Information pushing method and device, equipment, storage medium and product

Abstract

The invention discloses an information pushing method and device, equipment, a storage medium and a product, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the data preprocessing of user information data and art work data, and obtaining user features and art work features; the user features and the artistic work features are input into an information push model, an initial push result is obtained, and the information push model is constructed based on multi-modal data corresponding to historical users and information features corresponding to artistic works; and pushing the initial pushing result to the target user. According to the method provided by the embodiment of the invention, the obtained user features and the obtained artistic work features are input into the information push model to obtain the initial push result, so that the user can watch the real-time recommended artistic work, and the technical problem that the recommendation logic of the traditional recommendation method for the artistic work is solidified is solved; and the impression experience and the purchase desire of the user are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an information push method, device, equipment, storage medium and product. Background Art

[0002] With the development of artificial intelligence and digital display technology, the field of art screen display and home decoration has ushered in new changes. Traditional art platforms (such as TV art galleries, family albums and decorative paintings, etc.) face the lack of personalized and immersive experience. Specifically, the art platform's recommendation logic for artworks is relatively rigid, so that the art platform recommends artworks to users that are not what the users like, which in turn leads to poor user experience and lack of purchasing desire. Therefore, the industry is in urgent need of a method that can accurately recommend artworks that match the user's preferences. Summary of the invention

[0003] The main purpose of this application is to provide an information push method, device, equipment, storage medium and product, aiming to solve the technical problem that the existing technology cannot accurately recommend artworks that match the user's preferences to the user.

[0004] To achieve the above purpose, the present application provides an information push method, which includes the following steps:

[0005] Performing data preprocessing on the user information data and the artwork data to obtain user features and artwork features;

[0006] Inputting the user characteristics and the artwork characteristics into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information characteristics corresponding to artworks;

[0007] The initial push result is pushed to the target user.

[0008] In addition, to achieve the above purpose, the present application also proposes a method for constructing an information push model, which includes the following steps:

[0009] Constructing a historical user portrait based on historical user information data and artwork data, wherein the historical user portrait is used to indicate the matching status of the historical user with the artwork;

[0010] Establishing a model training set, a model testing set, and a model verification set according to the historical user information data, the artwork data, and the historical user portraits;

[0011] The initial model is trained and optimized based on the model training set, the model test set and the model verification set to obtain an information push model.

[0012] In addition, to achieve the above purpose, the present application also proposes an information push device, which includes:

[0013] A data processing module, used for pre-processing user information data and artwork data to obtain user features and artwork features;

[0014] A result generation module, used for inputting the user characteristics and the artwork characteristics into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information characteristics corresponding to artworks;

[0015] The result push module is used to push the initial push result to the target user.

[0016] In addition, to achieve the above-mentioned purpose, the present application also proposes an information push device, which includes: a memory, a processor, and an information push program stored in the memory and executable on the processor, wherein the information push program is configured to implement the steps of the information push method described above.

[0017] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and an information push program is stored on the storage medium. When the information push program is executed by the processor, the steps of the information push method described above are implemented.

[0018] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer program product, which includes an information push program, and when the information push program is executed by a processor, the steps of the information push method described above are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the structure of an information push device in a hardware operating environment involved in an embodiment of the present application;

[0020] Figure 2 This is a flow chart of the first embodiment of the information push method of this application;

[0021] Figure 3 This is a flow chart of the second embodiment of the method for pushing application information;

[0022] Figure 4 This is a flow chart of the first embodiment of the method for building an information push model for this application;

[0023] Figure 5 This is a structural block diagram of the first embodiment of the information push device of this application.

[0024] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0026] Reference Figure 1 , Figure 1 This is a schematic diagram of the information push device structure of the hardware operating environment involved in the embodiment of the present application.

[0027] like Figure 1 As shown, the information push device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0028] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the information push device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0029] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an information push program.

[0030] exist Figure 1In the information push device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the information push device of the present application can be set in the information push device, and the information push device calls the information push program stored in the memory 1005 through the processor 1001, and executes the information push method provided in the embodiment of the present application.

