaviors;

[0059] The implicit behaviors of users in each shopping are screened out from the interactive behavior data, and the implicit interaction characteristics of users in the virtual reality shopping scene are determined through all implicit behaviors, thereby obtaining the explicit interaction characteristics and implicit interaction characteristics of each user in the virtual reality shopping scene.

[0060] It should be noted that the explicit behavior in this application refers to the behavior of users clearly expressing their intentions and needs in the virtual reality device, which is usually an operation actively performed by the user that can be directly observed and quantified, such as clicking on a product or purchasing it; the implicit behavior in this application refers to the operation behavior of users who do not directly express their intentions in the virtual reality device, but indirectly reflect their interests and needs through their behavior trajectories; in addition, the explicit interaction features in this application

[0061] In specific implementation, the explicit behavior of the user in each shopping trip is screened out from the interactive behavior data, and the explicit interactive features of the user in the virtual reality shopping scene are determined through all the explicit behaviors. This can be achieved in the following manner: feature data that can reflect the user's behavior of clicking on and purchasing goods in each shopping trip is screened out from the interactive behavior data, and the click frequency and purchase frequency of the user in the shopping process are extracted from the feature data, and then all the click frequencies and purchase frequencies are converted into feature vectors in the order of recording time, and the feature vector is used as the explicit interactive feature of the user in the virtual reality shopping scene; the implicit behavior of the user in each shopping trip is screened out from the interactive behavior data, and the implicit interactive features of the user in the virtual reality shopping scene are determined through all the implicit behaviors. This can be achieved in the following manner: feature data that can reflect the user's browsing behavior in each shopping trip is screened out from the interactive behavior data, and the browsing time and visual field density of the user in the shopping process are extracted from the feature data, and then all the browsing time and visual field density are converted into feature vectors in the order of recording time, and the feature vector is used as the implicit interactive feature of the user in the virtual reality shopping scene.

[0062] In step 102, the explicit interaction features and implicit interaction features of each user are associated with interaction nodes, thereby determining the interaction attribute information of each user in the virtual reality shopping scene.

[0063] In some embodiments, the following steps may be used to associate the explicit interaction features and implicit interaction features of each user with interaction nodes, and then determine the interaction attribute information of each user in the virtual reality shopping scene:

[0064] For each user, extract all the interaction nodes of the user in the virtual reality shopping scene;

[0065] According to the explicit interaction features and implicit interaction features of the user, the interaction behavior of each interaction node is associated and perceived to obtain an associated attribute label of each interaction node;

[0066] The interactive attribute information of the user in the virtual reality shopping scene is determined through all associated attribute tags, and then the interactive attribute information of each user in the virtual reality shopping scene is obtained.

[0067] It should be noted that the interaction node in this application refers to the key node corresponding to each interaction behavior of the user in the virtual reality scene, and each interaction node carries the time, location and related attribute information of the behavior; the associated attribute label in this application refers to the attribute identifier assigned to each interaction node in the virtual reality scene; in addition, the interaction attribute information in this application refers to the characteristic information that can reflect the user's interaction behavior in the virtual reality scene, such as the user's behavior pattern and operation habits in a specific scene.

[0068] In specific implementation, extracting all the interaction nodes of the user in the virtual reality shopping scene can be achieved in the following manner, namely, obtaining the time node of each operation from the user's operation record, and defining the time node as the interaction node, so as to obtain all the interaction nodes of the user in the virtual reality shopping scene; performing association perception on the interaction behavior of each interaction node according to the user's explicit interaction features and implicit interaction features, and obtaining the association attribute label of each interaction node can be achieved in the following manner, namely, using the existing association rule mining algorithm (such as Apriori algorithm) to map the explicit interaction features and implicit interaction features to each specific interaction node, and using the association rule mining algorithm to obtain the association attribute label of each interaction node; The method learns the association information of explicit interaction features and implicit interaction features at the interaction nodes, and then outputs the association value of each interaction node through the association rule mining algorithm, and uses the association value of each interaction node as the association attribute label of the corresponding interaction node; determining the interaction attribute information of the user in the virtual reality shopping scene through all the association attribute labels can be achieved in the following way, that is, all the association attribute labels can be grouped according to the category of the interaction attribute, and the association attribute labels of each category group are counted, and then the statistical value of each category group is used as the attribute value of the corresponding interaction attribute category, and then all the attribute values ​​are combined into a set as the interaction attribute information of the user in the virtual reality shopping scene.

