Object recommendation method, computer terminal, storage medium and program product
By obtaining object evaluation data and images, determining the probability and association relationship of recommended, the problem of duplicate and low-quality products in the object recommendation system is solved, and the user experience is improved.
Patent Information
- Application Number
- CN202510350643.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-21
- Publication Date
- 2025-07-08
Smart Images
Figure CN120278787A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of image processing, and more particularly, to an object recommendation method, a computer terminal, a storage medium, and a program product. Background Art
[0002] In an e-commerce environment, an object recommendation system can accurately recommend products that a user may be interested in by analyzing user behavior, preferences, and product attributes. Thus, it can screen out the most relevant and attractive options for the user from a vast amount of product information, and is a key tool for enhancing the user experience and the platform's sales performance. However, to increase their exposure, some merchants may upload a large number of similar products, which often only have minor differences in the title or picture, while the actual product quality or characteristics are almost the same. As a result, when the object recommendation system recommends products to users, there may be a large number of duplicate and low-quality products, affecting the user experience.
[0003] Regarding the above problems, no effective solution has been proposed yet. Summary of the Invention
[0004] Embodiments of this application provide an object recommendation method, a computer terminal, a storage medium, and a program product to at least solve the technical problem in the related art that there are products with poor quality and duplicates among the products recommended to users.
[0005] According to one aspect of the embodiments of this application, an object recommendation method is provided, including: obtaining object evaluation data and object images of multiple objects to be recommended, where the object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; determining the recommended probability of the objects to be recommended based on the object evaluation data; constructing an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to represent the association relationship between the objects to be recommended; and selecting at least one object to be recommended from the multiple objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result.
[0006] According to one aspect of the embodiments of this application, an object recommendation method is further provided, including: in response to an input instruction acting on an operation interface, displaying object evaluation data and object images of multiple objects to be recommended on the operation interface, where the object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; and in response to an object recommendation instruction acting on the operation interface, displaying a target recommended object on the operation interface, where the target recommended object is selected from the multiple objects to be recommended based on the recommended probability and the object retrieval result of the objects to be recommended, the recommended probability is determined by the object evaluation data, the object retrieval result is constructed from the object images, and the object retrieval result is used to represent the association relationship between the objects to be recommended.
[0007] According to one aspect of the embodiments of the present application, there is also provided an object recommendation method, including: obtaining object evaluation data and object images of a plurality of objects to be recommended by calling a first interface, where the first interface includes a first parameter, and the parameter value of the first parameter includes the object evaluation data and the object images, and the object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; determining the recommended probability of the objects to be recommended based on the object evaluation data; constructing an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to represent the association relationship between the objects to be recommended; selecting at least one object to be recommended from the plurality of objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result; and outputting the target recommended object by calling a second interface, where the second interface includes a second parameter, and the parameter value of the second parameter includes the target recommended object.
[0008] According to one aspect of the embodiments of the present application, there is also provided an object recommendation device, including: a data acquisition module, configured to obtain object evaluation data and object images of a plurality of objects to be recommended, where the object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; a probability determination module, configured to determine the recommended probability of the objects to be recommended based on the object evaluation data; a result construction module, configured to construct an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to represent the association relationship between the objects to be recommended; and an object selection module, configured to select at least one object to be recommended from the plurality of objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result.
[0009] According to one aspect of the embodiments of the present application, there is also provided an object recommendation device, including: a first display module, configured to display object evaluation data and object images of a plurality of objects to be recommended on an operation interface in response to an input instruction acting on the operation interface, where the object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; and a second display module, configured to display a target recommended object on the operation interface in response to an object recommendation instruction acting on the operation interface, where the target recommended object is selected from the plurality of objects to be recommended based on the recommended probability and the object retrieval result of the objects to be recommended, the recommended probability is determined by the object evaluation data, the object retrieval result is constructed by the object images, and the object retrieval result is used to represent the association relationship between the objects to be recommended.
[0010] According to one aspect of the embodiments of the present application, there is also provided an object recommendation production, including: a first call module, configured to obtain object evaluation data and object images of multiple objects to be recommended by calling a first interface, where the first interface includes a first parameter, and the parameter value of the first parameter includes the object evaluation data and the object images, and the object evaluation data is used to characterize the data affecting the access popularity of the objects to be recommended; a recommended probability determination module, configured to determine the recommended probability of the objects to be recommended based on the object evaluation data; a retrieval result construction module, configured to construct an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to characterize the association relationship between the objects to be recommended; a recommended object selection module, configured to select at least one object to be recommended from the multiple objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result; a second call module, configured to output the target recommended object by calling a second interface, where the second interface includes a second parameter, and the parameter value of the second parameter includes the target recommended object.
[0011] According to another aspect of the embodiments of the present application, there is also provided a computer terminal, including: a memory storing an executable program; a processor configured to run the program, where when the program runs, it executes the methods in the various embodiments of the present application.
[0012] According to another aspect of the embodiments of the present application, there is also provided a computer-readable storage medium, where the computer-readable storage medium includes a stored executable program, and when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the methods in the various embodiments of the present application.
[0013] According to another aspect of the embodiments of the present application, there is also provided a computer program product, including a computer program, where when the computer program is executed by a processor, it implements the methods in the various embodiments of the present application.
[0014] According to another aspect of the embodiments of the present application, there is also provided a computer program product, including a non-volatile computer-readable storage medium storing a computer program, where when the computer program is executed by a processor, it implements the methods in the various embodiments of the present application.
[0015] According to another aspect of the embodiments of the present application, there is also provided a computer program, where when the computer program is executed by a processor, it implements the methods in the various embodiments of the present application.
[0016] In the embodiments of the present application, object evaluation data and object images of multiple objects to be recommended are obtained; based on the object evaluation data, the recommended probability of the objects to be recommended is determined; based on the object images, an object retrieval result of the objects to be recommended is constructed; based on the recommended probability and the object retrieval result, at least one object to be recommended is selected from the multiple objects to be recommended as the target recommended object. By accurately determining the association relationship and access popularity among different objects to be recommended based on the object images and object evaluation data of the objects to be recommended, the object recommendation system can remove the repeatedly appearing objects in the objects to be recommended according to the determined association relationship, and select objects with higher quality from the objects to be recommended according to the access popularity, so as to ensure the rationality of the target recommended objects selected from the multiple objects to be recommended, improve the user experience when browsing the target recommended objects, and further solve the technical problem that there are poor-quality and repeated products among the products recommended to users in the related art.
