Recommended commodity identification method, apparatus and device, and computer program product

Abstract

The embodiment of the invention provides a recommended commodity recognition method, device and equipment and a computer program product. The recommended commodity identification method comprises the steps that to-be-published information of a to-be-published commodity and a published commodity set are acquired, and the published commodity set comprises a plurality of published commodities and commodity publishing information corresponding to the published commodities; based on the published commodity set, determining an estimated publishing statistical feature of the to-be-published commodity; based on the to-be-published information, determining static attribute features of the to-be-published commodity; and identifying whether the to-be-released commodity is a recommended commodity based on the static attribute feature and the pre-estimated release statistical feature. In the embodiment of the invention, as the pre-estimated release statistical characteristics can represent the future performance of the to-be-released commodity, when the to-be-released commodity is analyzed and identified in combination with the static attribute characteristics and the pre-estimated release statistical characteristics, the accuracy and reliability of identifying whether the to-be-released commodity is the recommended commodity are effectively ensured.

Description

Technical Field

[0001] The present invention relates to the field of network technology, and in particular to a method, device, equipment and computer program product for identifying recommended commodities. Background Art

[0002] Product launch refers to launching new products on e-commerce platforms to attract consumers and increase sales. In order to successfully launch products to the market and achieve good sales performance, e-commerce platforms often select high-quality products from a large number of new products, and then help them grow rapidly through traffic support.

[0003] At present, high-quality products are often determined through tree models. Specifically, tree models can predict high-quality products and estimate product traffic based on the generalized characteristics of products. However, the above-mentioned prediction operation of high-quality products uses fewer reference indicators, which reduces the prediction accuracy of high-quality products. Summary of the invention

[0004] The embodiments of the present invention provide a method, apparatus, device and computer program product for identifying recommended commodities, which can accurately identify the recommended commodities.

[0005] In a first aspect, an embodiment of the present invention provides a method for identifying a recommended product, comprising:

[0006] Acquire information of a product to be released and a set of released products, wherein the set of released products includes multiple released products and product release information corresponding to the released products;

[0007] Determining estimated release statistical features of the to-be-released product based on the released product set;

[0008] Based on the information to be released, determining static attribute characteristics of the product to be released;

[0009] Based on the static attribute features and the estimated release statistical features, it is identified whether the to-be-released product is a recommended product.

[0010] In a second aspect, an embodiment of the present invention provides a device for identifying recommended products, including:

[0011] A first acquisition module is used to acquire information of a to-be-released commodity and a set of released commodities, wherein the set of released commodities includes a plurality of released commodities and commodity release information corresponding to the released commodities;

[0012] A first determination module, configured to determine estimated release statistical features of the to-be-released product based on the released product set;

[0013] The first determination module is further used to determine the static attribute characteristics of the commodity to be released based on the information to be released;

[0014] The first processing module is used to identify whether the product to be released is a recommended product based on the static attribute characteristics and the estimated release statistical characteristics.

[0015] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions, when executed by the processor, implement the method for identifying recommended products in the above-mentioned first aspect.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer storage medium for storing a computer program, wherein the computer program implements the method for identifying recommended products in the first aspect when executed by a computer.

[0017] In a fifth aspect, an embodiment of the present invention provides a computer program product, comprising: a computer-readable storage medium storing computer instructions, which, when executed by one or more processors, causes the one or more processors to execute the steps of the method for identifying recommended products shown in the first aspect above.

[0018] In a sixth aspect, an embodiment of the present invention provides a method for identifying recommended products, including:

[0019] Acquire a set of products to be released and a set of released products, wherein the set of products to be released includes a plurality of products to be released and to-be-released information corresponding to the products to be released, and the set of released products includes a plurality of released products and product release information corresponding to the released products;

[0020] Based on the set of released products, determining estimated release statistical features corresponding to the to-be-released products;

[0021] Determining static attribute features corresponding to the product to be released based on the to-be-released information corresponding to the product to be released;

[0022] At least one recommended product is determined in the set of products to be released based on the static attribute characteristics and the estimated release statistical characteristics corresponding to the products to be released.

[0023] In a seventh aspect, an embodiment of the present invention provides a device for identifying recommended products, including:

[0024] A second acquisition module is used to acquire a set of products to be released and a set of released products, wherein the set of products to be released includes a plurality of products to be released and the corresponding information to be released of the products to be released, and the set of released products includes a plurality of released products and the corresponding product release information of the released products;

[0025] A second determination module, configured to determine, based on the set of released products, estimated release statistical features corresponding to the product to be released;

[0026] The second determination module is further configured to determine the static attribute characteristics corresponding to the to-be-released product based on the to-be-released information corresponding to the to-be-released product;

[0027] The second processing module is used to determine at least one recommended product in the set of products to be released based on the static attribute characteristics and the estimated release statistical characteristics corresponding to the products to be released.

[0028] In an eighth aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions, when executed by the processor, implement the method for identifying recommended products in the above-mentioned sixth aspect.

[0029] In a ninth aspect, an embodiment of the present invention provides a computer storage medium for storing a computer program, wherein the computer program implements the method for identifying recommended products in the sixth aspect when executed by a computer.

[0030] In the tenth aspect, an embodiment of the present invention provides a computer program product, comprising: a computer-readable storage medium storing computer instructions, which, when executed by one or more processors, causes the one or more processors to execute the steps of the method for identifying recommended products shown in the sixth aspect above.

[0031] The recommended product identification method, apparatus, device and computer program product provided in the present embodiment obtains the to-be-released information of the to-be-released product and a set of released products, determines the estimated release statistical features of the to-be-released product based on the set of released products, and determines the static attribute features of the to-be-released product based on the to-be-released information, and then identifies whether the to-be-released product is a recommended product based on the static attribute features and the estimated release statistical features. Since the estimated release statistical features can characterize the future performance of the to-be-released product, when the static attribute features and the estimated release statistical features are combined to analyze and identify the to-be-released product, the accuracy and reliability of identifying the recommended product is effectively guaranteed, thereby further improving the practicality of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 A schematic diagram of a scenario of a method for identifying recommended products provided by an embodiment of the present invention;

[0034] Figure 2 A schematic diagram of a flow chart of a method for identifying recommended commodities provided by an embodiment of the present invention;

[0035] Figure 3 A schematic diagram of a process for determining estimated release statistical features of the to-be-released product based on the released product set provided in an embodiment of the present invention;

[0036] Figure 4 A schematic diagram of a process for identifying whether the product to be released is a recommended product based on the static attribute feature and the estimated release statistical feature provided by an embodiment of the present invention;

[0037] Figure 5 A schematic diagram of the principle of a new product selection method based on multimodal characterization retrieval of similar sequences provided in an application embodiment of the present invention;

[0038] Figure 6 A schematic diagram of the principle of obtaining sequence features of new products to be released provided in an application embodiment of the present invention;

[0039] Figure 7 A schematic diagram of a flow chart of a method for identifying recommended commodities provided by an embodiment of the present invention;

[0040] Figure 8 A schematic diagram of the structure of a device for identifying recommended products provided by an embodiment of the present invention;

[0041] Fig. 9 For Figure 8 A schematic diagram of the structure of an electronic device corresponding to the identification device for recommended commodities provided in an embodiment is shown;

[0042] Fig.10 A schematic diagram of the structure of a device for identifying recommended products provided by an embodiment of the present invention;

[0043] Fig.11 For Fig.10 A schematic diagram of the structure of an electronic device corresponding to the device for identifying recommended products provided in an embodiment is shown. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "an", and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings, and "multiple" generally includes at least two, but does not exclude the situation of including at least one.

[0046] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0047] As used herein, the words "if" and "if" may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0048] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a product or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such a product or system. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the product or system including the elements.

[0049] In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0050] Definition of terms:

[0051] Multimodal features: refers to features extracted from multiple different types of data (for example, text, images, audio, video, etc.), which can be used for various tasks such as classification, retrieval, generation, and understanding.

