Product recommendation methods, devices, electronic devices and storage media

Abstract

This disclosure provides a method, apparatus, electronic device, and storage medium for recommending products on an internet platform, relating to the field of computer technology, and particularly to the field of data processing. The implementation scheme simultaneously considers the first characteristic associated with the buyer, the second characteristic associated with the seller, and the third characteristic associated with the product during the stages of database creation, recall, and ranking of products to be recommended. This better balances the interests of the platform, buyers, and sellers, rationally allocates traffic distributed by the platform, and enhances the value of platform traffic.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more particularly to the field of data processing, specifically to a method, apparatus, electronic device, computer-readable storage medium, and computer program product for a product recommendation method. Background Technology

[0002] With the development of computer and internet technology, transactions between consumers and businesses on internet platforms are becoming increasingly frequent. An internet platform is a platform for buyers and sellers to exchange products, services, and information via the internet. Depending on the target buyer, it is divided into B2C (Business to Customer) internet platforms for consumers and B2B (Business to Business) internet platforms for businesses.

[0003] Product recommendations are common on e-commerce platforms, and the traffic they drive can even account for half of the platform's total traffic. Product recommendations are significant for platforms, buyers, and sellers. For platforms, product recommendation traffic, as a flexible source of traffic, is unaffected by relevance or other factors, allowing them to fully leverage their value and becoming one of the most flexibly allocated traffic resources. For buyers, product recommendations improve sourcing efficiency and satisfy the need for price comparison across multiple sources. For sellers, product recommendations provide more traffic support and product launch incentives, thereby improving product sales efficiency.

[0004] Existing product recommendation systems on these platforms prioritize user experience, focusing primarily on the diversity, novelty, surprise factor, and accuracy of recommendations from the buyer's perspective. However, they neglect the seller's experience, making it difficult to meet the needs of all parties and hindering the healthy development of the platform.

[0005] The methods described in this section are not necessarily methods that had been previously conceived or adopted. Unless otherwise specified, no method described in this section should be assumed to be prior art simply because it is included in this section. Similarly, unless otherwise specified, the issues mentioned in this section should not be considered to be accepted in any prior art. Summary of the Invention

[0006] This disclosure provides a product recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product.

[0007] According to one aspect of this disclosure, a method for recommending products on an internet platform is provided, comprising: constructing a product library to be recommended based on a first feature associated with a buyer, a second feature associated with a seller, and a third feature associated with a product on the internet platform; determining products to be recommended from the product library to be recommended; and sorting the products to be recommended to recommend the sorted products to the buyer, wherein the sorting is performed according to at least one of a first sorting strategy based on commercial domain traffic and a second sorting strategy based on non-commercial domain traffic.

[0008] According to another aspect of this disclosure, an internet platform product recommendation device is provided, comprising: a database construction module that constructs a product database to be recommended based on a first feature associated with a buyer, a second feature associated with a seller, and a third feature associated with a product on the internet platform; a recall module configured to determine products to be recommended from the product database and recall them; and a sorting module configured to sort the products to be recommended in order to recommend the sorted products to the buyer, wherein the sorting module includes a first sorting strategy execution module and a second sorting strategy execution module, the first sorting strategy execution module being configured to sort according to a first sorting strategy based on commercial domain traffic, and the second sorting strategy execution module being configured to sort according to a second sorting strategy based on non-commercial domain traffic.

[0009] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the Internet platform product recommendation method.

[0010] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to execute the product recommendation method of the Internet platform.

[0011] According to another aspect of this disclosure, a computer program product is provided, including a computer program, wherein the computer program, when executed by a processor, implements the Internet platform product recommendation method.

[0012] According to another aspect of this disclosure, one or more embodiments of this disclosure are provided that can balance the platform experience for both buyers and sellers and improve the traffic value of the platform.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0014] The accompanying drawings exemplify embodiments and form part of the specification, serving together with the textual description to explain exemplary implementations of the embodiments. The illustrated embodiments are for illustrative purposes only and do not limit the scope of the claims. Throughout the drawings, the same reference numerals refer to similar but not necessarily identical elements.

[0015] Figure 1 A schematic diagram of an exemplary system in which the various methods described herein may be implemented according to embodiments of the present disclosure is shown;

[0016] Figure 2 A flowchart of a product recommendation method according to an embodiment of the present disclosure is shown;

[0017] Figure 3 A flowchart illustrating the steps of calculating a sorting factor according to an embodiment of the present disclosure is shown;

[0018] Figure 4 A flowchart illustrating the steps of constructing a library of recommended products according to an embodiment of this disclosure is shown;

[0019] Figure 5 A scenario diagram is shown that can implement the product recommendation method according to embodiments of the present disclosure;

[0020] Figure 6 A structural block diagram of a product recommendation device according to an embodiment of the present disclosure is shown;

[0021] Figure 7 A structural block diagram of a product recommendation device according to another embodiment of the present disclosure is shown;

[0022] Figure 8 A structural block diagram of a second operation module according to another embodiment of the present disclosure is shown;

[0023] Figure 9 A structural block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure is shown. Detailed Implementation

[0024] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0025] In this disclosure, unless otherwise stated, the use of terms such as "first," "second," etc., to describe various elements is not intended to limit the positional, temporal, or importance relationships of these elements; such terms are merely used to distinguish one element from another. In some examples, the first element and the second element may refer to the same instance of that element, while in other cases, based on the context, they may refer to different instances.

[0026] The terminology used in the description of the various examples described in this disclosure is for the purpose of describing particular examples only and is not intended to be limiting. Unless the context explicitly indicates otherwise, an element may be one or more unless the number of elements is specifically limited. Furthermore, the term "and / or" as used in this disclosure covers any one of the listed items and all possible combinations thereof.

[0027] In product recommendation technologies, the essence is to revitalize secondary traffic, providing more transaction scenarios and spaces for both buyers and sellers, and greatly improving the platform experience for both parties. Existing B2C platform recommendation strategies are based on the effective matching of the relationship between people, goods, and places. The platform recommends products that meet users' needs by mining the characteristics of people, goods, and places. Therefore, existing platforms prioritize user experience when designing product recommendation systems, evaluating the diversity, novelty, surprise, and accuracy of recommendation results solely from a user experience perspective. However, sellers, as participants in online transactions on the platform, are essential for achieving a closed-loop transaction system, and their user experience on the platform cannot be ignored. Therefore, it is crucial to consider the interests of buyers, sellers, and the platform simultaneously in product recommendation methods.

[0028] To address the aforementioned technical problems, an embodiment of this disclosure provides a product recommendation method.

[0029] Before describing the method according to embodiments of this disclosure in detail, firstly, in conjunction with Figure 1 The description includes exemplary systems in which the various methods and apparatuses described herein may be implemented.

[0030] Figure 1 A schematic diagram of an exemplary system 100 in which the various methods and apparatus described herein can be implemented according to embodiments of this disclosure is shown. Reference Figure 1 The system 100 includes one or more client devices 101, 102, 103, 104, 105 and 106, a server 120, and one or more communication networks 110 coupling the one or more client devices to the server 120. The client devices 101, 102, 103, 104, 105 and 106 can be configured to execute one or more applications.

[0031] In embodiments of this disclosure, server 120 may run one or more services or software applications that enable the execution of methods for recommending products on an internet platform.

[0032] In some embodiments, server 120 may also provide other services or software applications, which may include non-virtual and virtual environments. In some embodiments, these services may be provided as web-based services or cloud services, such as to users of client devices 101, 102, 103, 104, 105, and / or 106 under a Software as a Service (SaaS) model.