[0031] The present application embodiment provides an information push method, referring to Figure 2 , Figure 2 This is a flow chart of the first embodiment of the information push method of this application.

[0032] In this embodiment, the information push method includes the following steps:

[0033] Step S10: pre-process the user information data and artwork data to obtain user features and artwork features.

[0034] It should be noted that the execution subject of the method of this embodiment can be a terminal device with model building, data processing and program running functions, such as a smart phone, a computer, etc., or an electronic device with the same or similar functions, such as the above-mentioned information push device. The following takes the information push device (hereinafter referred to as the recommendation device) as an example to illustrate this embodiment and the following embodiments.

[0035] It is understandable that the above user information data and artwork data can be collected from art platforms, and the above art platforms may include TV art galleries, art venues (such as museums, exhibition halls, etc.), home decoration exhibition halls, etc., and this embodiment does not limit this.

[0036] It should be understood that the above-mentioned user information data may include user behavior data and user attribute data, wherein user behavior data may include browsing history, search history, ratings, collections, purchase behavior, etc., and user attribute data may include age, gender, geographic location, device type, etc.; the above-mentioned artwork data may include the description, label, author, creation year, subject matter, style, size, price, etc. of the artwork.

[0037] In a specific implementation, the user information data and artwork data may be subjected to data cleaning, data conversion, data screening, feature extraction and other operations to implement data preprocessing, thereby obtaining the user features and artwork features.

[0038] Step S20: Inputting the user features and the artwork features into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information features corresponding to artworks.

[0039] It should be noted that the above information push model can be used to generate a list of artworks that meet the user's preferences (ie, the above initial push result).

[0040] Step S30: Push the initial push result to the target user.

[0041] In a specific implementation, the above initial push result can be sent to an art platform, and then the initial push result can be displayed to the target user through the art platform.

[0042] This embodiment obtains user features and artwork features by preprocessing user information data and artwork data; inputs the user features and artwork features into the information push model to obtain an initial push result, and the information push model is constructed based on the multimodal data corresponding to the historical user and the information features corresponding to the artwork; and pushes the initial push result to the target user. Since the above method of this embodiment inputs the acquired user features and artwork features into the information push model to obtain the initial push result, the user can watch the real-time recommended artwork, thereby solving the technical problem of the traditional recommendation method for the solidification of the recommendation logic of the artwork, thereby improving the user's viewing experience and purchasing desire. In addition, the information push method mentioned in this embodiment can also be applied to the commercial display field, deployed in museums, exhibition tubes, etc. to make mural TVs, automatically recommend different artworks and paintings to the visiting crowd, and can also provide users who love art for home use, etc., which can layout the high-end market of TV products and the commercial display field.

[0043] refer to Figure 3 , Figure 3 This is a flow chart of the second embodiment of the information push method of this application.

[0044] In a feasible implementation manner, the step S10 may include:

[0045] Step S101: performing data conversion and data cleaning on user information data and artwork data to obtain cleaned user information data and cleaned artwork data.

[0046] It should be understood that since the user information data and artwork data in different art platforms may have format differences, and these differences may lead to low data recognition efficiency or even data recognition errors. Therefore, in this embodiment, before data cleaning, the above-mentioned user information data and artwork data can be format converted to achieve data format unification. In particular, the target format of data conversion can be flexibly set based on actual conditions (for example, the format with the highest proportion in the current user information data and artwork data can be used as the target format to shorten the data conversion cycle), and this embodiment does not limit this.

[0047] Step S102: extracting features from the cleaned user information data and the cleaned artwork data to obtain user features and artwork features.

[0048] In a specific implementation, in the above-mentioned user information data and the above-mentioned artwork data, invalid records can be deleted and incomplete records can be filled with default values. More specifically, invalid records may include erroneous data (data caused by input errors, system failures or other reasons), outliers (caused by errors or data entry errors), duplicate records (identical or very similar data), irrelevant data (data that is irrelevant to the current analysis or business goals and may interfere with the analysis results), null values ​​(no data in the field), etc. Incomplete records may include missing values ​​(values ​​of certain fields in the record are missing), inconsistent data (certain fields in the data are inconsistent in format, type or content, such as the format of date and style fields is not uniform).