[0052] M6: Multi-Modality to Multi-Modality Multitask Mega-transformer, a super-large-scale Chinese pre-training model that can process and generate multiple types of data, such as text, images, videos, etc., and performs well on multiple tasks.

[0053] In order to facilitate understanding of the specific implementation process and implementation effect of the recommended product identification method, device, equipment and computer program product in this embodiment, the relevant technology is briefly described below:

[0054] Product launch refers to launching new products on e-commerce platforms to attract consumers and increase sales. In order to successfully bring products to the market and achieve good sales performance, e-commerce platforms often select high-quality products from a large number of new products, and then help them grow rapidly through traffic support, hoping that these high-quality products will be able to retain long-term retention in the first guessing scenario and have a strong ability to obtain long-term traffic after support.

[0055] In traditional method 1, high-quality products are usually selected through operation and investment promotion. In the process of operation and investment promotion, merchants can obtain a certain amount of cold-start support traffic by transferring rights and interests after signing up. However, since cold-start support traffic is only a short-term traffic support measure, it cannot be operated when new products need to be incubated continuously.

[0056] In traditional method 2, high-quality products are often determined through tree models. Specifically, the tree model can predict high-quality products and estimate product traffic through the generalized characteristics of the products. However, the above-mentioned prediction operation of high-quality products uses fewer indicator parameters, thereby reducing the prediction accuracy of high-quality products.

[0057] In order to solve the technical problems existing in the above-mentioned related technologies, this embodiment provides a method, device, equipment and computer program product for identifying recommended products. Figure 1As shown, the execution subject of the method for identifying recommended goods can be a device for identifying recommended goods 200, and the device for identifying recommended goods 200 can be implemented as a local server, a server in the cloud, or a preset device. In the case where the device for identifying recommended goods 200 is implemented as a server in the cloud, the method for identifying recommended goods can be executed in the cloud, and several computing nodes (cloud servers) can be deployed in the cloud, and each computing node has computing, storage and other processing resources. In the cloud, multiple computing nodes can be organized to provide a certain service. Of course, a computing node can also provide one or more services. The way to provide the service in the cloud can be to provide a service interface to the outside, and the user calls the service interface to use the corresponding service. The service interface includes software development kits (Software Development Kit, referred to as SDK), application programming interfaces (Application Programming Interface, referred to as API), and other forms.

[0058] The identification device 200 of the recommended product is connected to the client 100 in communication, wherein the client 100 is used for users to perform applications so as to trigger the identification operation of the recommended product. The above-mentioned client 100 can be any computing device with a certain ability to identify recommended products. In specific implementation, the client 100 can be a mobile phone, a personal computer PC, a tablet computer, a setting application, etc. In addition, the basic structure of the client 100 may include: at least one processor. The number of processors depends on the configuration and type of the client. The client 100 may also include a memory, which may be volatile, such as a random access memory (RAM), or non-volatile, such as a read-only memory (ROM), flash memory, etc., or may include both types. The memory usually stores an operating system (OS), one or more application programs, and may also store program data, etc. In addition to the processing unit and the memory, the client 100 also includes some basic configurations, such as a network card chip, an IO bus, a display component, and some peripheral devices, etc. Optionally, some peripheral devices may include, for example, a keyboard, a mouse, an input pen, a printer, etc. Other peripheral devices are well known in the art and will not be described in detail here.

[0059] The recommended product identification device 200 refers to a device that can implement the recommended product identification operation in a network virtual environment, and generally refers to a device that uses the network to perform information planning and recommended product identification operations. In physical implementation, the recommended product identification device 200 can be any device that can provide computing services and can perform corresponding recommended product identification operations, for example: it can be a cluster server, a conventional server, a cloud server, a cloud host, a virtual center, etc. The recommended product identification device 200 mainly includes a processor, a hard disk, a memory, a system bus, etc., which is similar to a general computer architecture.

[0060] In the above-mentioned embodiment, the client 100 is connected to the identification device 200 of the recommended product through a network, and the network connection can be a wireless or wired network connection. If the client 100 can be connected to the identification device 200 of the recommended product through a communication connection, the network standard of the mobile network can be any one of 2G (Global System for Mobile Communications GSM), 2.5G (General Packet Radio Service GPRS), 3G (Wideband Code Division Multiple Access WCDMA, Time Division Synchronous Code Division Multiple Access TD-SCDMA, Code Division Multiple Access 2000 CDMA2000, Universal Mobile Telecommunications System UTMS), 4G (Long Term Evolution LTE), 4G+ (Enhanced Long Term Evolution LTE+), Worldwide Interoperability for Microwave Access WiMax, 5G, 6G, etc.

[0061] In the embodiment of the present application, the client 100 is used by the user to generate an identification request for a recommended product, wherein the identification request for a recommended product can be generated by a human-computer interaction operation or a voice interaction operation. After obtaining the identification request for a recommended product, the identification request for a recommended product can be sent to the identification device 200 for recommending a product, so that the identification device 200 for recommending a product can perform a corresponding identification operation for the recommended product based on the identification request for the recommended product.

[0062] The recommended product identification device 200 is used to obtain a recommended product identification request through the client 100, and the obtained recommended product identification request includes at least a to-be-released product and a set of released products, wherein the to-be-released product corresponds to information to be released, and the set of released products may include multiple released products and product release information corresponding to the released products; in order to be able to identify whether the to-be-released product is a recommended product, after obtaining the to-be-released product and the set of released products, the estimated release statistical features of the to-be-released product can be determined based on the set of released products, and the static attribute features of the to-be-released product can be determined based on the information to be released; then, the static attribute features and the estimated release statistical features can be analyzed and processed to identify whether the to-be-released product is a recommended product, so as to obtain an identification result, which can include a first result for identifying that the to-be-released product is a recommended product; or, the identification result can include a second result for identifying that the to-be-released product is not a recommended product.

[0063] In this embodiment, by obtaining the to-be-released information of the to-be-released product and the set of released products, the estimated release statistical features of the to-be-released product are determined based on the set of released products, and the static attribute features of the to-be-released product are determined based on the to-be-released information, and then based on the static attribute features and the estimated release statistical features, it is identified whether the to-be-released product is a recommended product. Since the estimated release statistical features can characterize the future performance of the to-be-released product, and the static attribute features refer to features that are not dependent on historical data but are directly extracted from the attributes of the product itself or the category to which it belongs, therefore, when the static attribute features and the estimated release statistical features are combined to identify the to-be-released product, the accuracy and reliability of identifying whether the to-be-released product is a recommended product is effectively guaranteed, thereby further improving the practicality of the method.

[0064] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the case where there is no conflict between the embodiments, the following embodiments and the features in the embodiments can be combined with each other. In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0065] Figure 2 A flowchart of a method for identifying recommended products provided by an embodiment of the present invention; see attached Figure 2As shown, this embodiment provides a method for identifying recommended products. The execution subject of the method is a device for identifying recommended products. Specifically, the device for identifying recommended products can be implemented as software or a combination of software and hardware. When the device for identifying recommended products is implemented as hardware, it can be various electronic devices that can implement the identification operation of recommended products, including but not limited to personal computers, servers, databases, etc. When the device for identifying recommended products is implemented as software, it can be installed in the electronic devices listed above. Based on the above-mentioned device for identifying recommended products, the identification operation of recommended products can be implemented. Specifically, the method for identifying recommended products can include:

[0066] Step S201: obtaining information of a to-be-released commodity and a set of released commodities, wherein the set of released commodities includes a plurality of released commodities and commodity release information corresponding to the released commodities.

[0067] Step S202: based on the set of released products, determine the estimated release statistical features of the products to be released.

[0068] Step S203: Based on the information to be released, determine the static attribute characteristics of the product to be released.

[0069] Step S204: Based on the static attribute features and the estimated release statistical features, identify whether the product to be released is a recommended product.

