User shopping recommendation method and system based on scene recognition, medium and equipment

Abstract

The invention belongs to the technical field of self-service shopping, and provides a user shopping recommendation method and system based on scene recognition, a medium and equipment, and the method comprises the steps: building an environment scene library of a vending machine based on the historical shopping record of a user; based on a scene recognition assembly of the vending machine, obtaining a current environment scene where the vending machine is located according to the environment scene library; and according to the current environment scene, using the vending machine to make shopping recommendation to the user. The environment scene library of the vending machine is constructed according to the historical shopping records of the user, the current environment scene where the vending machine is located is obtained according to the environment scene library, and the vending machine is utilized to make shopping recommendation for the user, so that the efficiency and quality of making shopping recommendation for the user by the vending machine can be improved, the shopping experience of the user is improved, and the user experience is improved. And the intelligent level of shopping recommendation is improved.

Description

Technical Field

[0001] The present invention relates to the field of self-service shopping technology, and in particular to a method, system, medium and device for recommending user shopping based on scene recognition. Background Art

[0002] A vending machine is a machine that can automatically supply goods according to the coins inserted. Vending machines are commonly used equipment in commercial automation. They are not restricted by time and place, can save manpower and facilitate transactions.

[0003] Vending machines are set up in crowded places such as shopping malls, stations, and hospitals. The environments of these places are different, and the items sold by vending machines are also different, so it is necessary to recommend items based on different environmental scenarios. At present, it is not enough for vending machines to make targeted recommendations to users based on different layout environments and scenarios, which affects the users' shopping experience and also affects the sales of vending machines.

[0004] Therefore, it is necessary to provide a user shopping recommendation method, system, medium and device based on scene recognition. Summary of the invention

[0005] The present invention provides a user shopping recommendation method, system, medium and device based on scene recognition. According to the user's shopping history records, an environmental scene library of a vending machine is constructed. According to the environmental scene library, the current environmental scene of the vending machine is obtained, and the vending machine is used to make shopping recommendations to the user. This can improve the efficiency and quality of the vending machine's shopping recommendations to the user, which is beneficial to improving the user's shopping experience and improving the intelligence level of shopping recommendations.

[0006] The present invention provides a user shopping recommendation method based on scene recognition, comprising: Build a vending machine environment scene library based on user shopping history; Based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library; Based on the current environment, vending machines are used to make shopping recommendations to users.

[0007] Furthermore, based on the user's shopping history, an environmental scene library for vending machines is constructed, including: Get the user's shopping history; Based on historical records, a dataset of environmental scenarios used by users when shopping is established; According to the environmental scene data set, the environmental scene classification model is used to classify the environmental scenes, and an environmental scene library is constructed based on the classified environmental scenes.

[0008] Furthermore, based on historical records, a dataset of environmental scenarios when users shop is established, including: Obtain user shopping behavior information from historical records, and generate a behavior information attribute data set based on the user shopping behavior information; Obtain the environmental feature data of the vending machine corresponding to the time when the user's shopping behavior occurs, and generate an environmental feature data set; Based on the behavior information attribute dataset and the environmental feature dataset, an environmental scene dataset is generated.

[0009] Furthermore, the environmental feature data of the vending machine corresponding to the time when the user's shopping behavior occurs is obtained to generate an environmental feature data set, including: Based on the environmental sensor in the scene recognition component of the vending machine, the environmental temperature and humidity data are collected; Based on the camera device in the scene recognition component of the vending machine, the user's clothing data and user traffic data are obtained; Acquire user voice data based on the high-sensitivity microphone in the scene recognition component of the vending machine; Aggregate environmental temperature and humidity data, user clothing data, user traffic data, and user voice data to generate an environmental feature data set.

[0010] Furthermore, based on the behavior information attribute dataset and the environment feature dataset, an environment scene dataset is generated, including: Using the behavior information attribute dataset and the environmental feature dataset, several user shopping environment scenarios are generated based on the scenario generation model; among which, the scenario generation model is generated based on the training of the deep neural network model.

[0011] Furthermore, based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library, including: Obtain the user's current shopping behavior information and generate current behavior information attribute data; Based on the scene recognition component of the vending machine, identify and obtain the current environmental field feature data of the vending machine; According to the current behavior information attribute data and the current environment feature data, the environment scene data is matched using the environment scene library to obtain a matching current environment scene.