[0049] In a feasible implementation manner, the step S102 may include:

[0050] Step S1021: extracting features from the cleaned user information data based on the number of clicks, browsing time and evaluation of each artwork by the user to obtain user features.

[0051] It should be noted that the above user characteristics can be used to indicate what types of artworks the user often views or what artists or art styles the user is interested in.

[0052] In a specific implementation, it is possible to determine which types of artworks the user frequently views and which artists or art styles the user is interested in based on the number of clicks, browsing time, and evaluation of each artwork. Among them, it is possible to determine which types of artworks the user frequently views based on the number of clicks and browsing time of each artwork. For example, when a user clicks on a certain type of artwork more than 10 times in a week or stays on a certain type of artwork page for more than 30 minutes in a week, it can be considered that the user frequently views this type of artwork; it is possible to determine which types of artworks the user is interested in based on the number of clicks and evaluations of each artwork. For example, if the number of clicks on a certain type of artwork accounts for more than 30% of the total number of clicks of the user or the user collects or rates a certain type of artwork more than 5 times, it can be considered that the user is interested in this type of artwork. Of course, the above quantitative values ​​(such as 10 times, 30 minutes, 30%) are only used for illustration and not for limitation. The specific quantitative values ​​can be flexibly set based on the actual scenario, and this embodiment does not limit this.

[0053] Step S1022: extracting features from the cleaned artwork data according to the content information and visual information of the artwork to obtain artwork features.

[0054] It should be noted that the above-mentioned artwork features can be used to present the content information and visual information of the artwork. More specifically, the content information may include the description, label, author, creation year, subject matter, style, size, price and other attributes of the artwork, and the visual information may include the color distribution, shape, texture and the like of the artwork.

[0055] It should be understood that new features can also be created based on the above-mentioned user features and artwork features, such as the frequency of interaction between users and artworks, the popularity of artworks, etc., so as to establish relationship characteristics between user features and artwork features, thereby enabling a more comprehensive analysis of user preferences and interests.

[0056] In a feasible implementation manner, the information push method may further include:

[0057] Step S40: Acquire the real-time interaction data of the user with respect to the initial push result and the real-time monitoring data of the system with respect to the initial push result.

[0058] It should be noted that the above-mentioned real-time interaction data may refer to data used to indicate user opinions based on user feedback based on the initial push results. For example, users can classify artworks in the initial push results as disliked, favorited, etc. in the real-time interactive interface of the art platform in order to further optimize the recommendation. The above-mentioned real-time monitoring data may be data obtained by the system in the recommendation device continuously monitoring user behavior and the art platform. For example, the system can automatically optimize and iteratively upgrade through initial push results, user behavior, user feedback, etc. to monitor the performance and effect of the system and ensure the stable operation of the system.

[0059] Step S50: adjusting the initial push result in real time based on the real-time interaction data and the real-time monitoring data to obtain a real-time push result.

[0060] It is understandable that the above initial push results only include the artworks preferred by the user at the current moment. However, with the existence of objective factors such as changes in user needs and updates of artworks, the initial push results may not match the user's preferences at the next moment. Therefore, in this embodiment, it is necessary to continuously adjust the initial push results in real time to ensure that artworks that meet the user's preferences can be recommended to the user based on the real-time push results at every moment.

[0061] This embodiment converts and cleanses user information data and artwork data to obtain cleaned user information data and cleaned artwork data; extracts features from the cleaned user information data based on the number of clicks, browsing time and evaluations of each artwork by the user to obtain user features; extracts features from the cleaned artwork data based on the content information and visual information of the artwork to obtain artwork features; displays the initial push result to the user on the art platform, and obtains the real-time interaction data of the user on the initial push result and the real-time monitoring data of the system on the initial push result; adjusts the initial push result in real time based on the real-time interaction data and the real-time monitoring data to obtain a real-time push result. The above method of this embodiment extracts user features and artwork features, and analyzes the user's preferences and interests more comprehensively based on the relationship characteristics between user features and artwork features; in addition, the system corresponding to the recommendation device in this embodiment also has an automatic iterative optimization function, thereby ensuring the correctness of the recommendation effect and the stability of the system.