[0070] The specific implementation process and implementation effects of the above steps are described in detail below:

[0071] Step S201: obtaining information of a to-be-released commodity and a set of released commodities, wherein the set of released commodities includes a plurality of released commodities and commodity release information corresponding to the released commodities.

[0072] When a user has a need to identify recommended products, the identification device for recommended products can obtain the to-be-released information of the to-be-released products. Specifically, the identification device for recommended products can obtain the to-be-released information of the to-be-released products through the e-commerce platform. For the to-be-released products, they can refer to products that are waiting to be exposed on the e-commerce platform but have not been released. The number of to-be-released products can be one or more, and the to-be-released information can include text information and picture information of the to-be-released products. The above-mentioned text information can include at least one of the following: title information, product introduction information, and the picture information can include the main view of the to-be-released product or a partial release image or a full release image, etc. The number of to-be-released information corresponding to the to-be-released products is also one or more.

[0073] For the products to be released, since the products to be released have not been released on the e-commerce platform, the to-be-released information of the products to be released does not include the release statistical features corresponding to the products to be released. In order to accurately identify whether the products to be released are recommended products, in addition to obtaining the to-be-released information of the products to be released, a set of released products can also be obtained. The set of released products can include multiple released products and product release information corresponding to the released products, wherein the released products corresponding to the product release information are at least a part of the multiple released products. The product release information may include release statistical features, which may refer to features obtained by statistical analysis after the product is exposed and released, which can usually reflect the performance of the released products over a period of time based on historical data. In some instances, the release statistical features may include at least one of the following: sales volume, sales revenue, average price, product conversion rate, product return rate, product sales trend, etc. within a preset time period.

[0074] Step S202: based on the set of released products, determine the estimated release statistical features of the products to be released.

[0075] Since there are no corresponding release statistical features for the products to be released, in order to accurately analyze and process the products to be released, the set of released products can be analyzed and processed to determine the estimated release statistical features of the products to be released, wherein the estimated release statistical features may include at least one of the following: exposure volume, transaction ratio, and total transaction amount in a preset time period.

[0076] In some instances, determining the estimated release statistical features of a to-be-released product based on a set of released products may include: obtaining product details of the to-be-released product, which can reflect the specific features of the product; determining a target released product that matches the product details in the set of released products; and determining the released statistical features of the target released product as the estimated release statistical features of the to-be-released product, thereby effectively ensuring the accuracy and reliability of determining the estimated release statistical features of the to-be-released product.

[0077] Step S203: Based on the information to be released, determine the static attribute characteristics of the product to be released.

[0078] After obtaining the to-be-released information of the to-be-released product, in order to accurately identify whether the to-be-released product is a recommended product, it is necessary not only to determine the estimated release statistical characteristics of the to-be-released product, but also to determine the static attribute characteristics of the to-be-released product. The static attribute characteristics refer to characteristics that do not rely on specific historical data, but are extracted from the attributes of the product itself or the category to which it belongs. In some instances, the static attribute characteristics may include at least one of the following: product category, store information, store service capability, and the original price of the product.

[0079] In some instances, static attribute features can be obtained by performing feature extraction operations on the to-be-released information of the product to be released through a feature extraction model. At this time, determining the static attribute features of the product to be released based on the information to be released may include: obtaining a pre-trained feature extraction model; inputting the to-be-released information of the product to be released into the feature extraction model for feature extraction operations, thereby accurately obtaining the static attribute features of the product to be released.

[0080] Step S204: Based on the static attribute features and the estimated release statistical features, identify whether the product to be released is a recommended product.

[0081] After obtaining the static attribute features and the estimated release statistical features, the static attribute features and the estimated release statistical features can be analyzed and processed to identify whether the product to be released is a recommended product. In some instances, the identification operation of the product to be released can be implemented by a pre-trained deep learning model. At this time, based on the static attribute features and the estimated release statistical features, identifying whether the product to be released is a recommended product can include: obtaining a pre-trained deep learning model; inputting the static attribute features and the estimated release statistical features into the deep learning model to perform the identification operation of the recommended product, and obtaining an identification result for identifying whether the product to be released is a recommended product. The identification result can be a first result for identifying that the product to be released is a recommended product; or, the identification result can be a second result for identifying that the product to be released is not a recommended product.

[0082] The method for identifying recommended products provided in the present embodiment obtains the to-be-released information of the to-be-released products and a set of released products, determines the estimated release statistical features of the to-be-released products based on the set of released products, and determines the static attribute features of the to-be-released products based on the to-be-released information, and then identifies whether the to-be-released product is a recommended product based on the static attribute features and the estimated release statistical features. Since the estimated release statistical features can characterize the future performance of the to-be-released product, when the static attribute features and the estimated release statistical features are combined to analyze and identify the to-be-released products, the accuracy and reliability of identifying the recommended products is effectively guaranteed, thereby further improving the practicality of the method.

[0083] Figure 3 A schematic diagram of a process for determining estimated release statistical features of a product to be released based on a set of released products provided in an embodiment of the present invention; based on the above embodiment, refer to the attached Figure 3 As shown, this embodiment provides a method of determining the estimated release statistical features of the product to be released by using at least one similar product corresponding to the product to be released. At this time, based on the released product set, determining the estimated release statistical features of the product to be released may include:

[0084] Step S301: Determine at least one similar product corresponding to the product to be released in the set of released products.

[0085] Since the goods to be released are goods to be released in the future on the e-commerce platform, they may be brand new goods. At this time, there may not be any released goods in the released goods set that are exactly the same as the goods details of the goods to be released. Therefore, it may not be possible to determine similar goods corresponding to the goods to be released in the released goods set directly based on the goods details. At this time, similar goods can be determined based on the goods features of the goods to be released and the published features of the released goods in the released goods set, that is, similar goods can not only be determined by the goods categories similar to the goods to be released, that is, the released goods corresponding to the goods categories similar to the goods to be released in the released goods set are determined as similar goods; they can also be determined by analyzing and processing the goods features of the goods to be released and the published features of the released goods in the released goods set. At this time, determining at least one similar goods corresponding to the goods to be released in the released goods set may include: obtaining the first multimodal features of the goods to be released, and the second multimodal features corresponding to each of the multiple released goods in the released goods set; based on the first multimodal features and the multiple second multimodal features, determining at least one similar goods corresponding to the goods to be released in the released goods set.

[0086] Among them, after obtaining the set of products to be released and the set of released products, feature extraction operations can be performed on the products to be released and the set of released products respectively to obtain the first multimodal features of the products to be released and the second multimodal features corresponding to each of the multiple released products in the set of released products. In some instances, the first multimodal features and the second multimodal features can be determined by analyzing and processing the products to be released and the released products in the set of released products respectively through a pre-trained deep learning model (for example: the ultra-large-scale Chinese pre-training model M6), that is, the information to be released of the products to be released is input into the deep learning model for feature extraction operations to obtain the first multimodal features output by the deep learning model, and the product release information of the released products in the set of released products is input into the deep learning model for feature extraction operations to obtain the second multimodal features corresponding to the released products output by the deep learning model.

[0087] After obtaining the first multimodal features of the product to be released and the second multimodal features corresponding to each of the multiple released products in the set of released products, the first multimodal features and the multiple second multimodal features can be analyzed and processed to determine at least one similar product corresponding to the product to be released in the set of released products.

[0088] In some instances, at least one similar product can be realized by analyzing and processing the first multimodal features and the second multimodal features through a pre-trained machine learning model; or, at least one similar product can be determined by the similarity between the released product and the product to be released. At this time, based on the first multimodal feature and multiple second multimodal features, determining at least one similar product corresponding to the product to be released in the set of released products may include: determining the similarity between the released product and the product to be released based on the first multimodal feature and multiple second multimodal features; determining at least one similar product corresponding to the product to be released in the set of released products based on the similarity between the released product and the product to be released.