[0033] exist Figure 1 In the configuration shown, server 120 may include one or more components that implement the functions performed by server 120. These components may include software components, hardware components, or combinations thereof that can be executed by one or more processors. Users operating client devices 101, 102, 103, 104, 105, and / or 106 can sequentially interact with server 120 using one or more client applications to utilize the services provided by these components. It should be understood that various different system configurations are possible and may differ from system 100. Therefore, Figure 1 This is an example of a system used to implement the various methods described herein, and is not intended to be limiting.

[0034] Users can use client devices 101, 102, 103, 104, 105, and / or 106 to receive and browse product information recommended by the internet platform. The client devices can provide an interface that allows users to interact with the client devices. The client devices can also output information to users through this interface. Although... Figure 1 Only six client devices are described, but those skilled in the art will understand that this disclosure can support any number of client devices.

[0035] Client devices 101, 102, 103, 104, 105, and / or 106 may include various types of computer devices, such as portable handheld devices, general-purpose computers (such as personal computers and laptops), workstation computers, wearable devices, smart screen devices, self-service terminal devices, service robots, gaming systems, thin clients, various messaging devices, sensors, or other sensing devices. These computer devices can run various types and versions of software applications and operating systems, such as Microsoft Windows, Apple iOS, UNIX-like operating systems, Linux or Linux-like operating systems (such as Google Chrome OS); or include various mobile operating systems, such as Microsoft Windows Mobile OS, iOS, Windows Phone, and Android. Portable handheld devices may include cellular phones, smartphones, tablets, personal digital assistants (PDAs), etc. Wearable devices may include head-mounted displays (such as smart glasses) and other devices. Gaming systems may include various handheld gaming devices, internet-enabled gaming devices, etc. Client devices are capable of executing various applications, such as various internet-related applications, communication applications (such as email applications), short message service (SMS) applications, and can use various communication protocols.

[0036] Network 110 can be any type of network well known to those skilled in the art, and can use any of a variety of available protocols (including but not limited to TCP / IP, SNA, IPX, etc.) to support data communication. By way of example only, one or more networks 110 can be a local area network (LAN), an Ethernet-based network, a token ring network, a wide area network (WAN), the Internet, a virtual network, a virtual private network (VPN), an intranet, an extranet, a blockchain network, a public switched telephone network (PSTN), an infrared network, a wireless network (e.g., Bluetooth, WIFI), and / or any combination of these and / or other networks.

[0037] Server 120 may include one or more general-purpose computers, special-purpose server computers (e.g., PC (personal computer) servers, UNIX servers, mid-range servers), blade servers, mainframe computers, server clusters, or any other suitable arrangement and / or combination. Server 120 may include one or more virtual machines running a virtual operating system, or other computing architectures involving virtualization (e.g., one or more flexible pools of logical storage devices that can be virtualized to maintain virtual storage devices for servers). In various embodiments, server 120 may run one or more services or software applications that provide the functionality described below.

[0038] The computing unit in server 120 can run one or more operating systems, including any of the aforementioned operating systems and any commercially available server operating system. Server 120 can also run any of a variety of additional server applications and / or middleware applications, including HTTP servers, FTP servers, CGI servers, JAVA servers, database servers, etc.

[0039] In some implementations, server 120 may include one or more applications to analyze and merge data feeds and / or event updates received from users of client devices 101, 102, 103, 104, 105 and / or 106. Server 120 may also include one or more applications to display data feeds and / or real-time events via one or more display devices of client devices 101, 102, 103, 104, 105 and / or 106.

[0040] In some implementations, server 120 can be a server for a distributed system or a server integrated with blockchain. Server 120 can also be a cloud server, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology. A cloud server is a host product in the cloud computing service system, designed to address the shortcomings of traditional physical hosts and Virtual Private Server (VPS) services, such as high management difficulty and weak business scalability.

[0041] System 100 may also include one or more databases 130. In some embodiments, these databases may be used to store data and other information. For example, one or more of the databases 130 may be used to store information such as audio files and video files. Databases 130 may reside in various locations. For example, a database used by server 120 may be local to server 120, or it may be located away from server 120 and may communicate with server 120 via a network-based or dedicated connection. Databases 130 may be of different types. In some embodiments, the database used by server 120 may be, for example, a relational database. One or more of these databases may store, update, and retrieve data from and from the databases in response to commands.

[0042] In some embodiments, one or more of the databases 130 may also be used by an application to store application data. The databases used by the application may be of different types, such as key-value stores, object stores, or regular stores supported by a file system.

[0043] Figure 1The system 100 can be configured and operated in various ways to enable the application of the various methods and apparatus described in this disclosure.

[0044] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0045] Figure 2 A flowchart of a product recommendation method according to an embodiment of the present disclosure is shown.

[0046] refer to Figure 2 The embodiments of this disclosure provide an internet platform product recommendation method 200, including the following steps:

[0047] Step S201: Construct a database of products to be recommended based on the first feature associated with buyers, the second feature associated with sellers, and the third feature associated with products on the internet platform;

[0048] Step S202: Determine the products to be recommended from the product pool;

[0049] Step S203: Sort the products to be recommended to recommend the sorted products to the buyer, wherein the sorting is performed according to at least one of a first sorting strategy based on commercial domain traffic and a second sorting strategy based on non-commercial domain traffic.

[0050] The first feature is the characteristic associated with the buyer. In the example, the first feature may include buyer click preferences, buyer location, historical search records, and relationships between different buyers. By using the first feature, a buyer profile can be created, analyzing the buyer's shopping habits and preferences. This allows for more targeted product recommendations to meet the buyer's purchasing needs and promote transactions.

[0051] The second feature is seller-related. In the example, this could include the seller's membership level, store level, transaction badge level, the current period of the business, and the seller's paid advertising activity. This second feature allows analysis of the seller's privilege level, willingness to pay, ability to pay, and potential demand. This enables the development of targeted support strategies when setting product recommendation policies, leading to more precise and efficient platform traffic control, improved user experience for sellers, and ultimately, greater benefits for the platform.

[0052] In the example, a seller's privilege level can be determined based on paid benefits such as seller membership level, paid advertising level, and seller store level and transaction badge level. A seller's membership level is determined by the fee paid when joining the platform, including free membership, regular paid membership, and premium paid membership. Paid advertising level is determined by the type of advertising purchased for the product, such as external and internal advertising. External advertising can include bidding ads and hidden ads, while internal advertising refers to product promotion ads. Store level can be determined by the platform based on factors such as product quality, service quality, and operational quality. Transaction badge level can be determined by the platform based on order volume, number of traded products, and transaction service performance. Generally, the higher the seller's privilege level, the stronger their willingness and ability to pay when operating on the platform, and the more revenue they can generate for the platform.

[0053] In the example, the merchant's stage can refer to the length of time they have been on the platform and their current phase. Merchants at different stages have different activity patterns and motivations on the platform, requiring targeted guidance and support from the platform. For newly joined merchants, who are often highly active due to the novelty of the platform, their desire for traffic is stronger and more proactive, and they have high expectations for the platform's promotional effects. Therefore, when recommending products, the platform needs to consider the needs of these sellers and provide traffic support in its recommendation strategies to meet their traffic acquisition needs, help them quickly increase sales, and improve their seller experience. On the other hand, for sellers who have been on the platform for some time, their initiative and enthusiasm are often lower. Therefore, the platform needs to develop targeted strategies to activate their initiative. For example, for sellers whose memberships are about to expire, the platform can proactively provide traffic support to enhance their goodwill towards the platform and increase their willingness to renew their memberships.

[0054] The third feature is the product-related characteristic. In the example, the third feature includes one or more of the following: product parameters, category, origin, model, quality grade, historical click data, historical conversion data, and relationships between different products. Products are the objects recommended by the platform and are the carriers of interests for buyers, sellers, and the platform itself. Pushing suitable products to buyers in the right way is an important means of promoting transactions on the platform.