[0017] It is easy to notice that the above general description and the following detailed description are only for exemplifying and explaining the present application, and do not constitute a limitation to the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation to the present application. In the drawings:
[0019] Figure 1 is a schematic diagram of an application scenario of an object recommendation method shown according to an embodiment of the present application;
[0020] Figure 2 is a flowchart of an object recommendation method shown according to an embodiment of the present application;
[0021] Figure 3 is a schematic diagram of an object recommendation process shown according to an embodiment of the present application;
[0022] Figure 4 is a flowchart of another object recommendation method shown according to an embodiment of the present application;
[0023] Figure 5 is a flowchart of another object recommendation method shown according to an embodiment of the present application;
[0024] Figure 6 is a structural block diagram of an object recommendation device shown according to an embodiment of the present application;
[0025] Figure 7 is a structural block diagram of another object recommendation device shown according to an embodiment of the present application;
[0026] Figure 8 It is a structural block diagram of another object recommendation device shown according to an embodiment of the present application;
[0027] Figure 9 It is a structural block diagram of a computing device shown according to an embodiment of the present application;
[0028] Figure 10 It is a structural block diagram of an electronic device shown according to an embodiment of the present application. Detailed implementation manners
[0029] In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily need to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.
[0031] It should be noted that when the processing model used in the present application is actually applied, the pre-trained model can be fine-tuned with a small number of samples so that the processing model can be applied to different tasks. For example, it can be widely applied to fields such as Natural Language Processing (NLP), computer vision, and speech processing. Specifically, it can be applied to tasks in the field of computer vision such as Visual Question Answering (VQA), Image Caption (IC), and image generation, and can also be widely applied to tasks in the field of natural language processing such as text-based sentiment classification, text summary generation, and machine translation. Therefore, the main application scenarios of the processing model include but are not limited to digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc.
[0032] According to an embodiment of the present application, an object recommendation method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.
[0033] Considering the limited computing resources of mobile terminals, the above method provided by the embodiments of the present application can be applied to Figure 1 the application scenarios shown below, but not limited thereto. Figure 1 FIG. is a schematic diagram of an application scenario of an object recommendation method according to an embodiment of the present application. In Figure 1 the application scenario shown below, the processing model is deployed in server 10. Server 10 can be connected to one or more client devices 20 through a local area network connection, a wide area network connection, an Internet connection, or other types of data networks. Here, the client devices 20 may include but are not limited to: smart phones, tablet computers, laptop computers, palmtop computers, personal computers, smart home devices, vehicle-mounted devices, etc. The client device 20 can interact with the user through a graphical user interface to call the processing model, and further implement the method provided by the embodiments of the present application.
[0034] In the embodiments of the present application, the system composed of the client device and the server can execute the following steps: The client device executes to obtain object evaluation data and object images of multiple objects to be recommended. The server executes to determine the recommended probability of the object to be recommended based on the object evaluation data; construct an object retrieval result of the object to be recommended based on the object image; select a target recommended object from multiple objects to be recommended based on the recommended probability and the object retrieval result.
[0035] It should be noted that with the rapid development of high-performance computing units, in other application scenarios, the above method provided by the embodiments of the present application can also be applied to a model all-in-one machine. In an optional embodiment, multiple models are built in the model all-in-one machine. The user can select and adjust one model according to needs to obtain his own model. Thus, the high-performance computing unit built in the model all-in-one machine can directly call the adjusted model to execute the above method provided by the embodiments of the present application. In another optional embodiment, a trained model is built in the model all-in-one machine. Thus, the high-performance computing unit built in the model all-in-one machine can directly call the model to execute the above method provided by the embodiments of the present application.
[0036] Further, when the user needs to train their own model, they can also upload their own dataset through the client. The dataset is sent from the client to the server, enabling the server to adjust the pre-trained model with this dataset to obtain the user's own model, which is then deployed to the production environment. To facilitate the user's model adjustment requirements, the server can provide complete adjustment tools, development frameworks, and processes, supporting multiple adjustment strategies, so that the adjusted model can better adapt to different field applications and achieve high customization.
[0037] Under the above operating environment, the present application provides an object recommendation method as Figure 2 shown. Figure 2 It is a flowchart of an object recommendation method shown according to an embodiment of the present application. As Figure 2 shown, the method may include the following steps:
[0038] Step S202, obtain object evaluation data and object images of multiple objects to be recommended.
[0039] Among them, the object evaluation data is used to characterize the data that affects the access popularity of the objects to be recommended.
[0040] The above objects to be recommended may refer to objects with multimodal data, including but not limited to: products in a shopping platform, articles in a media platform, artworks equipped with descriptive information in the art market, digital collections, etc. For the convenience of understanding, the following takes products as an example of the above objects to be recommended for illustration. The above object evaluation data may be data related to the objects to be recommended and capable of evaluating the quality of the objects to be recommended, including but not limited to: historical sales data, user feedback, product information, market trends, etc. Through the object evaluation data, the object recommendation system can judge the access popularity of the objects to be recommended, so as to determine whether the objects to be recommended can be pushed to the client for the user to browse. The access popularity of the above objects to be recommended may refer to the popularity of the objects to be recommended.
[0041] In an alternative solution of this embodiment, considering that shopping platforms usually allocate more traffic to newly listed high-quality products of merchants to help these products enter the market faster, thereby motivating merchants to list more high-quality and distinctive products, promoting healthy competition among merchants, and then improving the overall product quality and service level. However, for newly listed products, whether they are truly high-quality products can usually only be determined by the product information provided by the merchants. If a merchant simply modifies the product information of an old product, for example, only modifies the picture or copywriting of the old product and deliberately indicates that this old product belongs to a new product with high quality when listing the product, the object recommendation system may misidentify this old product as a new product that requires traffic support, resulting in a large number of duplicate old products in the products recommended to users and unable to guarantee the quality of the recommended old products, thereby affecting the shopping experience of users. Therefore, in order to avoid pushing duplicate and low-quality objects to users, the object recommendation system can first determine multiple objects to be recommended that need to be recommended currently, such as multiple products newly listed or newly released by merchants through the shopping platform, and obtain the object evaluation data and object images of these multiple objects to be recommended, so as to judge whether these objects to be recommended are duplicate objects through the object images, and evaluate the access popularity of different objects to be recommended through the object evaluation data, so as to judge whether the object quality of different objects to be recommended is high and whether it is worth allocating more traffic to push the corresponding objects to be recommended to users.
[0042] For example, if the obtained object image shows that product A is different from other products among the products listed by the merchant or the objects to be recommended, the object recommendation system can determine that product A is a new product and requires certain traffic support; conversely, if the obtained object image shows that product A is the same as other products among the products listed by the merchant or the objects to be recommended, the object recommendation system can refuse to provide more traffic for product A to avoid reducing the shopping experience of users due to repeated browsing of the same product. Correspondingly, if the obtained object evaluation data shows that the user feedback of product A is good and the access popularity is high, it indicates that the quality of product A may be good and the possibility of users purchasing product A is relatively large. The corresponding object recommendation system can then allocate more push traffic for product A to push product A to users; conversely, if the obtained object evaluation data shows that the user feedback of product A is poor and the access popularity is low, it indicates that the quality of product A may be poor and the possibility of users purchasing product A is relatively low. The corresponding object recommendation system can then not push product A to users.