[0062] refer to Figure 4 , Figure 4 This is a flow chart of the first embodiment of the information push model construction method of this application.

[0063] In this embodiment, the information push model construction method includes the following steps:

[0064] Step S1: construct a historical user portrait based on historical user information data and artwork data, wherein the historical user portrait is used to represent the matching status of historical users with artworks.

[0065] It is understandable that the above-mentioned historical user information data may include historical browsing behavior, historical search behavior, historical ratings, historical collection behavior, historical purchase behavior, etc., and the above-mentioned artwork data may include the description, label, author, creation year, subject matter, style, size, price, etc. of the artwork.

[0066] It should be understood that the above-mentioned historical user information data and artwork data may be used to analyze which art styles, themes or artists the historical users have preferences for, thereby constructing the above-mentioned historical user portrait.

[0067] In a specific implementation, the historical user portrait can be further refined by identifying user behavior patterns (such as users regularly visiting specific types of artworks or artists) and user feedback (such as users rating or commenting on artworks).

[0068] Step S2: establishing a model training set, a model testing set and a model verification set according to the historical user information data, the artwork data and the historical user portraits.

[0069] Step S3: training and optimizing the initial model based on the model training set, the model test set and the model verification set to obtain an information push model.

[0070] It should be noted that the above initial model can be a neural network model (such as a collaborative filtering neural network model, a convolutional neural network model, a recurrent neural network, etc.), which includes an input layer, a hidden layer and an output layer.

[0071] In a specific implementation, the model training set can be used to optimize the model parameters of the initial model through the back propagation algorithm, and the model test set and the model validation set can be used to evaluate the model performance to adjust the model structure and parameters. More specifically, the optimization process of the initial model can be based on hyperparameter adjustment (for example, using grid search or random search to optimize the hyperparameters of the model), regularization (preventing overfitting, such as improving the generalization ability of the model through L1 or L2 regularization), cross-validation (for example, using K-fold cross-validation to evaluate model stability and performance) and feedback loops (for example, using user feedback data to continuously adjust and optimize the model).

[0072] In a feasible implementation manner, the step S1 may include:

[0073] Step S11: determining the artwork matching similarity between historical users and the artwork feature similarity between artworks based on the historical user information data and artwork data.

[0074] Step S12: constructing a historical user portrait based on the artwork matching similarity and the artwork feature similarity.

[0075] It is understandable that the above-mentioned similarity of artwork matching between historical users can be expressed as the degree of matching between different users of the same type of artwork. For example, assuming that the similarity threshold is 60% (i.e., the similarity of artwork matching between users is greater than or equal to 60%, indicating a high degree of matching, and less than 60%, indicating a low degree of matching), assuming that the similarity of artwork matching between user A and user B is 70%, then when constructing the historical user portrait of user A / B, the artwork matching with B / A can be referenced.

[0076] Similarly, the above-mentioned similarity of artwork features can be expressed as the degree of similarity of features between different artworks. For example, assuming that the similarity threshold is 70% (i.e., the similarity of artwork features between different artworks is greater than or equal to 70%, indicating a high degree of similarity, and less than 70% indicating a low degree of similarity), assuming that the similarity of artwork features between artwork A and artwork B is 80%, then when artwork A is recommended to a user, artwork B will also be recommended to the user, thereby obtaining a more comprehensive user portrait.

[0077] In a feasible implementation manner, the step S12 may include:

[0078] Step S121: determining a first matching artwork according to a collaborative filtering algorithm and the artwork matching similarity, and determining a second matching artwork according to a content-based recommendation algorithm and the artwork feature similarity.

[0079] Step S122: constructing a historical user portrait based on the first matching artwork and the second matching artwork.