[0089] After obtaining the first multimodal feature and the plurality of second multimodal features, the first multimodal feature and the second multimodal feature may be analyzed and processed to determine the similarity between the released product and the product to be released, and the similarity may be implemented as any one of the following: cosine similarity, Euclidean distance, Manhattan distance, Pearson correlation coefficient, etc. After obtaining the similarity between the released product and the product to be released, the similarity between the released product and the product to be released may be analyzed and processed to determine at least one similar product corresponding to the product to be released in the released product set.

[0090] In some instances, similar products can be determined by a preset threshold. In this case, based on the similarity between the published product and the product to be published, determining at least one similar product corresponding to the product to be published in the published product set may include: when the similarity between the published product and the product to be published is greater than or equal to the preset threshold, determining the published product corresponding to the similarity as a similar product corresponding to the product to be published; when the similarity between the published product and the product to be published is less than the preset threshold, determining the published product corresponding to the similarity as not a similar product corresponding to the product to be published, thereby effectively ensuring the accuracy and reliability of the determination of similar products.

[0091] In some other instances, similar products can be determined not only by a preset threshold but also by sorting the published products. In this case, based on the similarity between the published products and the products to be published, determining at least one similar product corresponding to the product to be published in the published product set may include: sorting the published products based on the similarity between the published products and the products to be published to obtain product sorting information; determining quantity information for limiting similar products; and determining at least one similar product corresponding to the product to be published in the published product set based on the quantity information and the product sorting information.

[0092] For example, after the published products include product a, product b, product c, product d and product e, the similarity between the product to be published and product a is similarity 1, the similarity between the product to be published and product b is similarity 2, the similarity between the product to be published and product c is similarity 3, the similarity between the product to be published and product d is similarity 4, and the similarity between the product to be published and product e is similarity 5. Then, the published products can be sorted from high to low based on the similarity between the published products and the products to be published. Assuming that the sorting information is product d-product c-product b-product e-product a, if the quantity information used to limit similar products is 3, then based on the above quantity information and product sorting information, at least one similar product corresponding to the product to be published in the published product set can be determined to include: product d, product c and product b, thereby effectively ensuring the accuracy and reliability of determining at least one similar product.

[0093] Step S302: Determine the publication statistical features corresponding to at least one similar product.

[0094] After determining at least one similar product corresponding to the product to be released, the product release information corresponding to the at least one similar product can be analyzed and processed to determine the release statistical characteristics corresponding to the at least one similar product. In some instances, the release statistical characteristics include at least one of the following: exposure volume, transaction ratio, total transaction amount, etc. in a preset time period, which can be specifically determined by the product identification of the similar products and the preset mapping relationship.

[0095] Step S303: determining the estimated release statistical feature of the product to be released based on the release statistical feature corresponding to at least one similar product.

[0096] Since the similarity between similar products and the product to be released is high, after obtaining the release statistical feature corresponding to at least one similar product, the estimated release statistical feature of the product to be released can be determined based on the release statistical feature corresponding to at least one similar product. In some examples, the release statistical feature corresponding to at least one similar product can be directly determined as the estimated release statistical feature of the product to be released.

[0097] In other instances, the release statistical features corresponding to at least one similar product can be analyzed and processed in combination with the weight information of the release statistical features to determine the estimated release statistical features of the product to be released. At this time, determining the estimated release statistical features of the product to be released based on the release statistical features corresponding to at least one similar product can include: obtaining the weight information corresponding to the release statistical features; determining the estimated release statistical features of the product to be released based on the release statistical features corresponding to at least one similar product and the weight information corresponding to the release statistical features.

[0098] Among them, since different similar products have different similarities with the product to be released, in order to accurately determine the estimated release statistical features of the product to be released, the weight information corresponding to the release statistical features can be obtained. In some instances, obtaining the weight information corresponding to the release statistical features can include: obtaining the similarity between the similar products and the product to be released; based on the similarity between the similar products and the product to be released, determining the weight information corresponding to the release statistical features, the weight information corresponding to the release statistical features is positively correlated with the similarity between the similar products and the product to be released.

[0099] After obtaining the weight information corresponding to the release statistical feature, the release statistical feature corresponding to at least one similar product and the weight information corresponding to the release statistical feature may be analyzed and processed to determine the estimated release statistical feature of the product to be released.

[0100] In some instances, determining the estimated release statistical features of a product to be released based on the release statistical features corresponding to each of at least one similar product and the weight information corresponding to the release statistical features may include: performing a weighted sum operation on the release statistical features corresponding to each of at least one similar product based on the weight information corresponding to the release statistical features, thereby stably obtaining the estimated release statistical features of the product to be released.

[0101] In other instances, determining the estimated release statistical features of the commodity to be released based on the release statistical features corresponding to each of at least one similar commodity and the weight information corresponding to the release statistical features may include: performing weighted averaging processing on the release statistical features corresponding to each of at least one similar commodity based on the weight information corresponding to the release statistical features, thereby stably obtaining the estimated release statistical features of the commodity to be released.

[0102] In this embodiment, in the set of released products, by determining at least one similar product corresponding to the product to be released and determining the release statistical features corresponding to each of the at least one similar product, and then determining the estimated release statistical features of the product to be released based on the release statistical features corresponding to each of the at least one similar product, the flexibility and reliability of determining the estimated release statistical features are effectively guaranteed.

[0103] Figure 4 A flowchart of identifying whether a product to be released is a recommended product based on static attribute features and estimated release statistical features provided by an embodiment of the present invention; based on any of the above embodiments, refer to the attached Figure 4 As shown, the identification of whether the product to be released is a recommended product can be achieved not only by a pre-trained deep learning model, but also by combining the global statistical features and multimodal features of the product to be released to perform an identification operation on the product to be released. At this time, based on the static attribute features and the estimated release statistical features, the identification of whether the product to be released is a recommended product in this embodiment may include:

[0104] Step S401: Obtain global statistical features and multimodal features of the products to be released.

[0105] The goods to be released may correspond to global statistical features, wherein global statistical features may refer to all statistical features corresponding to the goods to be released that can be queried in the network platform, and global statistical features may include statistical features of the merchant dimension and statistical features of the goods dimension. The statistical features of the merchant dimension may include at least one of the following: exposure within a preset time period, click-through rate, conversion rate, and relative price level in the same category; the statistical features of the goods dimension may include at least one of the following: product identification, product name, product category, product brand, product specification, product color and material, product applicable population, etc. It should be noted that since the goods to be released have not been released, the global statistical features of the goods to be released do not include the release statistical features of the goods to be released, that is, it can be considered that the features corresponding to the goods to be released in the release scenario are null values.

[0106] In order to accurately identify whether a product to be released is a recommended product, it is necessary not only to obtain the global statistical features of the product to be released, but also to obtain the multimodal features of the product to be released. The multimodal features of the product to be released can be obtained by performing an extraction operation on the product to be released through a feature extraction model. The above-mentioned feature extraction model can be an M6 model. After obtaining the to-be-released information of the product to be released, the to-be-released information of the product to be released can be input into the M6 ​​model for feature extraction operation to obtain the multimodal features of the product to be released output by the M6 ​​model.

[0107] Step S402: Based on static attribute features, estimated release statistical features, global statistical features and multimodal features, identify whether the product to be released is a recommended product.

[0108] After acquiring the static attribute features, the estimated release statistical features, the global statistical features, and the multimodal features, the static attribute features, the estimated release statistical features, the global statistical features, and the multimodal features may be analyzed and processed to identify whether the product to be released is a recommended product. In some instances, based on the static attribute features, the estimated release statistical features, the global statistical features, and the multimodal features, identifying whether the product to be released is a recommended product may include: determining the potential information of the product to be released to become a recommended product based on the static attribute features, the estimated release statistical features, the global statistical features, and the multimodal features; and identifying whether the product to be released is a recommended product based on the potential information.