[0069] In step 103, the customer attributes of the target user are detected, and the browsing interest points of other users with the same customer attributes as the target user when performing virtual reality shopping in the virtual reality device are extracted, and then the shopping behavior preference of the target user is confidently predicted based on all the browsing interest points to obtain a preference prediction value.

[0070] It should be noted that customer attributes refer to information that describes individual characteristics of customers. In this application, customer attributes are basic attribute information of personal users. Basic attribute information includes age, gender, and geographic location. In other embodiments, customer attributes may also include other information of the user, which is not specifically limited here.

[0071] In some embodiments, extracting browsing points of interest of other users who have the same customer attributes as the target user when performing virtual reality shopping in a virtual reality device may be achieved by using the following steps:

[0072] Other users with the same customer attributes as the target user are designated as the users;

[0073] For each designated user, the interaction behavior data of the designated user in the virtual reality mall is obtained, and then the gaze duration and gaze area of ​​the designated user in the process of browsing products are extracted;

[0074] The browsing interest points of the designated user when performing virtual reality shopping in the virtual reality device are determined by the gaze duration and the gaze area, and then the browsing interest points of each designated user when performing virtual reality shopping in the virtual reality device are obtained.

[0075] It should be noted that the gaze duration in this application refers to the duration of time that a user's gaze is focused on a specific target (such as a product) in a virtual reality environment; the gaze area refers to the specific spatial area where the user's gaze stays in the virtual reality environment; and the browsing points of interest refer to the information points to which the user shows a high degree of attention based on his or her interactive behavior in the virtual reality shopping scene. Browsing points of interest can reflect the user's immediate needs and potential preferences.

[0076] In specific implementation, other users with the same customer attributes as the target user may be designated as designated users in the following manner, namely: screen out a user group with the same customer attributes as the target user, and designate all users in the user group as designated users; for each designated user, obtain the interaction behavior data of the designated user in the virtual reality mall, and then extract the gaze duration and gaze area of ​​the designated user in the process of browsing goods, namely: obtain the interaction behavior data of the designated user in the virtual reality mall from the distributed storage system, and then extract the gaze duration and gaze area of ​​the designated user from the obtained interaction behavior data, wherein the gaze duration can be obtained by analyzing the length of time the designated user's gaze stays on a specific product or interface element, and the gaze area uses the eye tracker in the virtual reality device to capture the user's gaze trajectory in real time. The eye gaze point coordinates are obtained, and the eye gaze point coordinates are mapped with the display content (such as goods, interface elements, etc.) in the virtual reality environment to obtain the user's visual focus, and then the virtual reality environment is divided into a number of virtual areas (such as a product display area, a navigation button, an advertising column), and the user's visual focus is matched with these areas according to the position of the visual focus to obtain the gaze area; the browsing interest points of the designated user when performing virtual reality shopping in the virtual reality device are determined by the gaze duration and the gaze area, which can be achieved in the following manner, namely: combining the gaze duration and the gaze area, using a statistical analysis method (such as a heat map analysis) to identify the browsing interest points of the designated user, these interest points can represent goods, areas or information that the user pays more attention to, and then the browsing interest points of each designated user when performing virtual reality shopping in the virtual reality device are obtained.

[0077] In some embodiments, reference Figure 3 As shown, this figure is a schematic diagram of the process of determining the preference prediction value in some embodiments of the present application. In this embodiment, the shopping behavior preference of the target user is predicted based on all browsing interest points. The preference prediction value can be obtained by the following steps:

[0078] In step 1031, confidence association is performed on all browsing interest points to obtain association factors between browsing interest points;

[0079] In step 1032, valid points of interest are screened out from all browsed points of interest according to the association factor;

[0080] In step 1033, the shopping behavior preference of the target user is predicted through all valid points of interest to obtain a preference prediction value.

[0081] It should be noted that the correlation factor in this application is an indicator for measuring the degree of correlation between browsing interest points; the effective interest points in this application refer to information points that can truly reflect user behavior preferences; the preference prediction value in this application is an indicator for measuring the user's potential preference for predicted products or content.