[0080] It should be understood that the collaborative filtering algorithm is an algorithm for recommending artworks based on user behavior data. It predicts the user's preference for a specific artwork by analyzing the similarity between users, and then makes recommendations. The similarity between users is evaluated by determining the user's rating or purchase history. For example, if the similarity between user A and user B is very high, the system may recommend the artwork that user B likes to user A. The content-based recommendation algorithm is an algorithm for personalized recommendation by analyzing the content of items and user behavior. It can recommend artworks that match the user's preferences based on the user's historical preferences.

[0081] In a specific implementation, a hybrid recommendation algorithm (ie, combining the collaborative filtering algorithm and the content-based recommendation algorithm) may be used to construct the information push model, thereby improving the accuracy and coverage of the push results.

[0082] In the specific implementation, after determining the similarity of artwork matching between historical users and the similarity of artwork features between artworks, new artworks that historical users may be interested in can be discovered based on the similarity between users and the similarity between artworks, thus successfully constructing the above-mentioned historical user portrait.

[0083] This embodiment determines the artwork matching similarity between historical users and the artwork feature similarity between artworks based on historical user information data and artwork data; determines a first matching artwork based on a collaborative filtering algorithm and the artwork matching similarity, and determines a second matching artwork based on a content-based recommendation algorithm and the artwork feature similarity; constructs a historical user portrait based on the first matching artwork and the second matching artwork, and the historical user portrait is used to represent the matching of the artwork by the historical user; establishes a model training set, a model test set, and a model verification set based on the historical user information data, the artwork data, and the historical user portrait; trains and optimizes the initial model based on the model training set, the model test set, and the model verification set to obtain an information push model. The above method of this embodiment combines a collaborative filtering algorithm and a content-based recommendation algorithm to construct a historical user portrait, and then constructs an information push model by acquiring historical user information data, artwork data, and historical user portraits, so that the recommended artwork is more matched with the user's personal preferences, thereby improving the user's viewing experience and purchase desire.

[0084] In addition, an embodiment of the present application further proposes a storage medium, on which an information push program is stored. When the information push program is executed by a processor, the steps of the information push method described above are implemented.

[0085] Reference Figure 5 , Figure 5 This is a structural block diagram of the first embodiment of the information push device of this application.

[0086] like Figure 5 As shown, the information push device proposed in the embodiment of the present application includes:

[0087] The data processing module 501 is used to perform data preprocessing on the user information data and the artwork data to obtain user features and artwork features;

[0088] A result generation module 502, for inputting the user characteristics and the artwork characteristics into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information characteristics corresponding to artworks;

[0089] The result push module 503 is used to push the initial push result to the target user.

[0090] This embodiment performs data preprocessing on user information data and artwork data to obtain user features and artwork features; inputs the user features and artwork features into an information push model to obtain an initial push result, the information push model is constructed based on multimodal data corresponding to historical users and information features corresponding to artworks; and pushes the initial push result to the target user. Since the above method of this embodiment inputs the acquired user features and artwork features into the information push model to obtain the initial push result, and the initial push result will be adjusted in real time during the display to the user, the user can view the real-time recommended artwork, thereby solving the technical problem of the solidified recommendation logic of artworks in the traditional recommendation method, and further improving the user's viewing experience and purchasing desire.

[0091] Based on the first embodiment of the information pushing device of the present application, a second embodiment of the information pushing device of the present application is proposed.

[0092] In this embodiment, the data processing module 501 is also used to perform data conversion and data cleaning on the user information data and the artwork data to obtain cleaned user information data and cleaned artwork data; perform feature extraction on the cleaned user information data and the cleaned artwork data to obtain user features and artwork features.

[0093] Furthermore, the data processing module 501 is also used to extract features from the cleaned user information data based on the number of clicks, browsing time and evaluations of users on each artwork to obtain user features; and to extract features from the cleaned artwork data based on the content information and visual information of the artwork to obtain artwork features.

[0094] Furthermore, the result push module 503 is also used to display the initial push result to the user in the art platform, and obtain the real-time interaction data of the user for the initial push result and the real-time monitoring data of the system for the initial push result; based on the real-time interaction data and the real-time monitoring data, the initial push result is adjusted in real time to obtain a real-time push result.