[0012] Furthermore, according to the current environment scenario, the vending machine is used to make shopping recommendations to the user, including: According to the current environment scene, based on the set matching relationship database between the current environment scene and the item recommendation database, obtain the item recommendation database corresponding to the current environment scene; According to the item recommendation database, the item recommendation categories in the item recommendation database are recommended to the user by utilizing the item recommendation strategy in the item recommendation database; Among them, the item recommendation strategy is formulated based on the strategy formulation model; The strategy formulation model is obtained based on the user shopping demand prediction model and the user shopping correction model; The user shopping demand prediction model uses a neural network prediction model to predict the user's shopping demand based on the user's individual label data, the user's personalized shopping data, and the user's initial operation data. The user's individual label data includes whether the user is a new user or an old user; the user's personalized shopping data includes the user's preference for items and brands, as well as the user's attitude towards whether to accept other recommendations; the user's initial operation data is the user's initial operation of the vending machine based on his or her own shopping needs. The user shopping correction model is used to obtain the user's shopping correction value based on the user's progressive operation data and interaction data; obtain the shopping demand recommendation probability value based on the corresponding matching relationship library between the shopping correction value and the shopping demand recommendation probability value, and take the shopping demand corresponding to the shopping demand recommendation probability value greater than the set probability threshold as the user's shopping demand; The interactive data refers to the data generated by the interaction between the virtual AI digital human on the vending machine and the user; the user shopping correction model is expressed as follows: Among them, P represents the user's shopping correction value, represents the modified contribution value generated by the i-th progressive operation step performed by the user; i represents the i-th operation step, and k represents the number of operation steps performed by the user from the beginning of the operation to the completion of the shopping; Represents the modified contribution value generated by the j-th interaction process performed by the user; j represents the j-th interaction process, and m represents the number of interaction steps performed by the user from the beginning of the interaction to the completion of the shopping.

[0013] User shopping recommendation system based on scene recognition, including: The environment scene library construction module is used to build the environment scene library of the vending machine based on the user's shopping history; The current environment scene acquisition module is used to obtain the current environment scene of the vending machine based on the scene recognition component of the vending machine and the environment scene library; The shopping recommendation module is used to make shopping recommendations to users using vending machines based on the current environment scenario.

[0014] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing a user shopping recommendation method based on scene recognition.

[0015] A device includes a processor and a memory; the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory, so that the device executes a user shopping recommendation method based on scene recognition.

[0016] Compared with the prior art, the present invention has the following advantages and beneficial effects: based on the user's shopping history records, an environmental scene library of the vending machine is constructed, and based on the environmental scene library, the current environmental scene of the vending machine is obtained, and the vending machine is used to make shopping recommendations to the user, which can improve the efficiency and quality of the vending machine's shopping recommendations to the user, which is conducive to improving the user's shopping experience and improving the intelligence level of shopping recommendations.

[0017] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0018] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 A schematic diagram of steps for a user shopping recommendation method based on scene recognition; Figure 2 A schematic diagram of the method steps for constructing an environmental scene database for vending machines; Figure 3 Schematic diagram of the user shopping recommendation system based on scene recognition. DETAILED DESCRIPTION

[0020] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0021] The present invention provides a user shopping recommendation method based on scene recognition, such as Figure 1 As shown, including: Build a vending machine environment scene library based on user shopping history; Based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library; Based on the current environment, vending machines are used to make shopping recommendations to users.

[0022] The working principle of the above technical solution is: in order to realize the user shopping recommendation method based on scene recognition, the present invention first constructs an environmental scene library of the vending machine based on the user's shopping history records; then, based on the scene recognition component of the vending machine, according to the environmental scene library, the current environmental scene of the vending machine is obtained; finally, according to the current environmental scene, the vending machine is used to make shopping recommendations to the user.

[0023] The beneficial effects of the above technical solution are: by adopting the solution provided in this embodiment, an environmental scene library of the vending machine is constructed according to the user's shopping history records, and the current environmental scene of the vending machine is obtained according to the environmental scene library. The vending machine is used to make shopping recommendations to the user, which can improve the efficiency and quality of the vending machine's shopping recommendations to the user, which is conducive to improving the user's shopping experience and improving the intelligence level of shopping recommendations.