[0055] The rationality of the product recommendation sorting process directly affects the buyer's shopping experience on the platform, which in turn affects the buyer's subsequent clicks and conversions. Clicks and conversions are crucial steps for completing a transaction on an internet platform.

[0056] In terms of product recommendation ranking, theoretically, the higher a product ranks, the more exposure it receives. More exposure means a higher probability of clicks and conversions. Therefore, top-ranked positions are relatively scarce, and placing products that generate more revenue for the platform in these positions maximizes platform profits and effectively enhances traffic value. However, at the same time, top-ranked products have a stronger visual impact on buyers, directly affecting their platform experience. For example, if the first recommended product is exactly what the buyer wants, it greatly increases their positive impression of the platform. Conversely, if the top-ranked products are far from the buyer's expectations, the buyer may perceive the platform's recommendations as meaningless and even develop a negative attitude towards the platform's recommendation process. Furthermore, the value each recommended product can create for the platform varies. The product a buyer most wants might generate almost no revenue for the platform, resulting in low click and conversion value. Therefore, when ranking recommended products, it's insufficient to only consider buyer experience; the potential value each product brings to the platform must also be considered.

[0057] Commercial traffic refers to traffic that a platform allocates from public traffic, primarily through a paid subscription model. Commercial traffic directly generates revenue for the platform and is a crucial means of monetizing traffic. Non-commercial traffic, on the other hand, attracts more sellers, increases platform activity, and contributes to the platform's long-term sustainable development. When ranking products, different ranking strategies are applied to products from commercial and non-commercial traffic respectively. This effectively balances platform revenue and traffic sustainability, promoting the platform's healthy development. Therefore, given the limited ranking positions, it is essential and important to determine how to rationally and efficiently allocate traffic by properly ranking recommended products.

[0058] According to embodiments of this disclosure, when constructing the product library to be recommended, a first feature associated with the buyer, a second feature associated with the seller, and a third feature associated with the product are considered simultaneously. Furthermore, when sorting products, the platform can adopt at least one of a first sorting strategy based on commercial traffic and a second sorting strategy based on non-commercial traffic, depending on current needs and key support directions. This ensures that the recommended products fully understand seller needs while considering buyer experience, balancing the interests of buyers, sellers, and the platform. It also helps to accurately implement the platform's traffic control strategy, making the platform's traffic allocation more reasonable and increasing the value of the platform's traffic.

[0059] In some embodiments, the first ranking strategy may include ranking the products to be recommended based on their equity value, and the second ranking strategy may include calculating the ranking factors of the products to be recommended and ranking them according to the ranking factors.

[0060] The equity value of a product to be recommended can refer to the benefits the platform can obtain by promoting the product. In this embodiment, the first ranking strategy, which ranks products based on their equity value, prioritizes those that can bring more benefits to the platform, thereby increasing the platform's revenue. For non-commercial traffic, a ranking factor can be calculated for each product, and the products can be ranked according to the magnitude of this factor. Ranking based on the set ranking factor can effectively implement the platform's traffic control strategy.

[0061] The second ranking strategy sorts the products to be recommended based on the ranking factors. The ranking factors are mainly calculated by considering the click expectation coefficient (ctr), the conversion expectation coefficient (ct(cvr)), the click value (F1), and the conversion value (F2).

[0062] The click expectation coefficient (ctr) and conversion expectation coefficient (ct(cvr)) represent the probability that a product will be clicked and converted after being recommended. The click value (F1) and conversion value (F2) represent the impact of the product's clicks and conversions on the platform's value. For internet platforms, completing a closed loop of transactions is the most important goal. Successful transactions not only solve the practical needs of buyers and sellers on the platform but also increase their stickiness and boost platform traffic, forming the foundation for the platform's development and growth. From the platform's perspective, it aims to maximize the click-through rate and conversion rate of recommended products while also ensuring that each clicked and converted product generates as much revenue as possible. This allows the platform to maximize revenue while satisfying the needs of both buyers and sellers. Therefore, in this embodiment, the calculation of the ranking factor in the second ranking strategy needs to consider not only the click expectation coefficient (ctr) and conversion expectation coefficient (ct(cvr)) but also the product's click value (F1) and conversion value (F2). This ensures that the ranking result based on the second ranking strategy balances the interests of buyers, sellers, and the platform, resulting in a more reasonable allocation of platform traffic.

[0063] In some embodiments, the equity value of a product to be recommended may include at least one of ECPM (effective cost per mile, which refers to the advertising revenue that can be obtained per thousand impressions), advertising value, and product promotion quality.

[0064] In the example, the first ranking strategy prioritizes sorting by the ECPM value of the products to be recommended, followed by sorting by advertising value, and finally sorting by product promotion quality.

[0065] ECPM (Economic Value Per Mille) is an important indicator of a website's profitability. A higher ECPM indicates a higher level of recognition of the platform's value, and thus a higher value for the product's equity, making it a crucial factor to consider in ranking. Platform advertising value can be categorized by different ad types. In this example, on-site advertising is more valuable than off-site advertising. On-site advertising refers to ads placed by merchants on the platform to promote their products. Since both the advertiser and the product are on the platform, these ads have a very positive impact on the platform's transaction volume. Therefore, in this embodiment, the value of on-site advertising can be higher than that of off-site advertising. Off-site advertising can include auction ads and hidden ads. Since auction ads typically involve bidding for relatively scarce ad slots, with the highest bidder winning, in this example, the advertising value of auction ads is higher than that of hidden ads. Furthermore, the equity value of a product can also be reflected by the quality of its promotion. The quality of product promotion can be determined by whether the product and its seller meet the platform's support policies, thus helping the platform achieve more effective traffic control.

[0066] The first sorting strategy in this embodiment can give certain preference to products that conform to the platform's regulation or support strategies when sorting, while ensuring the platform's revenue. This allows the platform's regulation or support strategies to be better implemented, which helps the platform to play its regulatory role more effectively, making traffic allocation more reasonable and increasing traffic value.

[0067] In some embodiments, such as Figure 3 As shown, step 300 of calculating the ranking factor of the product to be recommended may include:

[0068] Step S301: Determine the expected click coefficient and expected conversion coefficient of the product to be recommended based on the historical data of the product and the first feature associated with the buyer;

[0069] Step S302: Based on the second characteristic associated with the seller to which the product to be recommended belongs and the platform traffic data, determine the click value and conversion value of the product to be recommended, wherein the platform traffic data includes click traffic data, search traffic conversion data, and recommendation traffic conversion data; and

[0070] Step S303: Determine the ranking factors for the products to be recommended based on the expected click coefficient, click value, expected conversion coefficient, and conversion value.

[0071] In the example, the historical data of the product to be recommended may include: a third feature associated with the product, and a first feature associated with the buyers involved in the clicks and conversions after the product was recommended each time. In this embodiment, the third feature includes the product's parameters, origin, model, quality, key features, historical click data, and historical conversion data. The first feature includes the buyer's regional preferences, brand preferences, price preferences, and click and conversion preferences. A transaction prediction model is established based on historical data to predict potential clicks and conversions between buyers and products. This model can predict different buyers' expected clicks and conversions for the recommended product and output corresponding click expectation coefficients and conversion expectation coefficients. Higher click expectation coefficients and conversion expectation coefficients indicate a higher probability of buyer clicks and conversions. In this embodiment, inputting the first feature associated with the buyer into the transaction prediction model yields the click expectation coefficient (ctr) and the conversion expectation coefficient (ct(cvr)).

[0072] Generally speaking, the operating revenue of internet platforms mainly includes commissions from transaction amounts and rights fees and advertising fees paid by sellers to promote products. Therefore, the seller's ability and willingness to pay have a significant impact on the click value and conversion value that the product can generate after it is recommended.