[0095] Furthermore, the information push model is constructed based on the following steps: constructing a historical user portrait based on historical user information data and artwork data, wherein the historical user portrait is used to represent the matching of historical users to artworks; establishing a model training set, a model test set and a model verification set based on the historical user information data, the artwork data and the historical user portrait; training and optimizing the initial model based on the model training set, the model test set and the model verification set to obtain the information push model.

[0096] Furthermore, the step of constructing a historical user portrait based on historical user information data and artwork data includes: determining the artwork matching similarity between historical users and the artwork feature similarity between artworks based on the historical user information data and artwork data; and constructing a historical user portrait based on the artwork matching similarity and the artwork feature similarity.

[0097] Other embodiments or specific implementations of the information push device of the present application can refer to the above-mentioned method embodiments and will not be described in detail here.

[0098] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0099] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0100] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0101] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An information push method, characterized in that: The method comprises the following steps: Performing data preprocessing on the user information data and the artwork data to obtain user features and artwork features; Inputting the user characteristics and the artwork characteristics into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information characteristics corresponding to artworks; The initial push result is pushed to the target user.

2. The information push method according to claim 1, characterized in that: The step of preprocessing the user information data and the artwork data to obtain user features and artwork features includes: Performing data conversion and data cleaning on the user information data and the artwork data to obtain cleaned user information data and cleaned artwork data; Feature extraction is performed on the cleaned user information data and the cleaned artwork data to obtain user features and artwork features.

3. The information push method according to claim 2, characterized in that: The step of extracting features from the cleaned user information data and the cleaned artwork data to obtain user features and artwork features includes: Extracting features from the cleaned user information data according to the number of clicks, browsing time and evaluation of each artwork by the user to obtain user features; Feature extraction is performed on the cleaned artwork data according to the content information and visual information of the artwork to obtain artwork features.

4. The information push method according to claim 1, characterized in that: The method further comprises: Acquire real-time interaction data of the user for the initial push result and real-time monitoring data of the system for the initial push result; The initial push result is adjusted in real time based on the real-time interaction data and the real-time monitoring data to obtain a real-time push result.

5. A method for constructing an information push model, characterized in that: The method comprises the following steps: Constructing a historical user portrait based on historical user information data and artwork data, wherein the historical user portrait is used to indicate the matching status of the historical user with the artwork; Establishing a model training set, a model testing set, and a model verification set according to the historical user information data, the artwork data, and the historical user portraits; The initial model is trained and optimized based on the model training set, the model test set and the model verification set to obtain an information push model.

6. The information push model construction method according to claim 5, characterized in that: The step of constructing a historical user portrait based on historical user information data and artwork data includes: Determine, based on the historical user information data and the artwork data, the artwork matching similarity between historical users and the artwork feature similarity between artworks; A historical user portrait is constructed based on the artwork matching similarity and the artwork feature similarity.

7. The information push model construction method according to claim 6, characterized in that: The step of constructing a historical user portrait based on the artwork matching similarity and the artwork feature similarity includes: Determine a first matching artwork according to a collaborative filtering algorithm and the artwork matching similarity, and determine a second matching artwork according to a content-based recommendation algorithm and the artwork feature similarity; A historical user portrait is constructed based on the first matching artwork and the second matching artwork.

8. An information push device, characterized in that: The information pushing device comprises: A data processing module, used for pre-processing user information data and artwork data to obtain user features and artwork features; A result generation module, used for inputting the user characteristics and the artwork characteristics into an information push model to obtain an initial push result, wherein the information push model is constructed based on multimodal data corresponding to historical users and information characteristics corresponding to artworks; The result push module is used to push the initial push result to the target user.

9. An information push device, characterized in that: The device comprises: a memory, a processor, and an information push program stored in the memory and executable on the processor, wherein the information push program is configured to implement the steps of the information push method according to any one of claims 1 to 4.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and an information push program is stored on the storage medium. When the information push program is executed by a processor, the steps of the information push method according to any one of claims 1 to 4 are implemented.

Images