[0024] In one embodiment, Figure 2 As shown in the figure, based on the user's shopping history, an environmental scene library for vending machines is constructed, including: Get the user's shopping history; Based on historical records, a dataset of environmental scenarios used by users when shopping is established; According to the environmental scene data set, the environmental scene classification model is used to classify the environmental scenes, and an environmental scene library is constructed based on the classified environmental scenes.

[0025] The working principle of the above technical solution is: in order to build an environmental scene database for vending machines based on the user's shopping history records, the present invention first obtains the user's shopping history records; then, based on the historical records, establishes an environmental scene data set for the user's shopping; finally, based on the environmental scene data set, uses the environmental scene classification model to classify the environmental scenes, and builds an environmental scene library based on the classified environmental scenes.

[0026] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, by constructing an environmental scene library based on the classified environmental scenes, the quality of the construction of the environmental scene library can be guaranteed.

[0027] In one embodiment, based on historical records, a data set of environmental scenes when a user is shopping is established, including: Obtain user shopping behavior information from historical records, and generate a behavior information attribute data set based on the user shopping behavior information; Obtain the environmental feature data of the vending machine corresponding to the time when the user's shopping behavior occurs, and generate an environmental feature data set; Based on the behavior information attribute dataset and the environmental feature dataset, an environmental scene dataset is generated.

[0028] The working principle of the above technical solution is as follows: in order to establish an environmental scene data set when a user is shopping, the present invention first obtains the user's shopping behavior information in the historical records, and generates a behavior information attribute data set based on the user's shopping behavior information; the user's shopping behavior information refers to the user's shopping behavior preferences and behavior habits, and the behavior preference refers to the type of goods that are preferred to be purchased, including beverages, drinking water, snacks, etc.; and the brand of goods that are preferred to be purchased; behavior habits refer to the time of purchasing goods, the quantity and type combination of purchased goods, the number of times goods are purchased, the operation selection time when purchasing goods, etc.; then the environmental feature data of the vending machine corresponding to the time when the user's shopping behavior occurs is obtained to generate an environmental feature data set; finally, based on the behavior information attribute data set and the environmental feature data set, an environmental scene data set is generated.

[0029] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, by generating an environmental scene data set, basic conditions are provided for the subsequent construction of an environmental scene database.

[0030] In one embodiment, the environmental characteristic data of the vending machine corresponding to the time when the user's shopping behavior occurs is obtained to generate an environmental characteristic data set, including: Based on the environmental sensor in the scene recognition component of the vending machine, the environmental temperature and humidity data are collected; Based on the camera device in the scene recognition component of the vending machine, the user's clothing data and user traffic data are obtained; Acquire user voice data based on the high-sensitivity microphone in the scene recognition component of the vending machine; Aggregate environmental temperature and humidity data, user clothing data, user traffic data, and user voice data to generate an environmental feature data set.

[0031] The working principle of the above technical solution is: in order to obtain the environmental characteristic data of the vending machine corresponding to the time when the user's shopping behavior occurs and generate an environmental characteristic data set, the present invention first collects and obtains the environmental temperature and humidity data based on the environmental sensor in the scene recognition component of the vending machine, so as to obtain the temperature and humidity of the environment of the vending machine; then, based on the camera device in the scene recognition component of the vending machine, obtains the user's clothing data and user flow data, which can reflect the weather conditions and traffic conditions of the user; finally, based on the high-sensitivity microphone in the scene recognition component of the vending machine, obtains the user's voice data, which can reflect the user's demand for the items discussed for purchase; summarizes the environmental temperature and humidity data, the user's clothing data, the user flow data and the user voice data to generate an environmental characteristic data set.

[0032] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, an environmental feature data set is generated based on environmental temperature and humidity data, user clothing data, user traffic data and user voice data, thereby ensuring the comprehensiveness of the environmental feature data set.

[0033] In one embodiment, generating an environmental scene dataset based on a behavior information attribute dataset and an environmental feature dataset includes: Using the behavior information attribute dataset and the environmental feature dataset, several user shopping environment scenarios are generated based on the scenario generation model; among which, the scenario generation model is generated based on the training of the deep neural network model.