[0082] In specific implementation, all browsing interest points are analyzed for association, and the association factor between browsing interest points can be obtained in the following manner, namely: a designated user is selected as the selected user, and the existing association rule algorithm (such as Apriori algorithm) can be used to perform association analysis on the browsing interest points between the selected user and all other designated users, and then an association value that can describe the browsing interest points between the selected user and all other designated users is output through the association rule algorithm, and the association value is used as the association factor between the selected user and the corresponding browsing interest points of all other designated users, and the above steps are repeated to obtain the association factor between the remaining designated users and the corresponding browsing interest points of all other designated users; the effective interest points can be screened out from all browsing interest points according to the association factor in the following manner, namely: a correlation threshold is pre-set, and the correlation threshold can reflect the correlation relationship between browsing interest points. average level, and then compare each correlation factor with the correlation threshold; when the correlation factor is greater than or equal to the correlation threshold, the browsing interest point of the selected user corresponding to the correlation factor is determined to be a valid browsing interest point, and the valid browsing interest point is used as a valid interest point; when the correlation factor is less than the correlation threshold, the browsing interest point of the selected user corresponding to the correlation factor is determined to be an invalid browsing interest point; all valid interest points can be obtained in the above manner; the shopping behavior preference of the target user is predicted through all valid interest points, and the preference prediction value can be obtained in the following manner, namely: initialize a preference prediction model based on a neural network, input all valid interest points into the preference prediction model, use the preference prediction value as the output label in the preference prediction model, use the preference prediction model to predict the shopping behavior preference of the target user, so that the prediction result of the preference prediction model can be used as the preference prediction value.

[0083] It should be noted that the training steps of the preference prediction model include the following stages: Data collection and preprocessing: Collect user behavior data (such as clicks, browsing, and purchases), clean and format the data, remove noise data, and standardize it, such as filling missing values ​​or eliminating invalid behaviors; Feature extraction: Extract the user's browsing interests from the user's interactive behavior data as the input features of the model; Model selection and initialization: Select a suitable preference prediction model, such as collaborative filtering, matrix decomposition, neural network, etc., initialize the model's structure and parameters, and if a deep learning model is used, initialize the weights and bias of the neural network; Model training and optimization: Use training data to train the model, and adjust the model parameters through optimization algorithms (such as gradient descent) to make the predicted value as close to the user's actual preference as possible. During the training process, cross-validation can be used to prevent overfitting; Model evaluation and adjustment: Evaluate the performance of the model on the validation set, and use evaluation indicators such as precision, recall, and F1 score to measure the accuracy of the prediction. If the model performs poorly, it may be necessary to adjust the model architecture, learning rate, or other hyperparameters, and retrain and optimize.

[0084] In step 104, the difference in the interaction behavior between the target user and other users in the virtual reality shopping scene is determined through all the interaction attribute information and the preference prediction value.

[0085] In some embodiments, determining the difference between the interaction behaviors of the target user and other users in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value can be achieved by using the following steps:

[0086] Screening out marked users who have the same preference behavior as the target user according to the preference prediction value;

[0087] Filter out the interactive attribute information of all marked users;

[0088] The interaction attribute information of the target user is differentiated from the interaction attribute information of all marked users to obtain the difference degree of interaction behavior between the target user and other users in the virtual reality shopping scene.

[0089] In specific implementation, screening out marked users having the same preference behavior as the target user according to the preference prediction value can be achieved in the following manner, namely, the preference prediction value of other users can be compared with that of the target user using the existing cosine similarity, all users having similar preference behaviors can be selected, and the screened out users can be used as marked users, wherein the marked users refer to the user groups with highly similar preference prediction values ​​to the target user; screening out the interaction attribute information of all marked users can be achieved in the following manner, namely, the interaction attribute information of all marked users can be screened out from all the interaction attribute information; performing differential analysis on the interaction attribute information of the target user and the interaction attribute information of all marked users, and obtaining the difference degree of interaction behavior between the target user and other users in the virtual reality shopping scene can be achieved in the following manner, namely, the average Euclidean distance of the interaction attribute information between the target user and each marked user can be used as the difference degree of interaction behavior between the target user and other users in the virtual reality shopping scene.

[0090] It should be noted that the difference in interactive behavior in the present application refers to the difference in interactive attributes between the target user and other users in the virtual reality shopping scenario.

[0091] In step 105, when the target user uses the virtual reality device to perform virtual reality shopping, the virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors.

[0092] In some embodiments, the virtual reality device can predict and display recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors by using the following steps:

[0093] Initialize a recommendation model;

[0094] Adjusting a recommendation strategy for product recommendations according to the differences in the interactive behaviors;

[0095] Updating the recommendation strategy built into the recommendation model through the adjusted recommendation strategy;

[0096] The target user’s current shopping intention is used as the input parameter of the recommendation model, and then the recommendation model is combined with the target user’s interactive behavior data to predict and generate a recommendation list for the target user;

[0097] The commodities in the recommendation list are displayed in sequence through a virtual reality device.

[0098] It should be noted that the recommendation model in this application refers to a machine learning model for personalized product recommendations, and the recommendation model is a recommendation model based on a neural network.