[0109] Specifically, after obtaining static attribute features, estimated release statistical features, all statistical features and multimodal features, the static attribute features, estimated release statistical features, all statistical features and multimodal features can be analyzed and processed to determine the potential information of the product to be released to become a recommended product. In some instances, the potential information can be determined by a pre-trained deep learning model. At this time, based on the static attribute features, estimated release statistical features, global statistical features and multimodal features, determining the potential information of the product to be released to become a recommended product can include: obtaining a pre-trained deep learning model; inputting the static attribute features, estimated release statistical features, global statistical features and multimodal features into the deep learning model for analysis and processing to obtain the potential information of the product to be released to become a recommended product, which effectively ensures the accuracy and reliability of determining the potential information of the recommended product.

[0110] In other instances, potential information can not only be obtained by unified analysis and processing of static attribute features, estimated release statistical features, global statistical features and multimodal features by a deep learning model, but can also be determined by the potential of the to-be-released product to obtain traffic and the potential of the to-be-released product to become a recommended product. At this time, determining the potential information of the to-be-released product to become a recommended product based on the static attribute features, estimated release statistical features, global statistical features and multimodal features may include: determining the first potential of the to-be-released product to obtain traffic and the second potential of the to-be-released product to become a transaction target based on the static attribute features, estimated release statistical features, global statistical features and multimodal features; determining the potential information of the to-be-released product to become a recommended product based on the first potential and the second potential.

[0111] Since the potential information of the to-be-released product to become a recommended product is related to the ability of the to-be-released product to obtain traffic and the potential of the to-be-released product to become a recommended product, in order to accurately determine the potential information of the to-be-released product to become a recommended product, the static attribute features, estimated release statistical features, global statistical features and multimodal features can be analyzed and processed to determine the first potential of the to-be-released product to obtain traffic and the second potential of the to-be-released product to become a transaction target. It should be noted that "the first potential of the to-be-released product to obtain traffic" can refer to the potential of the to-be-released product to obtain traffic within a preset time period, and the preset time period can be a pre-configured long time period, for example: the preset time period can be a pre-configured 30 days, 45 days or 60 days, etc.

[0112] After determining the first potential of the to-be-released product to obtain traffic and the second potential of the to-be-released product to become a transaction target, the first potential and the second potential can be analyzed and processed to determine the potential information of the to-be-released product to become a recommended product, wherein the potential information can be positively correlated with the first score, and the potential information can be positively correlated with the second score. In some instances, based on the first potential and the second potential, determining the potential information of the to-be-released product to become a recommended product may include: performing a weighted summation of the first potential and the second potential to obtain the potential information of the to-be-released product to become a recommended product. Alternatively, based on the first potential and the second potential, determining the potential information of the to-be-released product to become a recommended product may include: performing a weighted average operation on the first potential and the second potential to obtain the potential information of the to-be-released product to become a recommended product.

[0113] In other instances, determining the potential information of the product to be released to become a recommended product based on the first potential and the second potential may include: determining the third potential of the product to be released to become a successfully incubated product based on static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics; determining the potential information of the product to be released to become a recommended product based on the first potential, the second potential and the third potential.

[0114] Specifically, the potential information of the product to be released to become a recommended product is not only related to the ability of the product to be released to obtain traffic and the potential of the product to be released to become a recommended product, but also to the potential of the product to be released to become a successfully incubated product, where a successfully incubated product may refer to a product that has been successfully promoted in the market from an idea. Therefore, in order to accurately determine the potential information of the product to be released to become a recommended product, the static attribute features, estimated release statistical features, global statistical features and multimodal features may be analyzed and processed to determine the third potential of the product to be released to become a successfully incubated product. The method for determining the third potential is similar to the method for determining the first potential and the second potential mentioned above. For details, please refer to the above statements and will not be repeated here.

[0115] After determining the third potential of the product to be released to become a successfully incubated product, the first potential, the second potential and the third potential can be analyzed and processed to determine the potential information of the product to be released to become a recommended product. The potential information can be positively correlated with the first score, the potential information can be positively correlated with the second score, and the potential information can be positively correlated with the third score.

[0116] In some examples, based on the first potential, the second potential, and the third potential, determining the potential information of the to-be-released product to become a recommended product may include: performing a weighted sum operation on the first potential, the second potential, and the third potential to obtain the potential information of the to-be-released product to become a recommended product. Alternatively, based on the first potential, the second potential, and the third potential, determining the potential information of the to-be-released product to become a recommended product may include: performing a weighted average operation on the first potential, the second potential, and the third potential to obtain the potential information of the to-be-released product to become a recommended product.

[0117] After obtaining the potential information, the potential information can be analyzed and processed to identify whether the product to be released is a recommended product. In some instances, based on the potential information, identifying whether the product to be released is a recommended product can include: obtaining a potential threshold for analyzing and processing the potential information; when the potential information is greater than or equal to the preset threshold, it means that the probability of the product to be released becoming a recommended product is high, and the product to be released can be determined as a recommended product; when the potential information is less than the preset threshold, it means that the probability of the product to be released becoming a recommended product is low, and the product to be released can be determined as not a recommended product, thereby effectively achieving accurate and reliable identification of recommended products.

[0118] In this embodiment, by obtaining the global statistical features and multimodal features of the product to be released, and then identifying whether the product to be released is a recommended product based on static attribute features, estimated release statistical features, global statistical features and multimodal features, it is effectively achieved to accurately identify whether the product to be released is a recommended product.

[0119] In specific applications, take the first guess 30 days exposure product set as the released product set as an example, refer to the attached Figure 5-Figure 6 As shown, this application embodiment provides a method for selecting high-quality products based on multimodal representation. The method can retrieve similar product sequences based on the multimodal representation of released products, and enhance the sequence representation of new products to be released by using an attention mechanism and similar product sequences. Then, the selection operation of high-quality products can be implemented based on the sequence representation of new products to be released, further improving the accuracy and reliability of selecting high-quality products. Specifically, the selection method may include the following steps:

[0120] Step 1: Get the new products to be released and the collection of released products.

[0121] Among them, the new products to be released correspond to the information to be released, and the information to be released may include at least one of the following: text information and picture information of the products to be released, and the above text information may include at least one of the following: title information, product introduction information, and picture information may include the main view of the products to be released or part of the released image or all the released images, etc. The released product set may be a first guess 30-day exposure product set, and the first guess 30-day exposure product set may include multiple released products and release statistics corresponding to each released product, and the release statistics may include at least one of the following: sales volume, sales, average price, conversion rate of products, return rate of products, sales trend of products, etc. within a preset time period.

[0122] Step 2: Determine the first multimodal features of the new product and the second multimodal features corresponding to each released product in the released product set.

[0123] Obtain an M6 model for implementing a multimodal feature extraction operation, use the M6 ​​model to analyze and process the to-be-released information of new products, and obtain the first multimodal feature output by the M6 ​​model; use the M6 ​​model to analyze and process the release statistical information corresponding to each released product, and obtain the second multimodal feature output by the M6 ​​model.

[0124] Step 3: Use the retrieval module to obtain the feature similarity between the first multimodal feature and each second multimodal feature.

[0125] Among them, the retrieval module can provide a vector retrieval function, and then the first multimodal feature and each second multimodal feature can be input into the retrieval module to perform a similarity calculation operation to obtain the feature similarity between the first multimodal feature output by the retrieval module and each second multimodal feature. The above feature similarity can identify the similarity between the new product and each released product.

[0126] Step 4: Based on the feature similarity between the first multimodal feature and each second multimodal feature, determine at least one similar product corresponding to the new product and the feature similarity corresponding to each similar product in the released product set.