[0034] The working principle of the above technical solution is: in order to generate an environmental scene data set, the present invention uses a behavioral information attribute data set and an environmental feature data set to generate several user shopping environment scenes based on a scene generation model; wherein the scene generation model is generated based on deep neural network model training.

[0035] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, a comprehensive environmental scene can be easily generated through a scene generation model.

[0036] In one embodiment, based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library, including: Obtain the user's current shopping behavior information and generate current behavior information attribute data; Based on the scene recognition component of the vending machine, identify and obtain the current environmental feature data of the vending machine; According to the current behavior information attribute data and the current environment feature data, the environment scene data is matched using the environment scene library to obtain a matching current environment scene.

[0037] The working principle of the above technical solution is: in order to realize the scene recognition component based on the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library. The present invention first obtains the user's current shopping behavior information and generates current behavior information attribute data; then based on the scene recognition component of the vending machine, it identifies and obtains the current environment feature data of the vending machine; then, according to the current behavior information attribute data and the current environment feature data, the environment scene library is used to match the environment scene data to obtain a matching current environment scene.

[0038] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, by performing matching selection, it is possible to ensure that an accurate current environment scene is obtained.

[0039] In one embodiment, based on the current environment scenario, a vending machine is used to make shopping recommendations to a user, including: According to the current environment scene, based on the set matching relationship database between the current environment scene and the item recommendation database, obtain the item recommendation database corresponding to the current environment scene; According to the item recommendation database, the item recommendation categories in the item recommendation database are recommended to the user by utilizing the item recommendation strategy in the item recommendation database; Among them, the item recommendation strategy is formulated based on the strategy formulation model; The strategy formulation model is obtained based on the user shopping demand prediction model and the user shopping correction model; The user shopping demand prediction model uses a neural network prediction model to predict the user's shopping demand based on the user's individual label data, the user's personalized shopping data, and the user's initial operation data. The user's individual label data includes whether the user is a new user or an old user; the user's personalized shopping data includes the user's preference for items and brands, as well as the user's attitude towards whether to accept other recommendations; the user's initial operation data is the user's initial operation of the vending machine based on his or her own shopping needs. The user shopping correction model is used to obtain the user's shopping correction value based on the user's progressive operation data and interaction data; obtain the shopping demand recommendation probability value based on the corresponding matching relationship library between the shopping correction value and the shopping demand recommendation probability value, and take the shopping demand corresponding to the shopping demand recommendation probability value greater than the set probability threshold as the user's shopping demand; The interactive data refers to the data generated by the interaction between the virtual AI digital human on the vending machine and the user; the user shopping correction model is expressed as follows: Among them, P represents the user's shopping correction value, represents the modified contribution value generated by the i-th progressive operation step performed by the user; i represents the i-th operation step, and k represents the number of operation steps performed by the user from the beginning of the operation to the completion of the shopping; Represents the modified contribution value generated by the j-th interaction process performed by the user; j represents the j-th interaction process, and m represents the number of interaction steps performed by the user from the beginning of the interaction to the completion of the shopping.

[0040] The working principle of the above technical solution is as follows: in order to make shopping recommendations to users by using vending machines, the present invention obtains an item recommendation database corresponding to the current environment scene based on a matching relationship database between the current environment scene and the item recommendation database; then, based on the item recommendation database, the item recommendation strategy in the item recommendation database is used to recommend the item recommendation categories in the item recommendation database to the user; Among them, the item recommendation strategy is formulated based on the strategy formulation model; the strategy formulation model is obtained based on the user shopping demand prediction model and the user shopping correction model; wherein, the user shopping demand prediction model, based on the user's individual label data, the user's personalized shopping data, the user's initial operation data, uses a neural network prediction model to predict the user's shopping needs; wherein, the user's individual label data includes whether the user is a new user or an old user; the user's personalized shopping data includes the user's preference for items and brands, as well as the attitude towards whether to accept other recommendations; the user's initial operation data is: the user's initial operation of the vending machine according to his own shopping needs. The user shopping correction model is used to obtain the user's shopping correction value based on the user's progressive operation data and interaction data; according to the corresponding matching relationship library between the shopping correction value and the shopping demand recommendation probability value, the shopping demand recommendation probability value is obtained, and the shopping demand corresponding to the shopping demand recommendation probability value greater than the set probability threshold is used as the user's shopping demand; The interactive data refers to the data generated by the interaction between the virtual AI digital human on the vending machine and the user; the user shopping correction model is expressed as follows: Among them, P represents the user's shopping correction value, represents the modified contribution value generated by the i-th progressive operation step performed by the user; i represents the i-th operation step, and k represents the number of operation steps performed by the user from the beginning of the operation to the completion of the shopping; Represents the modified contribution value generated by the j-th interaction process performed by the user; j represents the j-th interaction process, and m represents the number of interaction steps performed by the user from the beginning of the interaction to the completion of the shopping.