[0099] In specific implementation, initializing a recommendation model can be achieved in the following manner, namely: a recommendation model can be obtained based on neural network training, and the recommendation model is initialized; adjusting the recommendation strategy for product recommendation according to the difference in the interactive behavior can be achieved in the following manner, namely: the existing reinforcement learning model can be used to automatically adjust the recommendation strategy for product recommendation. If the difference in the interactive behavior is less than the first threshold, the recommendation strategy can tend to recommend products similar to the marked user; if the difference is greater than the second threshold, the recommendation strategy needs to focus on the unique interests of the target user to improve the user's personalized needs; updating the recommendation strategy built into the recommendation model through the adjusted recommendation strategy can be achieved in the following manner, namely: replacing and updating the recommendation strategy built into the recommendation model with the adjusted recommendation strategy. During the updating process, the online learning mechanism can be used to update in real time. Recommendation model; taking the current shopping intention of the target user as the input parameter of the recommendation model, and then generating the recommendation list of the target user through the recommendation model combined with the interactive behavior data of the target user for prediction can be achieved in the following manner, namely: passing the current shopping intention of the target user as the input parameter into the recommendation model, the current shopping intention can be inferred through the specific behavior of the user in the virtual reality scene (such as searching for keywords, etc.), the recommendation model will combine the shopping intention information and the user's historical interactive behavior to predict the user's purchasing needs, and use the output list predicted by the recommendation model as the recommendation list of the target user; displaying the products in the recommendation list in sequence through the virtual reality device can be achieved in the following manner, namely: the virtual reality device can use an interactive interface to display the recommended products, such as virtual shelves, 3D product displays, etc., to ensure that the user can clearly see the product information and details.

[0100] On the other hand, in some embodiments, the present application provides an e-commerce transaction system based on virtual reality, the e-commerce transaction system based on virtual reality includes a commodity recommendation unit, referring to Figure 4 , which is a schematic diagram of the structure of a product recommendation unit according to some embodiments of the present application, the product recommendation unit 400 includes: a collection module 401, a processing module 402 and an execution module 403, which are described as follows:

[0101] Collection module 401, in this application, collection module 401 is mainly used to collect the interaction behavior data of all users in the virtual reality mall from the virtual reality device, and then extract the explicit interaction features and implicit interaction features of each user in the virtual reality shopping scene;

[0102] Processing module 402, in the present application, is used to associate the explicit interaction features and implicit interaction features of each user with interaction nodes, thereby determining the interaction attribute information of each user in the virtual reality shopping scene;

[0103] The processing module 402 in the present application is also used to detect the customer attributes of the target user, extract the browsing interest points of other users with the same customer attributes as the target user when performing virtual reality shopping in the virtual reality device, and then make a confidence prediction of the shopping behavior preference of the target user based on all the browsing interest points to obtain a preference prediction value;

[0104] The processing module 402 in the present application is also used to determine the difference between the target user and other users in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value;

[0105] Execution module 403, in the present application, execution module 403 is mainly used for predicting and displaying recommended products based on the target user's current shopping intention and the difference between the interactive behaviors when the target user uses the virtual reality device for virtual reality shopping.

[0106] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned virtual reality-based product recommendation method.

[0107] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing a commodity recommendation method based on virtual reality according to some embodiments of the present application. The commodity recommendation method based on virtual reality in the above embodiment can be performed by Figure 5 The computer device 500 shown in the figure is implemented, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503 and at least one communication interface 504.

[0108] The processor 501 may be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0109] The communication bus 502 may be used to transmit information between the above-mentioned components.

[0110] The memory 503 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 503 may exist independently and be connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0111] The memory 503 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The commodity recommendation method based on virtual reality in the above embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0112] The communication interface 504 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0113] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0114] The above-mentioned computer device can be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0115] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned commodity recommendation method based on virtual reality is implemented.

[0116] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0117] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A commodity recommendation method based on virtual reality, used for recommending commodities in an e-commerce transaction system based on virtual reality, characterized in that: The steps include: Collect all users' interactive behavior data in the virtual reality mall from the virtual reality devices, and then extract the explicit and implicit interactive features of each user in the virtual reality shopping scene; Associating the explicit interaction features and implicit interaction features of each user with interaction nodes, thereby determining the interaction attribute information of each user in the virtual reality shopping scene; Detect the customer attributes of the target user, extract the browsing interest points of other users with the same customer attributes as the target user when they perform virtual reality shopping in the virtual reality device, and then make a confidence prediction of the shopping behavior preference of the target user based on all the browsing interest points to obtain the preference prediction value; Determine the difference between the target user and other users' interaction behaviors in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value; When the target user uses the virtual reality device to perform virtual reality shopping, the virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors.