[0043] Step S204, based on the object evaluation data, determine the recommended probability of the object to be recommended.
[0044] In an alternative solution of this embodiment, considering that there are many types of object evaluation data related to the access popularity of the object to be recommended, it is usually difficult to reflect the recommended value of the corresponding object to be recommended based on only one type of data. Therefore, in order to be able to push objects of higher quality to users, the object recommendation system can determine the recommended probability of the corresponding object to be recommended according to the obtained object evaluation data, so as to intuitively show the value of the corresponding object to be recommended through the recommended probability, that is, whether it is worth allocating more support traffic to this object to be recommended, while improving the quality of goods and service levels provided by merchants and ensuring the shopping experience of users.
[0045] For example, the object recommendation system can first obtain a large amount of product data related to different products, such as sales records of different products, user feedback on products, data on the development trend of the product market, etc., and build a dynamic product recommendation probability prediction model based on these data. The object recommendation system can assign different prediction weights to different types of product data according to the importance of different types of product data in the product recommendation probability prediction model. After building the product recommendation probability prediction model, the object recommendation system can input the real-time obtained object evaluation data into the product recommendation probability prediction model to use the model to determine the recommended probability of different objects to be recommended in real time, so as to ensure the accuracy of the determined recommended probability.
[0046] Step S206: Based on the object image, construct the object retrieval result of the object to be recommended.
[0047] The object retrieval result is used to represent the association relationship between the objects to be recommended.
[0048] In an alternative solution of this embodiment, in order to accurately judge the association relationship between different objects to be recommended, such as whether different newly listed products are the same product, whether they belong to the same merchant, whether they are produced by the same manufacturer, etc., the object recommendation system can identify the object images corresponding to different objects to be recommended, extract image content that can reflect a series of information such as the type, object identifier, merchant identifier, manufacturer identifier, etc. of the corresponding object to be recommended from the object images, and then compare this information, so that the object recommendation system can accurately construct the object retrieval result of the above-mentioned object to be recommended on the basis of the object image, that is, determine the association relationship between different objects to be recommended.
[0049] Step S208: Based on the recommended probability and the object retrieval result, select at least one object to be recommended from multiple objects to be recommended as the target recommended object.
[0050] In an alternative solution of this embodiment, after determining the recommended probabilities of the objects to be recommended and the object retrieval results, the object recommendation system can, based on the recommended probabilities and the object retrieval results, select at least one high-quality target recommended object from the above-mentioned multiple objects to be recommended to ensure the user experience when browsing the target recommended objects. For example, the object recommendation system can first select different objects from the multiple objects to be recommended according to the determined association relationships of the objects to be recommended, so as to avoid a large number of duplicate objects in the finally recommended target recommended objects to the user, which may affect the user's browsing experience. After deduplicating the multiple objects to be recommended, the object recommendation system can further select objects with higher recommended probabilities as target recommended objects according to the recommended probabilities of the deduplicated objects, so as to ensure the quality of the selected target recommended objects and avoid the situation that the quality of the target recommended objects pushed to the user is low, resulting in the impact on the user experience.
[0051] In the embodiment of the present application, the method includes obtaining the object evaluation data and object images of multiple objects to be recommended; determining the recommended probabilities of the objects to be recommended based on the object evaluation data; constructing the object retrieval results of the objects to be recommended based on the object images; and selecting at least one object to be recommended from the multiple objects to be recommended as the target recommended object based on the recommended probabilities and the object retrieval results. By accurately determining the association relationships and access hotness among different objects to be recommended based on the object images and object evaluation data of the objects to be recommended, the object recommendation system can remove the repeatedly appearing objects in the objects to be recommended according to the determined association relationships and select objects with higher quality from the objects to be recommended according to the access hotness, so as to ensure the rationality of the target recommended objects selected from the multiple objects to be recommended, improve the user experience when browsing the target recommended objects, and further solve the technical problem in the related art that there are poor-quality and duplicate products among the products recommended to the user.
[0052] In the embodiment of the present application, the object retrieval results include: the same-image retrieval results and the same-source retrieval results. The same-image retrieval results are used to represent whether the types of the objects to be recommended are the same, and the same-source retrieval results are used to represent whether the object sources of the objects to be recommended are the same; constructing the object retrieval results of the objects to be recommended based on the object images includes: constructing the same-image retrieval results based on at least one visual element of the object images; constructing the same-source retrieval results based on at least one image content of the object images, where the data types of different image contents are different.
[0053] In an alternative solution of this embodiment, in order to accurately represent the association relationship between different objects to be recommended, the constructed object retrieval results may at least include same-graph retrieval results and same-source retrieval results. Correspondingly, the object recommendation system can determine whether the types of different objects to be recommended are the same through the same-graph retrieval results, and determine the object sources between different objects to be recommended through the same-source retrieval results, that is, whether the merchants providing the corresponding objects to be recommended are the same. Through these two retrieval results, the object recommendation system can more accurately determine whether there are duplicate objects among the objects to be recommended. Correspondingly, the object recommendation system can judge whether the types of the corresponding objects to be recommended are the same through the visual effects presented by different object images, that is, at least one visual element of the object image, such as elements like color, texture, and shape, to construct the above-mentioned same-graph retrieval results. At the same time, it can judge whether the object sources between the corresponding objects to be recommended are the same through at least one image content of different data types included in different object images, such as multi-modal data like text, pictures, and videos, to construct the above-mentioned same-source retrieval results.
[0054] In the embodiment of the present application, constructing a same-graph retrieval result based on at least one visual element of an object image includes: extracting features from at least one visual element to obtain at least one visual feature of the object image; matching multiple objects to be recommended based on at least one visual feature corresponding to different object images to obtain a first similarity between different objects to be recommended; constructing a same-graph retrieval result based on the first similarity, where the same-graph retrieval result includes: at least one image pair, and at least one image pair includes: a reference image and a similar image, and the reference image is used to represent the object image corresponding to the object to be recommended for which the same-graph retrieval result needs to be determined, and the first similarity between the reference image and the similar image is greater than a first preset value.