[0127] Among them, quantity information k for limiting similar products is pre-configured, and then the top k similar products with higher feature similarity can be determined in the released product set based on the quantity information k. The second multimodal features and similarities of similar products corresponding to the new product can be reflected in the form of a series:

[0128] Query item = [(doc item1, sim_score1), (doc item2, sim_score2)…(doc_item_k, sim_score_k)], wherein the doc item1 may be the second multimodal feature corresponding to similar product 1, sim_score1 may be the feature similarity between the first multimodal feature and the second multimodal feature of similar product 1, the doc item2 may be the second multimodal feature corresponding to similar product 2, sim_score2 may be the feature similarity between the first multimodal feature and the second multimodal feature of similar product 2, the doc item k may be the second multimodal feature corresponding to similar product k, and sim_score k may be the feature similarity between the first multimodal feature and the second multimodal feature of similar product k.

[0129] Step 5: Normalize the feature similarities between the first multimodal feature and each second multimodal feature to obtain similarity weight information between each similar product and the new product.

[0130] Step 6: Based on the first-guess statistical features and similar weight information corresponding to each similar product, determine the estimated sequence features of the new product. The estimated sequence features are used to identify the release statistical features of the new product. The release statistical features may include at least one of the following: exposure volume, transaction ratio, and total transaction amount in a preset time period.

[0131] Among them, based on the first-guess statistical features of the goods corresponding to each similar product and the similarity weight information, determining the estimated sequence features of the new product may include: performing a weighted sum operation or a weighted average operation on the first-guess statistical features of the goods corresponding to each similar product based on the similarity weight information corresponding to each similar product to obtain the estimated sequence features of the new product.

[0132] Step 7: Obtain static attribute features, global statistical features, and multimodal features of the new product, and concatenate the static attribute features, global statistical features, multimodal features, and estimated sequence features to obtain concatenated features.

[0133] Among them, static attribute features may include at least one of the following: the category of new products, the store / seller, the service of the store (free shipping, 7-day unconditional return), the original price, the e-commerce platform to which the new products belong, the seller's rating information, etc. The above static attribute features can be obtained through the e-commerce platform to which the new products belong. Global statistical features may include merchant dimension statistical features and commodity dimension statistical features. Merchant dimension statistical features may include at least one of the following: exposure in the past 30 days, click-through rate, conversion rate, price level information in the same category, etc. It should be noted that global statistical features do not include statistical features in the first guess scenario. The statistical features in the first guess scenario can be reflected through estimated sequence features.

[0134] Step 8: Obtain a deep learning model and input the concatenated features into the deep learning model for analysis and processing to identify whether the new product is a high-quality product that needs to be recommended in the first guess scenario.

[0135] Among them, the deep learning model can be a network model implemented by a (Multi-gate Mixture-of-Experts, MMoe for short) framework. The deep learning model can learn the performance information of the product in the growth target scenario of the cold start link, the long-term traffic acquisition scenario and the long-term transaction scenario. Specifically, the deep learning model can obtain the first potential of the new product in obtaining traffic in the long term (which can be expressed by a score, which can become the first score), the second potential of the new product to become a transaction target, and the third potential of the new product to become a successfully incubated product in the cold start link by analyzing and processing the spliced ​​notifications. Then, based on the first potential, the second potential and the third potential, the target potential of the new product to become a high-quality product recommended in the first guess scenario can be determined, and then based on the target potential, it can be identified whether the new product is a high-quality product that needs to be recommended in the first guess scenario.

[0136] Specifically, when the number of new products is one, identifying whether the new product is a high-quality product that needs to be recommended in the first-guess scenario based on target potential may include: when the target potential is greater than or equal to a preset threshold, it means that the probability of the new product becoming a high-quality product is relatively high, and the new product can be directly determined as a high-quality product that can be recommended in the first-guess scenario; when the target potential is less than the preset threshold, it means that the probability of the new product becoming a high-quality product is relatively low, and the new product can be directly determined as a high-quality product that cannot be recommended in the first-guess scenario.

[0137] When there are multiple new products, identifying whether the new products are high-quality products that need to be recommended in the first-guess scenario based on target potential may include: obtaining quantity information k used to limit high-quality products; sorting the multiple new products based on the target potential of each new product to obtain product sorting information; and selecting the first k new products from the multiple new products as high-quality products to be recommended in the first-guess scenario based on the product sorting information, thereby effectively ensuring the accuracy and reliability of determining the high-quality products.

[0138] The technical solution provided by this application embodiment obtains the first multimodal features of new products through the M6 ​​model, which can produce high-quality multimodal representations of products, and then searches for similar products corresponding to the new products based on the first multimodal features, determines the first-guess statistical features of similar products, and aggregates the historical performances of similar products (first-guess statistical features of products), so as to obtain the predicted sequence features of new products in future scenarios, and then can identify whether the new products are high-quality products recommended in the first-guess scenario based on the predicted sequence features as well as static attribute features, global statistical features, and multimodal features, thereby greatly improving the accuracy of analyzing and identifying whether new products are high-quality products, and improving the personalization ability of the deep learning model and the perception ability of the back-link, which is conducive to ensuring the harmonious development of the product pool ecology in the first-guess scenario, thereby improving the practicality of the method.

[0139] Figure 7 A flowchart of a method for identifying recommended products provided by an embodiment of the present invention; see attached Figure 7 As shown, this embodiment provides a method for identifying recommended products. The execution subject of the method is a device for identifying recommended products. Specifically, the device for identifying recommended products can be implemented as software or a combination of software and hardware. When the device for identifying recommended products is implemented as hardware, it can be various electronic devices that can implement the identification operation of recommended products, including but not limited to personal computers, servers, databases, etc. When the device for identifying recommended products is implemented as software, it can be installed in the electronic devices listed above. Based on the above-mentioned device for identifying recommended products, the identification operation of recommended products can be implemented. Specifically, the method for identifying recommended products can include:

[0140] Step S701: obtaining a set of commodities to be released and a set of released commodities, wherein the set of commodities to be released includes a plurality of commodities to be released, and the set of released commodities includes a plurality of released commodities.

[0141] Step S702: based on the set of released products, determining the estimated release statistical features corresponding to the product to be released.

[0142] Step S703: Determine the static attribute characteristics corresponding to the product to be released.

[0143] Among them, the specific implementation method and implementation effect of the above steps S701 to S703 in this embodiment are similar to the specific implementation method and implementation effect of steps S201 to S203 in the above embodiment. Please refer to the above statements for details, and no further details will be given here.

[0144] Step S704: determining at least one recommended product in the set of products to be released based on the static attribute characteristics and estimated release statistical characteristics corresponding to the products to be released.

[0145] After obtaining the static attribute characteristics and estimated release statistical characteristics corresponding to the to-be-released product, the static attribute characteristics and estimated release statistical characteristics corresponding to the to-be-released product can be analyzed and processed, so as to determine at least one recommended product in the to-be-released product set, thereby effectively achieving the accurate and reliable determination of at least one recommended product.

[0146] The method for identifying recommended products provided in the present embodiment obtains a set of products to be released and a set of released products, determines the estimated release statistical features corresponding to the products to be released based on the set of released products, and determines the static attribute features corresponding to the products to be released, and then determines at least one recommended product in the set of products to be released based on the static attribute features and the estimated release statistical features corresponding to the products to be released. Since the estimated release statistical features can characterize the future performance of the products to be released, when the static attribute features and the estimated release statistical features are combined to analyze and identify the products to be released, the accuracy and reliability of identifying the recommended products is effectively guaranteed, thereby further improving the practicality of the method.

[0147] Figure 8 A schematic diagram of a device for identifying recommended products provided by an embodiment of the present invention; Figure 8 As shown, this embodiment provides a recommended commodity identification device, which is used to perform the above Figure 2 The method for identifying recommended products shown in the figure, specifically, the identification device for the recommended products may include:

[0148] A first acquisition module 11 is used to acquire information of a product to be released and a set of released products, wherein the set of released products includes multiple released products and product release information corresponding to the released products;

[0149] A first determination module 12, configured to determine estimated release statistical features of the to-be-released product based on the released product set;

[0150] The first determination module 12 is further used to determine the static attribute characteristics of the product to be released based on the information to be released;

[0151] The first processing module 13 is used to identify whether the product to be released is a recommended product based on static attribute features and estimated release statistical features.