[0073] Therefore, this embodiment calculates the ranking factor of the recommended product based on the expected click coefficient, click value, expected conversion coefficient, and conversion value. It also considers the probability of subsequent buyer clicks and conversions, as well as the benefits that clicks and conversions can bring to the platform. The ranking factor is calculated as a measurement parameter. The larger the ranking factor, the more benefits the recommended product may bring to the platform. Such ranking results can identify the products that can bring the greatest value to the platform, maximize the platform's interests, and improve the value of the platform's traffic.

[0074] In some embodiments, determining the click value and conversion value of a product to be recommended, based on a second characteristic associated with the seller of the product, a third characteristic associated with the product, and platform traffic data, includes:

[0075] The seller rating coefficient M is determined based on the equity level in the second characteristic associated with the seller of the product to be recommended;

[0076] Determine the platform traffic control coefficient N based on platform traffic data;

[0077] The click value F1 is determined based on the seller rating coefficient M and the platform traffic control coefficient N. The click value F1 is configured as the sum of the seller equity coefficient M and the platform traffic control coefficient N.

[0078] F1 = M + N.

[0079] In the example, the privilege levels include at least one of the following: seller membership level, paid advertising level, seller store level, and transaction badge level.

[0080] In this embodiment, the click value is mainly determined based on the seller level coefficient and the platform traffic control coefficient. The higher the seller level coefficient, the higher the seller's rights level on the platform. Such products generate more revenue for the platform when clicked, making such sellers very important and valuable to the platform, and thus requiring priority consideration in recommendation ranking. The platform traffic control coefficient is a strategy formulated by the platform to control the current distribution of traffic based on its own interests or development needs. The platform traffic control strategy can proactively adjust the traffic allocation for products in different industries and stores during promotion, achieving a more reasonable allocation of traffic and increasing the value of platform traffic. Incorporating it into the click value calculation ensures that the clicks of the currently recommended products can take into account the platform's direct and indirect interests, short-term interests and long-term development, making the platform's traffic distribution more reasonable and promoting the platform's sustainable and healthy development.

[0081] In some embodiments, determining the platform traffic control coefficient N based on platform traffic data includes:

[0082] Determine the store's daily click traffic X based on click traffic data. i X the average daily click-through rate of stores with the same star rating in the same product category on the platform av ;

[0083] The store traffic control coefficient X is determined based on the daily click traffic of the store to which the recommended product belongs and the average daily click traffic of stores with the same star rating in the same product category on the platform.

[0084] The traffic quality coefficient C of the industry to which the recommended product belongs is determined based on search traffic conversion data and recommendation traffic conversion data.

[0085] The platform traffic control coefficient N is determined based on the store traffic control coefficient X and the traffic quality coefficient C of the industry to which the recommended product belongs.

[0086] In this embodiment, the platform traffic control coefficient N includes two aspects: the store traffic control coefficient X and the traffic quality coefficient C of the product's industry. This allows for the allocation of traffic between different industries and stores, thereby enabling the platform's traffic control strategy to be implemented, enhancing the platform's traffic control capabilities and accuracy, and increasing the value of the platform's traffic.

[0087] In some embodiments, the platform traffic control coefficient N satisfies the following conditions:

[0088] Response to the store's daily click traffic X of the product to be recommended i The daily click-through rate exceeded X times the average click-through rate of stores with the same star rating in the same product category on the platform. av In the case of a negative value,

[0089] Response to the store's daily click traffic X of the product to be recommended i The average daily click-through rate was lower than that of stores with the same star rating in the same product category on the platform. av In the case of a positive value,

[0090] Response to the store's daily click traffic X of the product to be recommended i Equals the average daily click-through rate of stores with the same star rating in the same product category on the platform X av In the case of , it is 0.

[0091] In this embodiment, by linking the platform traffic control coefficient with the click traffic of the store to which the product belongs on that day, the traffic of the store to which the recommended product belongs can be adjusted, i.e., traffic limiting or traffic supplementation, so that products of stores with fewer clicks can have a better exposure opportunity. The traffic distribution across the entire platform is controlled to be distributed in accordance with the store's star rating, thereby making the platform's traffic distribution more reasonable.

[0092] In some embodiments, the platform traffic control coefficient N is also configured to be positively correlated with the traffic quality coefficient C of the industry to which the product to be recommended belongs; the traffic quality coefficient C of the industry to which the product to be recommended belongs is the ratio of the conversion coefficient ct(cvr)1 of the search traffic of the industry to which the product to be recommended belongs to the conversion coefficient ct(cvr)2 of the recommended traffic of the industry to which the product to be recommended belongs.

[0093] Platform traffic can include search traffic, recommendation traffic, and other traffic. Since search traffic typically involves buyers actively searching for their desired products, it boasts higher click-through rates and conversion rates. However, because search traffic is passive, its revenue for the platform is actually lower than that of recommendation traffic. In this embodiment, considering the different purchasing habits across industries, to balance traffic distribution across different industries on the platform, the platform traffic control coefficient N is configured to be positively correlated with the traffic quality coefficient C of the industry to which the recommended product belongs. This makes the platform traffic distribution more reasonable. Furthermore, the traffic quality coefficient C is set as the ratio of the conversion coefficient ct(cvr)1 of the search traffic to the industry to which the recommended product belongs to to the conversion coefficient ct(cvr)2 of the recommendation traffic to the industry to which the recommended product belongs, i.e., C = ct(cvr)1 / ct(cvr)2. This allows for adjustments to the traffic quality coefficient based on the different conversion rates of search and recommendation traffic across different industries, and by combining it with the store's click traffic. This ensures that the traffic control objectives are accurately achieved and prevents control measures from becoming ineffective.

[0094] In some embodiments, determining the click value and conversion value of a product to be recommended, based on a second characteristic associated with the seller of the product to be recommended, a third characteristic associated with the product to be recommended, and platform traffic data, further includes:

[0095] The conversion value coefficient is determined based on the average traffic distribution and average conversion traffic across the entire market.

[0096] The product type coefficient is determined based on whether the product type of the product to be recommended is a preset product type.

[0097] The conversion value is determined based on the conversion value coefficient and the product type coefficient.

[0098] In the example, the conversion value coefficient is the ratio of the average traffic distribution to the average conversion traffic across the entire platform. Preset product types can be product types that the platform can independently set as needed to meet the platform's development direction and implement its support strategies. For example, for a B2B platform, its transaction matching methods mainly include: inquiries, telephone communication, categorized consultation, and online transactions. Among these, online transactions can complete the entire transaction matching loop on the platform, which is of great significance for improving platform traffic and user stickiness. Therefore, products that can be traded online are product types that the platform needs to support, and should be given certain support and preferential treatment in recommendation ranking.

[0099] By incorporating a product type coefficient into the conversion value calculation, products that align with the platform's support guidelines receive higher rankings, thus gaining more exposure and increasing the probability of clicks and conversions. This gives the platform greater autonomy in traffic control, resulting in more precise adjustments. It fully leverages the platform's subjective control capabilities, improving the rationality and efficiency of traffic allocation and enhancing traffic value.

[0100] The following example illustrates the method for calculating the ranking factor in the above implementation.

[0101] In the example, the ranking factor K of the product to be recommended is calculated according to the following formula:

[0102] K = ctr × F1 + ct(cvr) × F2,

[0103] Where ctr is the expected click coefficient, F1 is the click value, ct(cvr) is the expected conversion coefficient, and F2 is the conversion value.

[0104] In the example, the click value F1 is calculated according to the following formula:

[0105] F1 = M + N,

[0106] Where M is the seller level coefficient determined based on the rights level in the second feature, and N is the platform traffic control coefficient.