[0055] In an alternative solution of this embodiment, when constructing the same-image retrieval result, the object recommendation system may first extract features from at least one visual element included in the object image to obtain the corresponding visual features, and then match multiple objects to be recommended according to the at least one visual feature corresponding to different object images to determine the first similarity between different objects to be recommended. Through the first similarity, the object recommendation system can determine whether the corresponding objects to be recommended have the same object type, so as to accurately construct the above-mentioned same-image retrieval result. For easy viewing, the above-mentioned same-image retrieval result may represent the first similarity between the corresponding objects to be recommended in the form of image pairs, that is, the same-image retrieval result may include at least one image pair, and the at least one image pair may include: a reference image and a similar image, where the reference image may refer to the object image corresponding to the object to be recommended for which the same-image retrieval result needs to be determined, and the first similarity between the reference image and the similar image is greater than the first preset value, that is, the reference image and the similar image can be regarded as a pair of images of the same type.
[0056] In the embodiment of the present application, based on at least one image content of the object image, a homologous retrieval result is constructed, including: extracting features from at least one image content to obtain at least one content feature of the object image; matching at least one content feature corresponding to the reference image in the same-image retrieval result with at least one content feature corresponding to other images in the object image except the reference image to obtain the second similarity between the reference image and the other images; constructing a homologous retrieval result based on the second similarity, where the homologous retrieval result includes: at least one object source data pair, and the at least one object source data pair includes: the first object source corresponding to the reference image, and the second object source of the other image, and the second similarity between the first object source and the second object source is greater than the second preset value.
[0057] In an alternative solution of this embodiment, when constructing the homologous retrieval result, the object recommendation system may first extract features from at least one image content included in the object image to obtain at least one content feature of the object image, such as text features, picture features, etc. Then, it matches at least one content feature corresponding to the reference image in the same-image retrieval result with at least one content feature corresponding to other images except the reference image in the object image to determine the second similarity between the reference image and other images. The object recommendation system can use the second similarity to determine whether the corresponding objects to be recommended have the same object source, so as to accurately construct the above-mentioned homologous retrieval result. Corresponding to the same-image retrieval result, the homologous retrieval result may include at least one object source data pair. The at least one object source data pair includes the first object source corresponding to the above-mentioned reference image and the second object source of other images, where the second similarity between the first object source and the second object source is greater than a second preset value, that is, the first object source and the second object source can be regarded as the same object source.
[0058] In the embodiment of the present application, determining the recommended probability of an object to be recommended based on object evaluation data includes: evaluating the object to be recommended from at least one evaluation dimension based on the object evaluation data to obtain at least one evaluation score corresponding to the object to be recommended; performing a weighted process on the at least one evaluation score to obtain an object evaluation score corresponding to the object to be recommended; and performing a normalization process on the object evaluation scores corresponding to different objects to be recommended to obtain the recommended probability.
[0059] In an alternative solution of this embodiment, in order to accurately determine the recommended probability of an object to be recommended, in addition to using the aforementioned commodity recommendation probability prediction model, the object recommendation system can also directly evaluate the quality of the object to be recommended from at least one evaluation dimension according to the obtained object evaluation data to obtain at least one evaluation score corresponding to the object to be recommended, and then perform a weighted process on the at least one evaluation score to obtain an object evaluation score corresponding to the object to be recommended. Finally, by performing a normalization process on the object evaluation scores of different objects to be recommended, the object recommendation system can quickly and reasonably determine the corresponding recommended probability. Among them, the evaluation dimensions used by the object recommendation system can correspond to the data types of the obtained object evaluation data. For example, the object recommendation system can evaluate the quality of the object to be recommended through dimensions such as market development trends and user feedback to obtain the evaluation scores for the corresponding dimensions.
[0060] In an embodiment of the present application, based on the recommended probability and the object retrieval result, at least one object to be recommended is selected from multiple objects to be recommended as the target recommended object, including: based on the object retrieval result, constructing a first data representation of the object to be recommended, where the first data representation includes: the first object identifier of the first object to be recommended and the second object identifier of the second object to be recommended, the first object to be recommended is any object to be recommended, and the types and object sources between the first object to be recommended and the second object to be recommended are the same; updating the first data representation based on the first recommended probability corresponding to the second object to be recommended to obtain a second data representation; and determining the object to be recommended corresponding to the second object identifier in the second data representation as the target recommended object.
[0061] The above first data representation, second data representation, etc. may refer to data forms that can intuitively reflect the association relationships between different objects to be recommended, and may include but are not limited to: data tables, knowledge graphs, tree diagrams, etc.
[0062] In an alternative solution of this embodiment, when selecting the target recommended object, in order to ensure the quality of the selected target recommended object and avoid duplicate selection, the object recommendation system may first construct the above first data representation according to the aforementioned object retrieval result to clearly show the association relationships between different objects to be recommended by using the first data representation, such as whether they are the same object. Correspondingly, the first data representation may at least include: the first object identifier of the first object to be recommended and the second object identifier of the second object to be recommended, the first object to be recommended is any object to be recommended, and the types and object sources between the first object to be recommended and the second object to be recommended are the same. Considering that in the first data representation, the first object to be recommended and the corresponding second object to be recommended have the same object type and object source, this means that the object recommendation system only needs to deduplicate the second object to be recommended based on the first object to be recommended to achieve the purpose of avoiding duplicate objects. And in order to ensure the quality of the deduplicated objects, the object recommendation system may further introduce the recommended probabilities of different objects to be recommended on the basis of the first data representation, that is, directly consider the first recommended probability corresponding to the second object to be recommended, and update the first data representation by using the first recommended probability to retain the objects with higher recommended probabilities among the second objects to be recommended, so as to construct the above second data representation and implement the deduplication operation on the second object to be recommended. Finally, the object recommendation system may directly determine the object to be recommended corresponding to the second object identifier in the second data representation as the target recommended object to be pushed to the user, thereby ensuring the rationality of the determined target recommended object.
[0063] In the embodiment of the present application, based on the first recommended probability corresponding to the second object to be recommended, the first data representation is updated to obtain a second data representation, including: constructing a third data representation based on the first data representation and the first recommended probability, where the third data representation includes: a first object identifier, a second object identifier, and the first recommended probability; determining, from the third data representation, third object identifiers corresponding to different first object identifiers, and second recommended probabilities corresponding to the third object identifiers, where the third object identifiers belong to the second object identifiers, and the second recommended probabilities are input into the first recommended probability; selecting a target object identifier from the third object identifiers based on the second recommended probability, where the second recommended probability corresponding to the target object identifier is greater than the second recommended probabilities corresponding to other object identifiers except the target object identifier among the third object identifiers; constructing the second data representation based on the first object identifier, the target object identifier corresponding to the first object identifier, and the second recommended probability.