[0152] In some instances, when the first determination module 12 determines the estimated release statistical features of the product to be released based on the set of released products, the first determination module 12 is used to perform: determining at least one similar product corresponding to the product to be released in the set of released products; determining the release statistical features corresponding to at least one similar product based on the product release information corresponding to the released product; and determining the estimated release statistical features of the product to be released based on the release statistical features corresponding to at least one similar product.

[0153] In some instances, when the first determination module 12 determines the estimated release statistical characteristics of the commodity to be released based on the release statistical characteristics corresponding to at least one similar commodity, the first determination module 12 is used to execute: obtaining weight information corresponding to the release statistical characteristics; determining the estimated release statistical characteristics of the commodity to be released based on the release statistical characteristics corresponding to at least one similar commodity and the weight information corresponding to the release statistical characteristics.

[0154] In some instances, when the first determination module 12 obtains the weight information corresponding to the publication statistical feature, the first determination module 12 is used to execute: obtaining the similarity between the similar products and the products to be published; and determining the weight information corresponding to the publication statistical feature based on the similarity between the similar products and the products to be published.

[0155] In some instances, when the first determination module 12 determines at least one similar product corresponding to the product to be released in the set of released products, the first determination module 12 is used to execute: obtaining a first multimodal feature of the product to be released, and a second multimodal feature corresponding to each of a plurality of released products in the set of released products; and determining at least one similar product corresponding to the product to be released in the set of released products based on the first multimodal feature and the plurality of second multimodal features.

[0156] In some instances, when the first determination module 12 determines at least one similar product corresponding to the product to be released in the set of released products based on the first multimodal feature and the multiple second multimodal features, the first determination module 12 is used to perform: determining the similarity between the released product and the product to be released based on the first multimodal feature and the multiple second multimodal features; and determining at least one similar product corresponding to the product to be released in the set of released products based on the similarity between the released product and the product to be released.

[0157] In some instances, when the first determination module 12 determines at least one similar product corresponding to the product to be released in the set of released products based on the similarity between the released products and the product to be released, the first determination module 12 is used to execute: sorting the released products based on the similarity between the released products and the product to be released to obtain product sorting information; determining quantity information for limiting similar products; and determining at least one similar product corresponding to the product to be released in the set of released products based on the quantity information and the product sorting information.

[0158] In some instances, when the first processing module 13 identifies whether the product to be released is a recommended product based on static attribute features and estimated release statistical features, the first processing module 13 is used to execute: obtaining global statistical features and multimodal features of the product to be released; identifying whether the product to be released is a recommended product based on static attribute features, estimated release statistical features, global statistical features and multimodal features.

[0159] In some instances, when the first processing module 13 identifies whether a product to be released is a recommended product based on static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics, the first processing module 13 is used to execute: determining potential information of the product to be released to become a recommended product based on the static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics; and identifying whether the product to be released is a recommended product based on the potential information.

[0160] In some instances, when the first processing module 13 determines the potential information of the product to be released to become a recommended product based on static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics, the first processing module 13 is used to execute: determining the first potential of the product to be released to obtain traffic, and the second potential of the product to be released to become a transaction target based on the static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics; determining the potential information of the product to be released to become a recommended product based on the first potential and the second potential.

[0161] In some instances, when the first processing module 13 determines the potential information of the product to be released to become a recommended product based on the first potential and the second potential, the first processing module 13 is used to execute: determining the third potential of the product to be released to become a successfully incubated product based on static attribute characteristics, estimated release statistical characteristics, global statistical characteristics and multimodal characteristics; determining the potential information of the product to be released to become a recommended product based on the first potential, the second potential and the third potential.

[0162] Figure 8 The device shown can perform Figure 1-Figure 6For the method of the embodiment shown in the figure, the part not described in detail in this embodiment can be referred to Figure 1-Figure 6 The implementation process and technical effects of this technical solution refer to Figure 1-Figure 6 The description in the illustrated embodiment will not be repeated here.

[0163] In one possible design, Figure 8 The structure of the identification device for recommended products shown in the figure can be implemented as an electronic device, which can be a controller, a personal computer, a server, or other devices. Fig. 9 As shown, the electronic device may include: a first processor 21 and a first memory 22. The first memory 22 is used to store the corresponding electronic device to execute the above Figure 2-Figure 6 The program of the method for identifying recommended commodities provided in the illustrated embodiment, the first processor 21 is configured to execute the program stored in the first memory 22 .

[0164] The program includes one or more computer instructions, wherein the one or more computer instructions can implement the following steps when executed by the first processor 21: obtaining the to-be-released information of the to-be-released product and a set of released products, wherein the set of released products includes multiple released products and product release information corresponding to the released products; determining the estimated release statistical features of the to-be-released product based on the set of released products; determining the static attribute features of the to-be-released product based on the to-be-released information; and identifying whether the to-be-released product is a recommended product based on the static attribute features and the estimated release statistical features.

[0165] Furthermore, the first processor 21 is also used to execute the aforementioned Figure 2-Figure 6 All or part of the steps in the illustrated embodiments.

[0166] The structure of the electronic device may further include a first communication interface 23 for the electronic device to communicate with other devices or a communication network.

[0167] In addition, an embodiment of the present invention provides a computer storage medium for storing computer software instructions used by electronic devices, which includes instructions for executing the above Figure 2-Figure 6 The procedure involved in the method for identifying recommended products in the illustrated embodiment.

[0168] In addition, an embodiment of the present invention provides a computer program product, including: a computer-readable storage medium storing computer instructions, when the computer instructions are executed by one or more processors, causing the one or more processors to execute the above Figure 2-Figure 6 The steps in the method for identifying recommended commodities in the method embodiment are shown.

[0169] Fig.10A schematic diagram of a device for identifying recommended products provided by an embodiment of the present invention; Fig.10 As shown, this embodiment provides a device for identifying recommended products, and the generating device is used to perform the above Figure 7 The method for identifying recommended products shown in the figure, specifically, the identification device for the recommended products may include:

[0170] The second acquisition module 31 is used to acquire a set of commodities to be released and a set of released commodities, wherein the set of commodities to be released includes a plurality of commodities to be released and the corresponding information to be released of the commodities to be released, and the set of released commodities includes a plurality of released commodities and the commodity release information corresponding to the released commodities;

[0171] A second determination module 32 is used to determine the estimated release statistical features corresponding to the to-be-released product based on the released product set;

[0172] The second determining module 32 is further used to determine the static attribute characteristics corresponding to the to-be-released product based on the to-be-released information corresponding to the to-be-released product;

[0173] The second processing module 33 is used to determine at least one recommended product in the set of products to be released based on the static attribute characteristics and estimated release statistical characteristics corresponding to the products to be released.

[0174] Fig.10 The device shown can perform Figure 7 For the method of the embodiment shown in the figure, the part not described in detail in this embodiment can be referred to Figure 7 The implementation process and technical effects of this technical solution refer to Figure 7 The description in the illustrated embodiment will not be repeated here.

[0175] In one possible design, Fig.10 The structure of the identification device for recommended products shown in the figure can be implemented as an electronic device, which can be a controller, a personal computer, a server, or other devices. Fig.11 As shown, the electronic device may include: a second processor 41 and a second memory 42. The second memory 42 is used to store the corresponding electronic device executing the above Figure 7 The program of the method for identifying recommended commodities provided in the illustrated embodiment, the second processor 41 is configured to execute the program stored in the second memory 42 .