[0107] In the example, the seller rating coefficient M can include the store star rating A. n Membership Level Score D and Transaction Badge Level Score B m Where n represents the store's star rating, and m represents the transaction badge level. The specific formula for calculating the seller level coefficient M is:

[0108] M = A n +B m +D

[0109] In the example, the store's star rating is A. n Membership Level Score D and Transaction Badge Level Score B m The values ​​of are shown in Table 1.

[0110] Table 1 Examples of Seller Rating Coefficient Values

[0111]

[0112] In the example, the platform traffic control coefficient N is calculated according to the following formula:

[0113] N = (1-X) i / X av )×C

[0114] Among them, X i X represents the daily click traffic of the store to which the product to be recommended belongs. av C represents the average click-through rate of stores with the same star rating in the same product category as the product to be recommended on the platform on that day, and C represents the traffic quality coefficient of the industry to which the product to be recommended belongs.

[0115] In the example, X i / X av This represents the ratio of the daily click traffic of the store to which the current product belongs to the average daily click traffic of stores with the same star rating in the same product category on the platform. It can be used to determine whether the daily click traffic of the store to which the current product belongs exceeds or falls below the average click traffic. If it exceeds the average click traffic, then (1-X) i / X av (1-X) is a negative value, and the platform traffic control coefficient N is also a negative value. This will reduce the size of the ranking factor, causing the product to be ranked lower. Conversely, if the store's click traffic for the current product is lower than the average click traffic for the day, then (1-X) will be negative. i / X avThe platform traffic control coefficient N is a positive value, which increases the size of the ranking factor, so that the product will be ranked relatively high in the ranking. If the daily click traffic of the store to which the current product belongs is equal to the average click traffic, then the platform traffic control coefficient N is 0, and no control is needed during the ranking.

[0116] In the example, the conversion value F2 of the product to be recommended is calculated according to the following formula:

[0117] F2 = H + E,

[0118] Where H is the conversion value coefficient, which is the ratio of the average distribution traffic to the average conversion traffic of the market. Alternatively, in the example, H is 1000.

[0119] E represents the product type coefficient, which is determined based on whether the product type to be recommended is a preset product type. If the product type to be recommended is a preset product type, the product type coefficient is 1; otherwise, the product type coefficient is 0.

[0120] Figure 4 A schematic diagram illustrating the steps of constructing a library of recommended products according to an embodiment of the present disclosure is shown.

[0121] In some embodiments, such as Figure 4 As shown, step 400 of constructing the product library to be recommended (for example, combining...) Figure 2 Step S201 may include:

[0122] Step S401: Determine the product category to be recommended based on the first characteristic associated with the buyer;

[0123] Step S402: For products under the product category to be recommended, determine whether the second feature associated with the seller of the product and the third feature associated with the product meet the corresponding database creation conditions;

[0124] Step S403: After deduplicating the products that meet the database creation conditions, add them to the product database to be recommended.

[0125] In the example, the database creation conditions can be selectively set for the characteristics (including the first characteristic, the second characteristic, and the third characteristic) of the product. If the conditions are met, the product will be created as a database for subsequent recommendations to the user.

[0126] In the example, the database creation conditions can be set based on the second characteristics related to the seller of the product. The second characteristics include the seller's membership level and the seller's store level. The corresponding database creation conditions can be set to the seller's membership level reaching a predetermined membership level and the seller's store level reaching a predetermined store level.

[0127] In the example, the platform categorizes sellers' membership levels based on the amount of benefits paid to join the platform: free members, regular paid members, and premium paid members. The database creation criteria can then be set to regular paid members; that is, only sellers whose membership level in the second characteristic related to the product seller is regular paid members or higher will meet the database creation criteria.

[0128] In the example, the platform classifies sellers' store ratings based on a comprehensive assessment of factors such as product quality, service quality, and operational quality, defining them from highest to lowest as five-star stores, four-star stores, three-star stores, two-star stores, and one-star stores.

[0129] The database creation criteria can also include a third characteristic related to the product as a constraint. In the example, the third characteristic includes the product's quality level. Whether the product's quality level meets a preset level is used as a database creation condition. It can be understood that the product's quality level is a rating given by the platform based on factors such as product type, quality parameters, and after-sales service, defined from highest to lowest as five-star, four-star, three-star, two-star, and one-star products.

[0130] In the example, the database creation conditions can be set to restrict any one of the three conditions: merchant membership level, merchant store level, and product quality level. Alternatively, conditions can be set for two or even all three conditions simultaneously. Multi-dimensional restrictions based on different conditions can promote products that better meet the platform's expectations.

[0131] Understandably, the restrictions determined by the database establishment conditions can serve not only as a basis for deciding whether to establish a database, but also as a basis for deciding the quantity of databases to be established.

[0132] In the example, the database creation conditions can be set to meet the following conditions simultaneously:

[0133] (1) Seller's membership level: Ordinary paid members and above will have their database created, while free members will not have their database created;

[0134] (2) Product quality level: Products with three stars or above are added to the inventory, while products with one star and two stars are not added to the inventory.

[0135] (3) From the perspective of merchant store level: The maximum inventory building ratio of products is different for stores with different star ratings. The higher the store star rating, the higher the maximum inventory building ratio. See Table 2 for details.

[0136] Table 2 Relationship between Store Star Rating and Inventory Construction Ratio

[0137] Store Star Rating Maximum database construction ratio One star 30％ Two stars 40％ Samsung 50％ Four stars 75％ Five stars 100％

[0138] Deduplication of products that meet the database creation criteria means selecting only one or a portion of duplicate products for database creation. Duplicate products are defined as having identical header images and identical product parameters. The deduplication criteria can be set by the platform itself. In the example, the criteria can be set based on one or more of the seller's membership level, product quality level, and merchant store star rating, or other conditions that align with the platform's support policies.

[0139] In this embodiment, by using the second features related to the seller and the second features related to the product as the criteria for determining whether a product should be included in the recommended product library during the database construction stage, the platform takes into account the seller experience and the types of products that the platform hopes to support when constructing the recommended products. This effectively ensures that the final recommendation results can improve the seller's platform experience and improve the accuracy of the platform's traffic adjustment.

[0140] In some embodiments, products to be recommended are determined from a library of products to be recommended (e.g., in conjunction with...). Figure 2 Step S202) may include: recalling recommended products from the recommended product library according to at least one of a first recall strategy and a second recall strategy, wherein the first recall strategy includes recalling at least one of collinear recall, similar recall and third-level category recall, and the second recall strategy includes recalling products in the recommended product library that meet the individual recall conditions individually.

[0141] Understandably, the categories and quantity of products included in the product recommendation database are relatively large, necessitating further selection during the recommendation process. Recall is this selection process. In this embodiment, the recall strategy includes a first recall strategy and a second recall strategy. The first recall strategy includes at least one of the following: collinear recall, similarity recall, and third-level category recall. This first recall strategy is primarily designed around buyer experience, creating a buyer profile based on primary characteristics relevant to the buyer on the platform, and recalling products that the buyer might be interested in or related to them, thereby increasing buyer trust in the platform and enhancing user stickiness. The second recall strategy is a separate recall strategy, where the platform sets specific recall conditions and recalls products that meet these conditions. These separate recall conditions are set by the platform based on its own interests or development direction, enhancing the platform's freedom in traffic control and ensuring the implementation of its traffic control strategies. Therefore, the recall strategy in this embodiment can be either the first recall strategy which prioritizes buyer experience, or the second recall strategy which recalls products that meet the platform's expectations. Alternatively, considering that the number of products that meet the individual recall conditions set by the platform will not be large, a combination of the first and second recall strategies can be adopted. This way, products that meet buyer expectations can be selected from the product pool to be recommended, while also taking into account seller experience and platform interests.