[0064] In an alternative solution of this embodiment, to facilitate understanding of the process of updating the first data representation using the first recommended probability, the object recommendation system may first add the first recommended probability to the first data representation according to the second object identifier of the corresponding second object to be recommended, so as to construct a corresponding third data representation based on the first data representation and the first recommended probability. The third data representation may include: the first object identifier, the second object identifier, and the first recommended probability. After constructing the third data representation, the object recommendation system may determine, from the third data representation, third object identifiers corresponding to different first object identifiers, and second recommended probabilities corresponding to the third object identifiers, where the third object identifiers belong to the second object identifiers, and the second recommended probabilities are input into the first recommended probability. After determining the second recommended probability, the object recommendation system may further select a target object identifier from the third object identifiers according to the second recommended probability. Correspondingly, the second recommended probability corresponding to the target object identifier is greater than the second recommended probabilities corresponding to other object identifiers except the target object identifier among the third object identifiers. Finally, according to the first object identifier, the target object identifier corresponding to the first object identifier, and the second recommended probability, the object recommendation system may construct the above-mentioned second data representation.
[0065] In the embodiment of the present application, the above method further includes: obtaining object browsing data of a target client; constructing object recommendation parameters of a target recommended object based on the object browsing data; determining a traffic allocation weight of the target recommended object based on the recommended probability corresponding to the target recommended object; and pushing the target recommended object to the target client based on the traffic allocation weight and the object recommendation parameters.
[0066] In an alternative solution of this embodiment, when pushing a target recommended object, the picture recommendation system may first obtain the object browsing data of the target client used by the user, such as historical browsing records, user portraits, and other information, and then construct the object recommendation parameters of the target recommended object on the basis of the object browsing data, such as the display position and display times of the target recommended picture, etc. At the same time, according to the recommended probability corresponding to the target recommended object, the traffic allocation weight of the target recommended object is determined. The greater the recommended probability, the greater the corresponding traffic allocation weight. Finally, after determining the object recommendation parameters and the traffic allocation weight, the object recommendation system can push the target recommended object to the target client according to the traffic allocation weight and the object recommendation parameters to improve the user experience.
[0067] For ease of understanding, Figure 3 is a schematic diagram of an object recommendation process shown according to an embodiment of the present application. As Figure 3 shown, in order to avoid repeatedly pushing low-quality products to users, the object recommendation system may first obtain the new products currently on the shelves by merchants, as well as the product evaluation data and product images corresponding to these new products. Then, the object recommendation system can extract features from the product evaluation data to obtain the corresponding product features, and input these data features into a pre-trained evaluation model, such as the aforementioned product recommendation probability prediction model, to evaluate the quality of the new products from multiple different dimensions to obtain the corresponding recommended probability. At the same time, the object recommendation system can perform same-image retrieval on the new products based on the visual elements of the obtained product images to obtain the corresponding same-image retrieval results, and then extract multi-modal data from the product images to obtain at least one multi-modal data included in the product images. Based on the reference images included in the same-image retrieval results and the corresponding at least one multi-modal data, homologous retrieval is performed on the new products to obtain the corresponding homologous retrieval results. Finally, by combining the recommended probability of the new products evaluated above, as well as the same-image retrieval results and homologous retrieval results of the new products, the object recommendation system can perform deduplication processing on the new products to screen out non-repetitive and high-quality target products, and allocate corresponding support traffic to the target products to reasonably push the target products to users.
[0068] Among them, the data representation corresponding to the recommended probability obtained by evaluating the product can be:
[0069] Table 1
[0070] item score a 0.9 b 0.8 c 0.7 d 0.6 e 0.5 f 0.4
[0071] Among them, item represents the product identifier of the new product, which may include a, b, c, d, e, f. Score may refer to the corresponding evaluation score or recommended probability.
[0072] The first data representation constructed based on the same-image retrieval results and the homologous retrieval results can be:
[0073] Table 2
[0074] source_item sim_item a b a c b c c a d d e e e f f e
[0075] Among them, source_item can refer to the object identifier corresponding to the reference image, and sim_item can refer to the object identifier corresponding to the similar image. As can be seen from Table 2, a, b, and c can be regarded as duplicate products, d can be regarded as a non-duplicate product, and e and f can be regarded as duplicate products.
[0076] The third data representation constructed based on the first data representation and the second recommended probability can be:
[0077] Table 3
[0078]
[0079]
[0080] The second data representation obtained by screening the third data representation can be:
[0081] Table 4
[0082] source_item sim_item sim_item-score a a 0.9 b a 0.9 c a 0.9 d d 0.6 e e 0.5 f e 0.5
[0083] Through the second data representation, the object recommendation system can determine that the representations of the products that currently need traffic support can be a, d, and e.
[0084] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. And the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or reject.
[0085] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.
[0086] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions for causing a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present application.
[0087] According to an embodiment of the present application, there is also provided another object recommendation method. Figure 4 It is a flowchart of another object recommendation method shown according to an embodiment of the present application, as Figure 4 shown. The method may include the following steps:
[0088] Step S402, in response to an input instruction acting on the operation interface, display object evaluation data and object images of multiple objects to be recommended on the operation interface.
[0089] Among them, the object evaluation data is used to characterize the data affecting the access popularity of the objects to be recommended.
[0090] Step S404, in response to an object recommendation instruction acting on the operation interface, display a target recommended object on the operation interface.
[0091] Among them, the target recommended object is selected from multiple objects to be recommended according to the recommended probability of the objects to be recommended and the object retrieval result. The recommended probability is determined by the object evaluation data, and the object retrieval result is constructed from the object images. The object retrieval result is used to characterize the association relationship between the objects to be recommended.
[0092] In an alternative solution of this embodiment, when receiving an input instruction acting on the operation interface, the object recommendation system can first display the object evaluation data and object images of multiple objects to be recommended on the operation interface to reflect the access popularity of the objects to be recommended through the object evaluation data. When receiving an object recommendation instruction acting on the operation interface, the object recommendation system can first determine the recommended probability of the objects to be recommended according to the object evaluation data, and construct the object retrieval result of the objects to be recommended according to the object images to reflect the association relationship between the objects to be recommended through the object retrieval result, and then select at least one object to be recommended from multiple objects to be recommended according to the recommended probability and the object retrieval result as the target recommended object, and display the target recommended object in the operation interface for the user to view conveniently.
[0093] It should be noted that the preferred implementation solutions involved in the above embodiments of the present application are the same as the solutions, application scenarios, and implementation processes provided by the embodiments, but are not limited to the solutions provided by the embodiments.
[0094] According to an embodiment of the present application, there is also provided another object recommendation method. Figure 5 It is a flowchart of another object recommendation method shown according to an embodiment of the present application, as Figure 5 shown. The method may include the following steps:
[0095] Step S502, obtain the object evaluation data and object images of multiple objects to be recommended by calling a first interface.
[0096] Among them, the first interface includes a first parameter, and the parameter value of the first parameter includes object evaluation data and object images. The object evaluation data is used to represent the data affecting the access popularity of the objects to be recommended.