[0176] The program includes one or more computer instructions, wherein the one or more computer instructions can implement the following steps when executed by the second processor 41: obtaining a set of products to be released and a set of released products, wherein the set of products to be released includes multiple products to be released and to-be-released information corresponding to the products to be released, and the set of released products includes multiple released products and product release information corresponding to the released products; based on the set of released products, determining the estimated release statistical features corresponding to the products to be released; based on the to-be-released information corresponding to the products to be released, determining the static attribute features corresponding to the products to be released; based on the static attribute features and the estimated release statistical features corresponding to the products to be released, determining at least one recommended product in the set of products to be released.

[0177] Furthermore, the second processor 41 is also used to execute the aforementioned Figure 7 All or part of the steps in the illustrated embodiments.

[0178] The structure of the electronic device may further include a second communication interface 43 for the electronic device to communicate with other devices or a communication network.

[0179] In addition, an embodiment of the present invention provides a computer storage medium for storing computer software instructions used by electronic devices, which includes instructions for executing the above Figure 7 The procedure involved in the method for identifying recommended products in the illustrated embodiment.

[0180] In addition, an embodiment of the present invention provides a computer program product, including: a computer-readable storage medium storing computer instructions, when the computer instructions are executed by one or more processors, causing the one or more processors to execute the above Figure 7 The steps in the method for identifying recommended commodities in the method embodiment are shown.

[0181] 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 used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present 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 must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0182] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0183] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by adding a necessary general hardware platform, and of course can also be implemented by combining hardware and software. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a computer product, and the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0184] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable device to produce a machine, so that the instructions executed by the processor of the computer or other programmable device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0185] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable device to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable device so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide the functions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps of the functions specified in the blocks or multiple blocks. In a typical configuration, the computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0186] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0187] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying recommended products, characterized in that: include: Acquire information of a product to be released and a set of released products, wherein the set of released products includes multiple released products and product release information corresponding to the released products; Determining estimated release statistical features of the to-be-released product based on the released product set; Based on the information to be released, determining static attribute characteristics of the product to be released; Based on the static attribute features and the estimated release statistical features, it is identified whether the to-be-released product is a recommended product.

2. The method according to claim 1, characterized in that: Determining estimated release statistical features of the to-be-released product based on the released product set includes: Determine, in the set of released products, at least one similar product corresponding to the product to be released; Determine, based on the product release information corresponding to the released product, a release statistical feature corresponding to at least one similar product; Based on the publishing statistical feature corresponding to at least one similar product, the estimated publishing statistical feature of the product to be published is determined.

3. The method according to claim 2, characterized in that Determining the estimated release statistical feature of the product to be released based on the release statistical feature corresponding to at least one similar product, including: Obtaining weight information corresponding to the release statistical features; Based on the publishing statistical feature corresponding to each of at least one similar product and the weight information corresponding to the publishing statistical feature, the estimated publishing statistical feature of the product to be released is determined.

4. The method according to claim 3, characterized in that Obtaining weight information corresponding to the release statistical feature includes: Obtaining the similarity between the similar product and the product to be released; Based on the similarity between the similar products and the product to be released, weight information corresponding to the release statistical feature is determined.

5. The method according to claim 2, characterized in that: Determining at least one similar product corresponding to the product to be released in the set of released products includes: Acquire a first multimodal feature of the to-be-released product and a second multimodal feature corresponding to each of a plurality of released products in the released product set; At least one similar product corresponding to the product to be released is determined in the set of released products based on the first multimodal feature and a plurality of second multimodal features.

6. The method according to claim 5, characterized in that Determining at least one similar product corresponding to the product to be released in the set of released products based on the first multimodal feature and the plurality of second multimodal features includes: Determining the similarity between the published product and the product to be published based on the first multimodal feature and the plurality of second multimodal features; Based on the similarity between the published product and the product to be published, at least one similar product corresponding to the product to be published is determined in the published product set.

7. The method according to claim 6, characterized in that Determining at least one similar product corresponding to the product to be released in the set of released products based on the similarity between the released product and the product to be released, comprising: Sorting the published products based on the similarity between the published products and the products to be published, and obtaining product sorting information; Determining quantity information for limiting the similar products; Based on the quantity information and the commodity ranking information, at least one similar commodity corresponding to the commodity to be released is determined in the set of released commodities.

8. The method according to any one of claims 1 to 7, characterized in that: Based on the static attribute feature and the estimated release statistical feature, identifying whether the to-be-released product is a recommended product includes: Obtaining global statistical features and multimodal features of the product to be released; Based on the static attribute features, the estimated release statistical features, the global statistical features and the multimodal features, it is identified whether the to-be-released product is a recommended product.

9. The method according to claim 8, characterized in that Based on the static attribute feature, the estimated release statistical feature, the global statistical feature and the multimodal feature, identifying whether the to-be-released product is a recommended product includes: Based on the static attribute features, the estimated release statistical features, the global statistical features and the multimodal features, determining the potential information of the to-be-released product to become a recommended product; Based on the potential information, it is identified whether the product to be released is a recommended product.

10. The method according to claim 9, characterized in that Based on the static attribute features, the estimated release statistical features, the global statistical features and the multimodal features, determining the potential information of the to-be-released product to become a recommended product, including: Based on the static attribute features, the estimated release statistical features, the global statistical features and the multimodal features, determining a first potential of the to-be-released product to obtain traffic and a second potential of the to-be-released product to become a transaction target; Based on the first potential and the second potential, potential information of the to-be-released product becoming a recommended product is determined.

11. The method according to claim 10, characterized in that Determining the potential information of the to-be-released product to become a recommended product based on the first potential and the second potential includes: Determining the third potential of the to-be-released product to become a successfully incubated product based on the static attribute features, the estimated release statistical features, the global statistical features, and the multimodal features; Based on the first potential, the second potential and the third potential, potential information of the to-be-released product becoming a recommended product is determined.

12. A method for identifying recommended products, characterized in that: include: Acquire a set of products to be released and a set of released products, wherein the set of products to be released includes a plurality of products to be released and to-be-released information corresponding to the products to be released, and the set of released products includes a plurality of released products and product release information corresponding to the released products; Based on the set of released products, determining estimated release statistical features corresponding to the to-be-released products; Determining static attribute features corresponding to the product to be released based on the to-be-released information corresponding to the product to be released; At least one recommended product is determined in the set of products to be released based on the static attribute characteristics and the estimated release statistical characteristics corresponding to the products to be released.

13. A device for identifying recommended products, characterized in that: include: A first acquisition module is used to acquire information of a to-be-released commodity and a set of released commodities, wherein the set of released commodities includes a plurality of released commodities and commodity release information corresponding to the released commodities; A first determination module, configured to determine estimated release statistical features of the to-be-released product based on the released product set; The first determination module is further used to determine the static attribute characteristics of the commodity to be released based on the information to be released; The first processing module is used to identify whether the product to be released is a recommended product based on the static attribute characteristics and the estimated release statistical characteristics.

14. A device for identifying recommended products, characterized in that: include: A second acquisition module is used to acquire a set of commodities to be released and a set of released commodities, wherein the set of commodities to be released includes a plurality of commodities to be released and the corresponding information to be released of the commodities to be released, and the set of released commodities includes a plurality of released commodities and the commodity release information corresponding to the released commodities; A second determination module, configured to determine, based on the set of released products, estimated release statistical features corresponding to the product to be released; The second determination module is further configured to determine the static attribute characteristics corresponding to the to-be-released product based on the to-be-released information corresponding to the to-be-released product; The second processing module is used to determine at least one recommended product in the set of products to be released based on the static attribute characteristics and the estimated release statistical characteristics corresponding to the products to be released.

15. An electronic device, characterized in that: include: A memory, a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions, when executed by the processor, implement the method as described in any one of claims 1-12.

16. A computer storage medium, characterized in that: Used to store a computer program, which enables a computer to implement the method of any one of claims 1 to 12 when executed.

17. A computer program product, characterized in that include: A computer program, when executed by a processor of an electronic device, causes the processor to execute the steps of any one of the methods of claims 1 to 12.

Images