[0142] In some embodiments, the individual recall criteria in the second recall strategy include the fact that the recommended product belongs to the commercial domain material.

[0143] E-commerce domain materials are paid product materials that sellers pay for on the platform for promotion. Compared to the vast total amount of materials on the platform, e-commerce domain materials account for a relatively small proportion, and the corresponding merchants are also quite scattered. Without other intervention, these products may not get exposure in the product recommendation recall process because they do not meet the recall criteria, thus creating a "long tail effect." However, the sellers of these e-commerce domain materials often have a strong willingness and ability to pay, and activating these products can bring more considerable revenue to the platform. This embodiment adopts a secondary recall strategy of separate recall for e-commerce domain materials, which can effectively meet the traffic needs of the corresponding products and avoid the "long tail effect."

[0144] The order in which products are recommended is a crucial aspect of product recommendations on online platforms. Whether the order is reasonable or not directly affects the buyer's shopping experience on the platform, and consequently affects the buyer's subsequent clicks and conversions. Clicks and conversions are essential steps for a transaction to be completed on an online platform.

[0145] In terms of product recommendation ranking, theoretically, the higher a product ranks, the more exposure it receives. More exposure means a higher probability of clicks and conversions. Therefore, top-ranked positions are relatively scarce, and placing products that generate more revenue for the platform in these positions maximizes platform profits and effectively enhances traffic value. However, at the same time, top-ranked products have a stronger visual impact on buyers, directly affecting their platform experience. For example, if the first recommended product is exactly what the buyer wants, it greatly increases their positive impression of the platform. Conversely, if the top-ranked products are far from the buyer's expectations, the buyer may perceive the platform's recommendations as meaningless and even develop a negative attitude towards the platform's recommendation process. Furthermore, the value each recommended product can create for the platform varies. The product a buyer most wants might generate almost no revenue for the platform, resulting in low click and conversion value. Therefore, when ranking recommended products, it's insufficient to only consider buyer experience; the potential value each product brings to the platform must also be considered.

[0146] Figure 5 A scenario diagram is shown that can implement a product recommendation method according to embodiments of the present disclosure.

[0147] like Figure 5 As shown, different sorting strategies are adopted according to different traffic types. For recommended products under commercial traffic, the first sorting strategy is used, prioritizing products with high equity value and arranging them in order as Q1, Q2, Q3, Q4... For recommended products under non-commercial traffic, the second sorting strategy is used, prioritizing products with larger sorting factors and arranging them in order as F1, F2, F3, F4... Then, considering the user's purchase experience and the platform's value to different positions, some positions in the final recommendation interface presented to buyers are used to recommend products under commercial traffic, such as P2, P3, P5, P6, etc. These positions do not have the same strong impact on the buyer's experience as the first product, and the non-contiguous, intermittent arrangement avoids causing buyer aversion. At the same time, these positions are relatively high, which can give the recommended products a good exposure opportunity, as well as click and conversion opportunities, thereby creating more revenue for the platform.

[0148] Figures 6-7 A structural block diagram of a product recommendation device according to an embodiment of the present disclosure is shown.

[0149] According to embodiments of this disclosure, such as Figure 6 As shown, an internet platform product recommendation device 600 is also provided, which includes:

[0150] The database building module 610 is configured to build a database of products to be recommended based on the first feature, the second feature, and the third feature on the platform; the first feature is a feature associated with buyers on the internet platform, the second feature is a feature associated with sellers on the internet platform, and the third feature is a third feature associated with products on the internet platform.

[0151] The recall module 620 is configured to determine the products to be recommended from the product pool;

[0152] The sorting module 630 is configured to sort the products to be recommended in order to recommend the sorted products to buyers. The sorting module 630 includes a first sorting strategy execution module and a second sorting strategy execution module. The first sorting strategy execution module is configured to sort according to a first sorting strategy based on commercial domain traffic, and the second sorting strategy execution module is configured to sort according to a second sorting strategy based on non-commercial domain traffic.

[0153] In some embodiments, such as Figure 7 As shown, the internet platform product recommendation device 700 includes a database creation module 710, a recall module 720, and a sorting module 730. The sorting module 730 includes a first sorting strategy execution module 731 and a second sorting strategy execution module 732. The first sorting strategy execution module 731 includes a first calculation module 7311 and a first processing module 7312. The first calculation module 7311 is configured to calculate the equity value of the products to be recommended under commercial domain traffic, and the first processing module 7312 is configured to sort according to the equity value of the products to be recommended. The second sorting strategy execution module 732 includes a second calculation module 7321 and a second processing module 7322. The second calculation module 7321 is configured to calculate the sorting factor of the products to be recommended under non-commercial domain traffic, and the second processing module 7322 is configured to sort according to the sorting factor of the products to be recommended.

[0154] In some embodiments, the first calculation module 7311 is configured to calculate the equity value of the product to be recommended based on at least one of the ECPM value, advertising value, and promotion quality.

[0155] In some embodiments, such as Figure 7 As shown, the second computing module 7321 includes:

[0156] The first operation module 7321A is configured to calculate the expected click coefficient and expected conversion coefficient of the product to be recommended based on the historical data of the product to be recommended.

[0157] The second operation module 7321B is configured to calculate the click value and conversion value of the product to be recommended based on a second characteristic associated with the seller of the product to be recommended, a third characteristic associated with the product to be recommended, and platform traffic data. The platform traffic data includes at least one of click traffic data, search traffic conversion data, recommendation traffic conversion data, average overall traffic distribution, and average overall conversion traffic.

[0158] The third operation module 7321C is configured to calculate the ranking factor of the product to be recommended based on the expected click coefficient, click value, expected conversion coefficient, and conversion value.

[0159] In the example, the third operation module 7321C is configured to calculate the ranking factor of the product to be recommended according to the following formula:

[0160] K = ctr × F1 + ct(cvr) × F2,

[0161] Where K is the ranking factor, ctr is the expected click coefficient, F1 is the click value, ct(cvr) is the expected conversion coefficient, and F2 is the conversion value.

[0162] In some embodiments, such as Figure 8 As shown, the second operation module 800 (see...) Figure 7 The second operation module 7321B includes a first calculation module 801 for determining the click value. The first calculation module 801 includes a first submodule 8011, a second submodule 8012, and a third submodule 8013.

[0163] The first submodule 8011 is configured to determine a seller level coefficient based on the privilege level in the second feature associated with the seller of the product to be recommended, the privilege level including at least one of seller membership level, paid advertising level, seller store level, and transaction badge level.

[0164] The second submodule 8012 is configured to determine the platform traffic control coefficient based on the platform traffic data.

[0165] The third submodule 8013 is configured to determine the click value based on the seller equity coefficient and the platform traffic control coefficient, wherein the click value is the sum of the seller equity coefficient and the platform traffic control coefficient.

[0166] In the example, the privilege levels include at least one of the following: seller membership level, paid advertising level, seller store level, and transaction badge level.

[0167] In the example, the third submodule 8013 calculates the click value of the product to be recommended according to the following formula:

[0168] F1 = M + N,

[0169] Where M is the seller level coefficient determined based on the rights level in the second feature, and N is the platform traffic control coefficient.

[0170] In some embodiments, such as Figure 8 As shown, the second submodule 8012 includes:

[0171] The first control processing module 8012A is configured to determine the daily click traffic of the store to which the recommended product belongs and the daily average click traffic of stores with the same star rating under the same product category on the platform, based on the click traffic data.

[0172] The second control processing module 8012B is configured to determine the store traffic control coefficient based on the daily click traffic of the store to which the recommended product belongs and the daily average click traffic of stores with the same star rating under the same product category on the platform.