[0097] Step S504, determine the recommended probability of the objects to be recommended based on the object evaluation data.
[0098] Step S506, construct an object retrieval result of the objects to be recommended based on the object images.
[0099] Among them, the object retrieval result is used to represent the association relationship between the objects to be recommended.
[0100] Step S508, select at least one object to be recommended from the multiple objects to be recommended as the target recommended object based on the recommended probability and the object retrieval result.
[0101] Step S510, output the target recommended object by calling a second interface.
[0102] Among them, the second interface includes a second parameter, and the parameter value of the second parameter includes the target recommended object.
[0103] In an alternative solution of this embodiment, when recommending objects, the object recommendation system may first call the first interface to obtain the object evaluation data and object images of multiple objects to be recommended, so as to reflect the access popularity of the objects to be recommended through the object evaluation data, determine the recommended probability of the objects to be recommended according to the object evaluation data, and at the same time construct an object retrieval result of the objects to be recommended according to the object images, so as to reflect the association relationship between the objects to be recommended through the object retrieval result. Then, according to the recommended probability and the object retrieval result, select at least one object to be recommended from the multiple objects to be recommended as the target recommended object, and finally call the second interface to output the target recommended object for the convenience of the user to view.
[0104] According to an embodiment of the present application, there is also provided an object recommendation device for implementing the above object recommendation method.Figure 6 is a structural block diagram of an object recommendation device shown according to an embodiment of the present application. As Figure 6 shown, the device includes: a data acquisition module 602, a probability determination module 604, a result construction module 606, and an object selection module 608.
[0105] Among them, the data acquisition module 602 is used to acquire object evaluation data and object images of multiple objects to be recommended. Among them, the object evaluation data is used to characterize the data affecting the access popularity of the objects to be recommended; the probability determination module 604 is used to determine the recommended probability of the objects to be recommended based on the object evaluation data; the result construction module 606 is used to construct an object retrieval result of the objects to be recommended based on the object images. Among them, the object retrieval result is used to characterize the association relationship between the objects to be recommended; the object selection module 608 is used to select at least one object to be recommended from multiple objects to be recommended based on the recommended probability and the object retrieval result as the target recommended object.
[0106] In the embodiment of the present application, the object retrieval result includes: a same-image retrieval result and a same-source retrieval result. The same-image retrieval result is used to characterize whether the types of the objects to be recommended are the same, and the same-source retrieval result is used to characterize whether the object sources of the objects to be recommended are the same; the result construction module 606 includes: a first construction unit, which is used to construct a same-image retrieval result based on at least one visual element of the object image; a second construction unit, which is used to construct a same-source retrieval result based on at least one image content of the object image, where the data types of different image contents are different.
[0107] In the embodiment of the present application, the first construction unit is further used to: extract features from at least one visual element to obtain at least one visual feature of the object image; match multiple objects to be recommended based on at least one visual feature corresponding to different object images to obtain a first similarity between different objects to be recommended; construct a same-image retrieval result based on the first similarity, where the same-image retrieval result includes: at least one image pair, and at least one image pair includes: a reference image and a similar image, and the reference image is used to characterize the object image corresponding to the object to be recommended for which the same-image retrieval result needs to be determined, and the first similarity between the reference image and the similar image is greater than a first preset value.
[0108] In an embodiment of the present application, the second construction unit is further configured to: extract features from at least one image content to obtain at least one content feature of the object image; match at least one content feature corresponding to the reference image in the same-image retrieval result with at least one content feature corresponding to other images in the object image except the reference image to obtain a second similarity between the reference image and the other images; and construct a homologous retrieval result based on the second similarity, where the homologous retrieval result includes: at least one pair of object source data, and at least one pair of object source data includes: a first object source corresponding to the reference image, and a second object source of the other image, and the second similarity between the first object source and the second object source is greater than a second preset value.
[0109] In an embodiment of the present application, the probability determination module 604 includes: an object evaluation unit, configured to evaluate a to-be-recommended object from at least one evaluation dimension based on object evaluation data to obtain at least one evaluation score corresponding to the to-be-recommended object; a score weighting unit, configured to perform a weighting process on the at least one evaluation score to obtain an object evaluation score corresponding to the to-be-recommended object; and a probability determination unit, configured to perform a normalization process on the object evaluation scores corresponding to different to-be-recommended objects to obtain a recommended probability.
[0110] In an embodiment of the present application, the object selection module 608 includes: a data representation construction unit, configured to construct a first data representation of a to-be-recommended object based on an object retrieval result, where the first data representation includes: a first object identifier of a first to-be-recommended object and a second object identifier of a second to-be-recommended object, the first to-be-recommended object is any to-be-recommended object, and the types and object sources between the first to-be-recommended object and the second to-be-recommended object are the same; a data representation update unit, configured to update the first data representation based on a first recommended probability corresponding to the second to-be-recommended object to obtain a second data representation; and a recommended object determination unit, configured to determine the to-be-recommended object corresponding to the second object identifier in the second data representation as a target recommended object.
[0111] In an embodiment of the present application, the data representation update unit is further configured to: construct a third data representation based on the first data representation and the first recommended probability, where the third data representation includes: the first object identifier, the second object identifier, and the first recommended probability; determine, from the third data representation, a third object identifier corresponding to different first object identifiers, and a second recommended probability corresponding to the third object identifier, where the third object identifier belongs to the second object identifier, and the second recommended probability is input to the first recommended probability; select a target object identifier from the third object identifiers based on the second recommended probability, where the second recommended probability corresponding to the target object identifier is greater than the second recommended probabilities corresponding to other object identifiers except the target object identifier in the third object identifiers; and construct the second data representation based on the first object identifier, the target object identifier corresponding to the first object identifier, and the second recommended probability.
[0112] In an embodiment of the present application, the above device further includes: a browsing data acquisition module, configured to acquire object browsing data of a target client; a recommendation parameter construction module, configured to construct object recommendation parameters of a target recommended object based on the object browsing data; a traffic weight determination module, configured to determine a traffic allocation weight of the target recommended object based on the recommended probability corresponding to the target recommended object; and a recommended object push module, configured to push the target recommended object to the target client based on the traffic allocation weight and the object recommendation parameters.
[0113] It should be noted here that the above data acquisition module 602, probability determination module 604, result construction module 606, and object selection module 608 correspond to steps S202 to S208. The instances and application scenarios implemented by the three modules and the corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules or units may be hardware components or software components stored in a memory and processed by one or more processors, and the above modules may also be part of the device and can run in the server 10 provided in the above embodiments.
[0114] According to an embodiment of the present application, there is also provided an object recommendation device for implementing the above object recommendation method. Figure 7 is a structural block diagram of another object recommendation device shown according to an embodiment of the present application, as Figure 7 shown, the device includes: a first display module 702 and a second display module 704.