[0041] The beneficial effect of the above technical solution is: by adopting the solution provided by this embodiment, based on the current environmental scene, based on the matching relationship database between the set current environmental scene and the item recommendation database, an item recommendation database corresponding to the current environmental scene is obtained; then, based on the item recommendation database, the item recommendation strategy in the item recommendation database is used to recommend the item recommendation categories in the item recommendation database to the user, thereby ensuring the accuracy and flexibility of shopping recommendations.

[0042] User shopping recommendation system based on scene recognition, such as Figure 3 As shown, including: The environment scene library construction module is used to build the environment scene library of the vending machine based on the user's shopping history; The current environment scene acquisition module is used to obtain the current environment scene of the vending machine based on the scene recognition component of the vending machine and the environment scene library; The shopping recommendation module is used to make shopping recommendations to users using vending machines based on the current environment scenario.

[0043] The working principle of the above technical solution is as follows: in order to realize a user shopping recommendation system based on scene recognition, the present invention proposes an environmental scene library construction module, which is used to construct an environmental scene library of a vending machine based on the user's shopping history records; a current environmental scene acquisition module is proposed, which is used to obtain the current environmental scene of the vending machine based on the environmental scene library based on the scene recognition component of the vending machine; a shopping recommendation module is proposed, which is used to make shopping recommendations to the user using the vending machine based on the current environmental scene; based on the same idea as the method in the above embodiment, the user shopping recommendation system based on scene recognition provided by the present application can realize the method in the above embodiment.

[0044] The beneficial effects of the above technical solution are: by adopting the solution provided in this embodiment, an environmental scene library of the vending machine is constructed according to the user's shopping history records, and the current environmental scene of the vending machine is obtained according to the environmental scene library. The vending machine is used to make shopping recommendations to the user, which can improve the efficiency and quality of the vending machine's shopping recommendations to the user, which is conducive to improving the user's shopping experience and improving the intelligence level of shopping recommendations.

[0045] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing a user shopping recommendation method based on scene recognition.

[0046] The working principle of the above technical solution is: instructions can be stored through a computer-readable storage medium, and the instructions can be loaded by a processor to execute a user shopping recommendation method based on scene recognition.

[0047] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, the storage of instructions can be realized through a computer-readable storage medium.

[0048] A device includes a processor and a memory; the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory, so that the device executes a user shopping recommendation method based on scene recognition.

[0049] The working principle of the above technical solution is: instructions can be stored in the memory, and the instructions stored in the memory can be executed by the processor, so that the device executes the user shopping recommendation method based on scene recognition.

[0050] The beneficial effect of the above technical solution is: by adopting the solution provided in this embodiment, through the memory and the processor, it can be implemented on the device to execute the user shopping recommendation method based on scene recognition.

[0051] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A user shopping recommendation method based on scene recognition, characterized in that: include: Build a vending machine environment scene library based on user shopping history; Based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library; Based on the current environment, vending machines are used to make shopping recommendations to users.

2. The user shopping recommendation method based on scene recognition according to claim 1, characterized in that: Based on the user's shopping history, build an environmental scene library for vending machines, including: Get the user's shopping history; Based on historical records, a dataset of environmental scenarios used by users when shopping is established; According to the environmental scene data set, the environmental scene classification model is used to classify the environmental scenes, and an environmental scene library is constructed based on the classified environmental scenes.

3. The method for user shopping recommendation based on scene recognition according to claim 2, characterized in that: Based on historical records, a dataset of environmental scenarios of users when shopping is established, including: Obtain user shopping behavior information from historical records, and generate a behavior information attribute data set based on the user shopping behavior information; Obtain the environmental feature data of the vending machine corresponding to the time when the user's shopping behavior occurs, and generate an environmental feature data set; Based on the behavior information attribute dataset and the environmental feature dataset, an environmental scene dataset is generated.