[0173] The third control and processing module 8012C is configured to determine the traffic quality coefficient of the industry to which the product to be recommended belongs based on search traffic conversion data and recommendation traffic conversion data.

[0174] The fourth regulation and processing module 8012D is configured to determine the platform traffic regulation coefficient based on the store traffic control coefficient and the traffic quality coefficient of the industry to which the recommended product belongs.

[0175] In some embodiments, the platform traffic control coefficient determined by the second submodule 8012 satisfies the following conditions:

[0176] If the daily click traffic of the store containing the product to be recommended exceeds the average daily click traffic of stores with the same star rating in the same product category on the platform, the platform's traffic control coefficient will be negative.

[0177] If the click-through rate of the store containing the product to be recommended is lower than the average click-through rate of stores with the same star rating in the same product category on the platform on that day, the platform's traffic control coefficient will be set to a positive value.

[0178] If the click traffic of the store to which the recommended product belongs on that day is equal to the average click traffic of stores with the same star rating in the same product category on the platform on that day, the platform traffic control coefficient is 0.

[0179] In some embodiments, the platform traffic control coefficient determined by the second submodule 8012 satisfies the following conditions:

[0180] The platform's traffic control coefficient is positively correlated with the traffic quality coefficient of the industry to which the recommended product belongs;

[0181] Among them, the traffic quality coefficient of the industry to which the recommended product belongs is the ratio of the conversion coefficient of the search traffic of the industry to which the recommended product belongs to to the conversion coefficient of the recommended traffic of the industry to which the recommended product belongs.

[0182] In the example, the platform traffic control coefficient N is calculated according to the following formula:

[0183] N = (1-X) i / X av )×C

[0184] Among them, X i X represents the daily click traffic of the store to which the product to be recommended belongs. av C represents the average click-through rate of stores with the same star rating in the same product category as the product to be recommended on the platform on that day, and C represents the traffic quality coefficient of the industry to which the product to be recommended belongs.

[0185] In some embodiments, the second operation module 800 further includes a second calculation module 802 for determining the conversion value of the product to be recommended. The second calculation module 802 includes a fourth submodule 8021, a fifth submodule 8022, and a sixth submodule 8023.

[0186] The fourth submodule 8021 is configured to determine the conversion value coefficient based on the average distribution traffic and the average conversion traffic of the overall market.

[0187] The fifth submodule 8022 is configured to determine the product type coefficient based on whether the product type of the product to be recommended is a preset product type.

[0188] The sixth submodule 8023 is configured to determine the conversion value of the product to be recommended based on the conversion value coefficient and the product type coefficient.

[0189] In the example, the second calculation module 802 is configured to calculate the conversion value of the product to be recommended according to the following formula:

[0190] F2 = H + E,

[0191] Where F2 is the conversion value, H is the conversion value coefficient, and E is the product type coefficient. The conversion value coefficient is determined based on the average distribution traffic and average conversion traffic of the overall market, and the product type coefficient is determined based on the type of product to be recommended.

[0192] In some embodiments, such as Figure 7As shown, the database construction module 710 includes: a determination module 711, configured to determine the product category to be recommended based on a first feature associated with the buyer; a judgment module 712, for products under the product category to be recommended, judging whether the second feature associated with the seller and the third feature associated with the product meet the corresponding database construction conditions, wherein the database construction conditions for the second feature associated with the seller include the seller having a predetermined membership level and store level, and the database construction conditions for the third feature associated with the product include the product having a predetermined quality level; and a third processing module 713, configured to deduplicate the products that meet the database construction conditions and include the deduplicated products in the product database to be recommended.

[0193] In some embodiments, the recall module 720 is configured to recall recommended products from the recommended product library according to at least one of a first recall strategy and a second recall strategy, wherein the first recall strategy includes recall based on at least one of collinear recall, similar recall and third-level category recall, and the second recall strategy includes recalling products in the recommended product library that meet the individual recall conditions individually.

[0194] In some embodiments, the individual recall conditions of the second recall strategy in the recall module 720 include that the recommended product belongs to the commercial domain material.

[0195] The internet platform product recommendation device 700 in the technical solution disclosed herein can be used to implement, for example... Figure 2-4 The product recommendation method shown here balances the interests of buyers, sellers, and the platform, facilitating the precise implementation of the platform's traffic control strategy, resulting in a more rational allocation of platform traffic and increasing the value of platform traffic. Further details will not be elaborated upon here.

[0196] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0197] According to embodiments of this disclosure, an electronic device, a readable storage medium, and a computer program product are also provided.

[0198] refer to Figure 9The present invention describes a structural block diagram of an electronic device 900 that can serve as a server or client of the present disclosure, which is an example of a hardware device that can be applied to various aspects of the present disclosure. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0199] like Figure 9 As shown, the electronic device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded into a random access memory (RAM) 903 from a storage unit 908. The RAM 903 may also store various programs and data required for the operation of the electronic device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0200] Multiple components in electronic device 900 are connected to I / O interface 905, including: input unit 906, output unit 907, storage unit 908, and communication unit 909. Input unit 906 can be any type of device capable of inputting information to electronic device 900. Input unit 906 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of electronic device, and can include, but is not limited to, a mouse, keyboard, touchscreen, trackpad, trackball, joystick, microphone, and / or remote control. Output unit 907 can be any type of device capable of presenting information, and can include, but is not limited to, a monitor, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 908 can include, but is not limited to, disk and optical disk. Communication unit 909 allows electronic device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and can include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers and / or chipsets, such as Bluetooth™ devices, 802.11 devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0201] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as a product recommendation method. For example, in some embodiments, the product recommendation method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the product recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the product recommendation method by any other suitable means (e.g., by means of firmware).

[0202] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0203] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0204] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0205] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0206] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), the Internet, and blockchain networks.

[0207] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0208] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0209] While embodiments or examples of this disclosure have been described with reference to the accompanying drawings, it should be understood that the methods, systems, and devices described above are merely exemplary embodiments or examples, and the scope of the invention is not limited by these embodiments or examples, but only by the granted claims and their equivalents. Various elements in the embodiments or examples may be omitted or replaced by their equivalents. Furthermore, the steps may be performed in a different order than that described in this disclosure. Further, various elements in the embodiments or examples may be combined in various ways. Importantly, as the technology evolves, many elements described herein can be replaced by equivalents that appear after this disclosure.

Claims

1. An Internet platform commodity recommendation method, comprising: constructing a recommended commodity library according to first features associated with buyers, second features associated with sellers, and third features associated with commodities on an Internet platform, including: determining a commodity category to be recommended according to the first features associated with the buyers, judging whether the second features associated with the sellers of the commodities and the third features associated with the commodities under the commodity category to be recommended meet corresponding library construction conditions, wherein the library construction conditions of the second features associated with the sellers include that the sellers have predetermined membership levels and store levels, and the library construction conditions of the third features associated with the commodities include that the commodities have predetermined quality levels; de-duplicating the commodities that meet the library construction conditions and including the de-duplicated commodities in the recommended commodity library; determining recommended commodities from the recommended commodity library, including: recalling the recommended commodities from the recommended commodity library according to a first recall strategy based on buyer experience and a second recall strategy based on platform traffic strategy, wherein the first recall strategy includes recalling based on at least one of co-linear recall, similar recall, and three-level category recall, and the second recall strategy includes separately recalling commodities in the recommended commodity library that meet separate recall conditions; ranking the recommended commodities to recommend the ranked recommended commodities to the buyers, wherein the ranking is performed according to a first ranking strategy based on commercial domain traffic and a second ranking strategy based on non-commercial domain traffic, the first ranking strategy includes ranking according to the benefit value of the recommended commodities, and the second ranking strategy includes determining a ranking factor of the recommended commodities and ranking according to the ranking factor, wherein determining the ranking factor of the recommended commodities in the second ranking strategy includes: determining a click expectation coefficient and a conversion expectation coefficient of the recommended commodities according to historical data of the recommended commodities and the first features associated with the buyers; determining a click value and a conversion value of the recommended commodities according to the second features associated with the sellers of the recommended commodities, the third features associated with the recommended commodities, and platform traffic data, wherein the platform traffic data includes click traffic data, search traffic conversion data, recommendation traffic conversion data, average distribution traffic of the market, and average conversion traffic of the market; and determining the ranking factor of the recommended commodities according to the click expectation coefficient, the click value, the conversion expectation coefficient, and the conversion value.