2. The method according to claim 1, characterized in that The explicit and implicit interaction features of each user in the virtual reality shopping scene are extracted, including: For each user, obtain the user's interactive behavior data in the virtual reality mall; Filter out the user's explicit behaviors in each shopping session from the interactive behavior data, and determine the user's explicit interactive features in the virtual reality shopping scenario through all explicit behaviors; The implicit behaviors of users in each shopping are screened out from the interactive behavior data, and the implicit interaction characteristics of users in the virtual reality shopping scene are determined through all implicit behaviors, thereby obtaining the explicit interaction characteristics and implicit interaction characteristics of each user in the virtual reality shopping scene.

3. The method according to claim 1, characterized in that The explicit interaction features and implicit interaction features of each user are associated with interaction nodes to thereby determine the interaction attribute information of each user in the virtual reality shopping scene, specifically including: For each user, extract all the interaction nodes of the user in the virtual reality shopping scene; According to the explicit interaction features and implicit interaction features of the user, the interaction behavior of each interaction node is associated and perceived to obtain an associated attribute label of each interaction node; The interactive attribute information of the user in the virtual reality shopping scene is determined through all associated attribute tags, and then the interactive attribute information of each user in the virtual reality shopping scene is obtained.

4. The method according to claim 1, characterized in that Extracting browsing points of interest of other users who have the same customer attributes as the target user when they perform virtual reality shopping in a virtual reality device specifically includes: Other users with the same customer attributes as the target user are designated as the users; For each designated user, the interaction behavior data of the designated user in the virtual reality mall is obtained, and then the gaze duration and gaze area of ​​the designated user in the process of browsing products are extracted; The browsing interest points of the designated user when performing virtual reality shopping in the virtual reality device are determined by the gaze duration and the gaze area, and then the browsing interest points of each designated user when performing virtual reality shopping in the virtual reality device are obtained.

5. The method according to claim 1, characterized in that Based on all browsing interests, the shopping behavior preferences of the target user are predicted with confidence, and the preference prediction values ​​obtained include: Perform confidence association on all browsing interest points to obtain the association factors between browsing interest points; Filtering out valid points of interest from all browsing points of interest according to the association factor; The shopping behavior preference of the target user is predicted through all valid points of interest to obtain the preference prediction value.

6. The method according to claim 1, characterized in that The virtual reality device predicts and displays recommended products based on the current shopping intention of the target user and the difference between the interactive behaviors, specifically including: Initialize a recommendation model; Adjusting a recommendation strategy for product recommendations according to the differences in the interactive behaviors; Updating the recommendation strategy built into the recommendation model through the adjusted recommendation strategy; The target user’s current shopping intention is used as the input parameter of the recommendation model, and then the recommendation model is combined with the target user’s interactive behavior data to predict and generate a recommendation list for the target user; The commodities in the recommendation list are displayed in sequence through a virtual reality device.

7. The method according to claim 1, characterized in that The virtual reality device is a head-mounted display device.

8. An e-commerce transaction system based on virtual reality, the e-commerce transaction system based on virtual reality includes a commodity recommendation unit, characterized in that: The product recommendation unit includes: The collection module is used to collect the interactive behavior data of all users in the virtual reality mall from the virtual reality device, and then extract the explicit interaction features and implicit interaction features of each user in the virtual reality shopping scene; A processing module, used to associate the explicit interaction features and implicit interaction features of each user with interaction nodes, and then determine the interaction attribute information of each user in the virtual reality shopping scene; The processing module is further used to detect the customer attributes of the target user, extract the browsing interest points of other users with the same customer attributes as the target user when performing virtual reality shopping in the virtual reality device, and then make a confidence prediction of the shopping behavior preference of the target user based on all the browsing interest points to obtain a preference prediction value; The processing module is further used to determine the difference between the interaction behaviors of the target user and other users in the virtual reality shopping scene through all the interaction attribute information and the preference prediction value; The execution module is used for predicting and displaying recommended commodities according to the current shopping intention of the target user and the difference degree of the interactive behavior when the target user uses the virtual reality device for virtual reality shopping.

9. A computer device, comprising a memory and a processor, wherein the memory stores a code, characterized in that: The processor is configured to obtain the code and execute the virtual reality-based commodity recommendation method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the commodity recommendation method based on virtual reality as described in any one of claims 1 to 7 is implemented.

Images