4. The user shopping recommendation method based on scene recognition according to claim 3 is characterized in that: Obtain the environmental feature data of the vending machine corresponding to the user's shopping behavior, and generate an environmental feature data set, including: Based on the environmental sensor in the scene recognition component of the vending machine, the environmental temperature and humidity data are collected; Based on the camera device in the scene recognition component of the vending machine, the user's clothing data and user traffic data are obtained; Acquire user voice data based on the high-sensitivity microphone in the scene recognition component of the vending machine; Aggregate environmental temperature and humidity data, user clothing data, user traffic data, and user voice data to generate an environmental feature data set.

5. The user shopping recommendation method based on scene recognition according to claim 3 is characterized in that: Based on the behavior information attribute dataset and the environmental feature dataset, an environmental scene dataset is generated, including: Using the behavior information attribute dataset and the environmental feature dataset, several user shopping environment scenarios are generated based on the scenario generation model; among which, the scenario generation model is generated based on the training of the deep neural network model.

6. The method for user shopping recommendation based on scene recognition according to claim 4, characterized in that: Based on the scene recognition component of the vending machine, the current environment scene of the vending machine is obtained according to the environment scene library, including: Obtain the user's current shopping behavior information and generate current behavior information attribute data; Based on the scene recognition component of the vending machine, identify and obtain the current environmental feature data of the vending machine; According to the current behavior information attribute data and the current environment feature data, the environment scene data is matched using the environment scene library to obtain a matching current environment scene.

7. The method for user shopping recommendation based on scene recognition according to claim 1, characterized in that: Based on the current environment, vending machines are used to make shopping recommendations to users, including: According to the current environment scene, based on the set matching relationship database between the current environment scene and the item recommendation database, obtain the item recommendation database corresponding to the current environment scene; According to the item recommendation database, the item recommendation categories in the item recommendation database are recommended to the user by utilizing the item recommendation strategy in the item recommendation database; Among them, the item recommendation strategy is formulated based on the strategy formulation model; The strategy formulation model is obtained based on the user shopping demand prediction model and the user shopping correction model; The user shopping demand prediction model uses a neural network prediction model to predict the user's shopping demand based on the user's individual label data, the user's personalized shopping data, and the user's initial operation data. The user's individual label data includes whether the user is a new user or an old user; the user's personalized shopping data includes the user's preference for items and brands, as well as the user's attitude towards whether to accept other recommendations; the user's initial operation data is the user's initial operation of the vending machine based on his or her own shopping needs. The user shopping correction model is used to obtain the user's shopping correction value based on the user's progressive operation data and interaction data; obtain the shopping demand recommendation probability value based on the corresponding matching relationship library between the shopping correction value and the shopping demand recommendation probability value, and take the shopping demand corresponding to the shopping demand recommendation probability value greater than the set probability threshold as the user's shopping demand; The interactive data refers to the data generated by the interaction between the virtual AI digital human on the vending machine and the user; the user shopping correction model is expressed as follows: Among them, P represents the user's shopping correction value, represents the modified contribution value generated by the i-th progressive operation step performed by the user; i represents the i-th operation step, and k represents the number of operation steps performed by the user from the beginning of the operation to the completion of the shopping; Represents the modified contribution value generated by the j-th interaction process performed by the user; j represents the j-th interaction process, and m represents the number of interaction steps performed by the user from the beginning of the interaction to the completion of the shopping.

8. A user shopping recommendation system based on scene recognition, characterized in that: include: The environment scene library construction module is used to build the environment scene library of the vending machine based on the user's shopping history; The current environment scene acquisition module is used to obtain the current environment scene of the vending machine based on the scene recognition component of the vending machine and the environment scene library; The shopping recommendation module is used to make shopping recommendations to users using vending machines based on the current environment scenario.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the user shopping recommendation method based on scene recognition as described in any one of claims 1-7.

10. A device, characterized in that: It includes a processor and a memory; the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory, so that the device executes the user shopping recommendation method based on scene recognition as described in any one of claims 1-7.

Images