[0115] Among them, the first display module 702 is configured to respond to an input instruction acting on the operation interface and display object evaluation data and object images of a plurality of objects to be recommended on the operation interface, where the object evaluation data is used to characterize data affecting the access popularity of the objects to be recommended; the second display module 704 is configured to respond to an object recommendation instruction acting on the operation interface and display a target recommended object on the operation interface, where the target recommended object is selected from a plurality of objects to be recommended according to the recommended probability of the objects to be recommended and the object retrieval result, the recommended probability is determined by the object evaluation data, the object retrieval result is constructed from the object images, and the object retrieval result is used to characterize the association relationship between the objects to be recommended.
[0116] It should be noted here that the above first display module 702 and second display module 704 correspond to steps S602 to S604. The instances and application scenarios implemented by the two modules and the corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules or units may be hardware components or software components stored in a memory and processed by one or more processors, and the above modules may also be part of the device and can run in the server 10 provided in the above embodiments.
[0117] According to an embodiment of the present application, there is also provided an object recommendation device for implementing the above object recommendation method. Figure 8 It is a structural block diagram of another object recommendation device shown according to an embodiment of the present application. As Figure 8 shown, the device includes: a first call module 802, a recommended probability determination module 804, a retrieval result construction module 806, a recommended object selection module 808, and a second call module 810.
[0118] Among them, the first call module 802 is used to obtain object evaluation data and object images of multiple objects to be recommended by calling a first interface. Among them, the first interface includes a first parameter, and the parameter value of the first parameter includes object evaluation data and object images. The object evaluation data is used to represent data affecting the access popularity of the objects to be recommended; the recommended probability determination module 804 is used to determine the recommended probability of the objects to be recommended based on the object evaluation data; the retrieval result construction module 806 is used to construct an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to represent the association relationship between the objects to be recommended; the recommended object selection module 808 is used to select at least one object to be recommended from multiple objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result; the second call module 810 is used to output the target recommended object by calling a second interface. Among them, the second interface includes a second parameter, and the parameter value of the second parameter includes the target recommended object.
[0119] It should be noted here that the above first call module 802, recommended probability determination module 804, retrieval result construction module 806, recommended object selection module 808, and second call module 810 correspond to steps S502 to S510. The instances and application scenarios implemented by the five modules and the corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors. The above modules can also be part of a device and can run in the server 10 provided in the above embodiments.
[0120] It should be noted that the preferred implementation schemes involved in the above embodiments of the present application are the same as the schemes, application scenarios, and implementation processes provided in the above embodiments, but are not limited to the schemes provided in the above embodiments.
[0121] An embodiment of the present application can provide a computing device. Figure 9 It is a structural block diagram of a computing device shown according to an embodiment of the present application. As shown in the figure, the computing device 900 may include: one or more (only one is shown in the figure) processors 902, a memory 904, a storage controller, and a peripheral interface.
[0122] The above-mentioned computing device can be understood as an integrated intelligent terminal, including but not limited to servers, desktop computers, PCs (Personal Computers), model all-in-ones, etc. Moreover, the model described in the above embodiments of the present application can be preset in the computing device.
[0123] Specifically, the computing device can preset various types of models, including but not limited to models in the fields of natural language processing, visual processing, speech processing, code processing, multi-modal task processing, etc., so as to provide diverse model selections. In different product forms, the computing device can support one or more model usage methods, including but not limited to model training, model invocation, model fine-tuning, model deployment, model inference and application, etc. In some product forms, the computing device also supports model management, including but not limited to multi-type model management (supporting the management of various types of models such as discriminative and generative models), model version control (supporting the control of different model versions), model evaluation (evaluating the performance and effect of the model based on model evaluation tools), etc. In other product forms, the computing device can also create applications based on the model, provide API invocation capabilities, and can call the model into the created application through the API interface, while providing application management tools to achieve the management and monitoring of the application.
[0124] Furthermore, the computing device can also include data management (supporting the creation and management of model tuning data sets), a training center (providing rich training resources to help users learn and master AI technologies), and basic control capabilities (providing enterprise-level basic control capabilities to ensure the security and efficient operation of the system). Through the above functions, a comprehensive and integrated AI development, training, deployment, and application device is provided.
[0125] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and devices in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, to implement the methods in the above embodiments. The memory can include high-speed random access memory, and can also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory can further include a memory remotely set relative to the processor, and these remote memories can be connected to terminal A through a network. Examples of the above network include but are not limited to the Internet, enterprise intranets, local area networks, mobile communication networks, and their combinations.
[0126] The processor can call the executable program stored in the memory through the transmission device to execute the method described in any one of the above embodiments.
[0127] Embodiments of the present application can provide an electronic device. Figure 10 It is a structural block diagram of an electronic device shown according to an embodiment of the present application. As shown in the figure, the electronic device may include: an input / output device 1002; a memory 1004 and a processor 1006, wherein the processor 1006 is connected to the input / output device 1002 and the memory 1004 through a bus 1008.
[0128] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and devices in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implements the methods in the above embodiments. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory may further include a memory remotely provided with respect to the processor, and these remote memories can be connected to the terminal A through a network. Examples of the above network include but are not limited to the Internet, enterprise intranets, local area networks, mobile communication networks, and combinations thereof.
[0129] The processor can call the executable program stored in the memory through a transmission device to execute the method described in any one of the above embodiments.
[0130] Those of ordinary skill in the art can understand that the structure shown as Figure 10 is only schematic. The computing device can also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, and terminal devices such as Mobile Internet Devices (MID), PAD, etc. The Figure 10 does not limit the structure of the above computing device. For example, the computing device 900 may further include more or fewer components (such as a network interface, a display device, etc.) than those shown in the Figure 10 , or have a different configuration from that shown in the Figure 10 .
[0131] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the relevant hardware of the terminal device through a program, and the program can be stored in a computer-readable storage medium. The storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.
[0132] Embodiments of the present application also provide a computer-readable storage medium. Optionally, in this embodiment, the above computer-readable storage medium may be used to store the program code executed by the method provided in the above embodiment.
[0133] Optionally, in this embodiment, the above storage medium may be located in a computing device.
[0134] Optionally, in this embodiment, the computer-readable storage medium is configured to store an executable program, and when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the method described in any one of the above embodiments.
[0135] Embodiments of the present application also provide a computer program product. Optionally, in this embodiment, the above computer program product may include a computer program, and when the computer program is executed by a processor, it implements the method provided in the above embodiment.
[0136] Embodiments of the present application also provide a computer program product. Optionally, the above computer program product may include a non-volatile computer-readable storage medium, and the non-volatile computer-readable storage medium may be used to store a computer program, and when the computer program is executed by a processor, it implements the method provided in the above embodiment.
[0137] Embodiments of the present application also provide a computer program. Optionally, in this embodiment, when the above computer program is executed by a processor, it implements the method provided in the above embodiment.
[0138] In the above embodiments of the present application, the descriptions of the various embodiments each have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.
[0139] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the units or modules can be in an electrical or other form.
[0140] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0141] In addition, each functional unit in various embodiments of the present application can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
[0142] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: USB flash drive, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disc and other various media that can store program codes.
[0143] The above description is only a preferred embodiment of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.
Claims
1. An object recommendation method, characterized in that, Including: Obtaining object evaluation data and object images of multiple objects to be recommended, where the object evaluation data is used to characterize data that affects the access popularity of the objects to be recommended; Determining the recommended probability of the objects to be recommended based on the object evaluation data; Constructing an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to characterize the association relationship between the objects to be recommended; Selecting at least one object to be recommended from the multiple objects to be recommended as a target recommended object based on the recommended probability and the object retrieval result.
2. The method according to claim 1, characterized in that, The object retrieval result includes: a same-image retrieval result and a same-source retrieval result. The same-image retrieval result is used to characterize whether the types of the objects to be recommended are the same, and the same-source retrieval result is used to characterize whether the object sources of the objects to be recommended are the same. Constructing the object retrieval result of the objects to be recommended based on the object images includes: Constructing the same-image retrieval result based on at least one visual element of the object images; Constructing the same-source retrieval result based on at least one image content of the object images, where the data types of different image contents are different.
3. The method according to claim 2, characterized in that, Constructing the same-image retrieval result based on at least one visual element of the object images includes: Performing feature extraction on the at least one visual element to obtain at least one visual feature of the object images; Matching the multiple objects to be recommended based on the at least one visual feature corresponding to different object images to obtain a first similarity between different objects to be recommended; Constructing the same-image retrieval result based on the first similarity, where the same-image retrieval result includes: at least one image pair, and the at least one image pair includes: a reference image and a similar image. The reference image is used to characterize the object image corresponding to the object to be recommended for which the same-image retrieval result needs to be determined, and the first similarity between the reference image and the similar image is greater than a first preset value.
4. The method according to claim 2, characterized in that, Constructing the same-source retrieval result based on at least one image content of the object images includes: Performing feature extraction on the at least one image content to obtain at least one content feature of the object images; Matching at least one content feature corresponding to the reference image in the same-image retrieval result with at least one content feature corresponding to other images except the reference image in the object images to obtain a second similarity between the reference image and the other images; Constructing the same-source retrieval result based on the second similarity, where the same-source retrieval result includes: at least one object source data pair, and the at least one object source data pair includes: a first object source corresponding to the reference image and a second object source of the other images, and the second similarity between the first object source and the second object source is greater than a second preset value.
5. The method according to claim 1, characterized in that, Determining the recommended probability of the objects to be recommended based on the object evaluation data includes: Based on the object evaluation data, evaluate the object to be recommended from at least one evaluation dimension to obtain at least one evaluation score corresponding to the object to be recommended; Perform a weighting process on the at least one evaluation score to obtain an object evaluation score corresponding to the object to be recommended; Perform a normalization process on the object evaluation scores corresponding to different objects to be recommended to obtain the recommended probability; 6. The method according to claim 1, wherein The step of selecting at least one object to be recommended from the multiple objects to be recommended as the target recommended object based on the recommended probability and the object retrieval result includes: Based on the object retrieval result, construct a first data representation of the object to be recommended, where the first data representation includes: a first object identifier of a first object to be recommended and a second object identifier of a second object to be recommended, the first object to be recommended is any one of the objects to be recommended, and the types and object sources between the first object to be recommended and the second object to be recommended are the same; Update the first data representation based on the first recommended probability corresponding to the second object to be recommended to obtain a second data representation; Determine the object to be recommended corresponding to the second object identifier in the second data representation as the target recommended object.
7. The method according to claim 6, wherein The step of updating the first data representation based on the first recommended probability corresponding to the second object to be recommended to obtain a second data representation includes: Based on the first data representation and the first recommended probability, construct a third data representation, where the third data representation includes: the first object identifier, the second object identifier, and the first recommended probability; From the third data representation, determine third object identifiers corresponding to different first object identifiers, and second recommended probabilities corresponding to the third object identifiers, where the third object identifiers belong to the second object identifiers, and the second recommended probabilities are input to the first recommended probability; Based on the second recommended probability, select a target object identifier from the third object identifiers, where the second recommended probability corresponding to the target object identifier is greater than the second recommended probabilities corresponding to other object identifiers except the target object identifier in the third object identifiers; Based on the first object identifier, the target object identifier corresponding to the first object identifier, and the second recommended probability, construct the second data representation.
8. The method according to claim 1, wherein The method further includes: Obtain the object browsing data of the target client; Based on the object browsing data, construct object recommendation parameters of the target recommended object; Based on the recommended probability corresponding to the target recommended object, determine the traffic allocation weight of the target recommended object; Push the target recommended object to the target client based on the traffic allocation weight and the object recommendation parameters.
9. An object recommendation method, characterized in that, It includes: Respond to an input instruction on the operation interface, and display the object evaluation data and object images of multiple objects to be recommended on the operation interface, where the object evaluation data is used to characterize the data affecting the access popularity of the object to be recommended; In response to an object recommendation instruction acting on the operation interface, display a target recommended object on the operation interface, where the target recommended object is selected from the multiple objects to be recommended according to the recommended probability of the object to be recommended and the object retrieval result, the recommended probability is determined from the object evaluation data, the object retrieval result is constructed from the object image, and the object retrieval result is used to characterize the association relationship between the objects to be recommended.
10. An object recommendation method, characterized in that, It includes: Obtain the object evaluation data and object images of multiple objects to be recommended by calling a first interface, where the first interface includes a first parameter, and the parameter value of the first parameter includes the object evaluation data and the object images, and the object evaluation data is used to characterize the data affecting the access popularity of the objects to be recommended; Determine the recommended probability of the objects to be recommended based on the object evaluation data; Construct an object retrieval result of the objects to be recommended based on the object images, where the object retrieval result is used to characterize the association relationship between the objects to be recommended; Select at least one object to be recommended from the multiple objects to be recommended as the target recommended object based on the recommended probability and the object retrieval result; Output the target recommended object by calling a second interface, where the second interface includes a second parameter, and the parameter value of the second parameter includes the target recommended object.
11. A computer terminal, characterized in that, It includes: A memory storing an executable program; A processor for running the program, where when the program runs, it executes the method according to any one of claims 1 to 10.
12. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, where when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the method according to any one of claims 1 to 10.
13. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 10.