2. The method of claim 1, wherein, The benefit value of the recommended commodities in the first ranking strategy includes at least one of ECPM, advertising value, and promotion quality of the recommended commodities.

3. The method of claim 1, wherein, The determination of the click value and the conversion value of the recommended commodities according to the second features associated with the sellers of the recommended commodities, the third features associated with the recommended commodities, and the platform traffic data includes: a seller level coefficient determined according to a benefit level in the second feature associated with a seller of the to-be-recommended commodity, the benefit level comprising at least one of a seller member level, a paid advertisement level, a store level of the seller, and a transaction medal level; a platform traffic regulation coefficient determined according to the platform traffic data; the click value is determined according to the seller level coefficient and the platform traffic regulation coefficient; wherein the click value is a sum of the seller benefit coefficient and the platform traffic regulation coefficient.

4. The method of claim 3, wherein, the determination of the platform traffic regulation coefficient according to the platform traffic data comprises: determination of a click traffic of a store to which the to-be-recommended commodity belongs on the day and an average click traffic of stores of the same star level under the same commodity category on the platform on the day according to the click traffic data; determination of a store traffic control coefficient according to the click traffic of the store to which the to-be-recommended commodity belongs on the day and the average click traffic of the stores of the same star level under the same commodity category on the platform on the day; determination of a traffic quality coefficient of an industry to which the to-be-recommended commodity belongs according to the search traffic conversion data and the recommendation traffic conversion data; determination of the platform traffic regulation coefficient according to the store traffic control coefficient and the traffic quality coefficient of the industry to which the to-be-recommended commodity belongs.

5. The method of claim 4, wherein, the platform traffic regulation coefficient satisfies the following conditions: in response to the click traffic of the store to which the to-be-recommended commodity belongs on the day being higher than the average click traffic of the stores of the same star level under the same commodity category on the platform on the day, the platform traffic regulation coefficient is a negative value, in response to the click traffic of the store to which the to-be-recommended commodity belongs on the day being lower than the average click traffic of the stores of the same star level under the same commodity category on the platform on the day, the platform traffic regulation coefficient is a positive value, in response to the click traffic of the store to which the to-be-recommended commodity belongs on the day being equal to the average click traffic of the stores of the same star level under the same commodity category on the platform on the day, the platform traffic regulation coefficient is 0.

6. The method of claim 4 or 5, wherein, the platform traffic regulation coefficient further satisfies the following conditions: the platform traffic regulation coefficient is positively correlated with the traffic quality coefficient of the industry to which the to-be-recommended commodity belongs; wherein the traffic quality coefficient of the industry to which the to-be-recommended commodity belongs is a ratio of a conversion coefficient of search traffic of the industry to which the to-be-recommended commodity belongs to a conversion coefficient of recommendation traffic of the industry to which the to-be-recommended commodity belongs.

7. The method of claim 1, wherein, the determination of the click value and the conversion value of the to-be-recommended commodity according to the second feature associated with the seller of the to-be-recommended commodity, the third feature related to the to-be-recommended commodity, and the platform traffic data further comprises: determination of a conversion value coefficient according to the big board average distribution traffic and the big board average conversion traffic, determination of a commodity type coefficient according to whether the commodity type of the to-be-recommended commodity is a preset commodity type, determination of the conversion value according to the conversion value coefficient and the commodity type coefficient.

8. The method of claim 1, wherein, the separate recall condition in the second recall strategy comprises that the to-be-recommended commodity belongs to a commercial domain material.

9. An internet platform commodity recommendation device, comprising: The building module is configured to build a recommended commodity library according to first features associated with buyers, second features associated with sellers and third features associated with commodities on the Internet platform, and the building module comprises: The determining module is configured to determine the commodity category to be recommended according to the first features associated with the buyers; The judging module is configured to judge whether the second features associated with sellers of commodities under the commodity category to be recommended and the third features associated with the commodities satisfy corresponding building conditions, wherein the building condition of the second features associated with the sellers comprises that the sellers have predetermined membership levels and store levels, and the building condition of the third features associated with the commodities comprises that the commodities have predetermined quality levels; The third processing module is configured to deduplicate the commodities satisfying the building conditions and incorporate the deduplicated commodities into the recommended commodity library; The recalling module is configured to determine recommended commodities from the recommended commodity library, and the recalling module comprises: The recalling module is configured to determine recommended commodities from the recommended commodity library, and the recalling module comprises: The sorting module is configured to sort the recommended commodities to recommend the sorted recommended commodities to the buyers, wherein the sorting module comprises a first sorting strategy execution module and a second sorting strategy execution module, the first sorting strategy execution module is configured to sort according to a first sorting strategy based on commercial domain traffic, and the second sorting strategy execution module is configured to sort according to a second sorting strategy based on non-commercial domain traffic, the first sorting strategy execution module comprises a first calculation module and a first processing module, the first calculation module is configured to calculate the benefit value of the recommended commodities under the commercial domain traffic, and the first processing module is configured to sort according to the benefit value of the recommended commodities; the second sorting strategy execution module comprises a second calculation module and a second processing module, the second calculation module is configured to calculate a sorting factor of the recommended commodities under the non-commercial domain traffic, and the second processing module is configured to sort according to the sorting factor of the recommended commodities, wherein the second calculation module comprises: The first operation module is configured to determine a click expectation coefficient and a conversion expectation coefficient of the recommended commodities according to historical data of the recommended commodities and the first features associated with the buyers; a second operation module configured to determine a click value and a conversion value of the to-be-recommended commodity according to the second feature associated with a seller to which the to-be-recommended commodity belongs and platform traffic data, wherein the platform traffic data comprises at least one of click traffic data, search traffic conversion data, recommendation traffic conversion data, average distribution traffic of a market, and average conversion traffic of the market; and a third operation module configured to determine the ranking factor of the to-be-recommended commodity according to the click expectation coefficient, the click value, the conversion expectation coefficient, and the conversion value.

10. The apparatus of claim 9, wherein, The first calculation module is configured to determine the benefit value of the to-be-recommended commodity according to at least one of an ECPM, an advertisement value, and a promotion quality of the to-be-recommended commodity.

11. The apparatus of claim 9, wherein, The second operation module comprises a first operation module configured to determine the click value, and the first operation module comprises: a first submodule configured to determine a seller level coefficient according to a benefit level in the second feature associated with the seller of the to-be-recommended commodity, wherein the benefit level comprises at least one of a seller membership level, a paid advertisement level, a store level of the seller, and a transaction medal level; a second submodule configured to determine a platform traffic regulation coefficient according to the platform traffic data; a third submodule configured to determine the click value according to the seller benefit coefficient and the platform traffic regulation coefficient, wherein the click value is a sum of the seller benefit coefficient and the platform traffic regulation coefficient.

12. The apparatus of claim 11, wherein, The second submodule comprises: a first regulation processing module configured to determine a click traffic of a store to which the to-be-recommended commodity belongs on a current day and an average click traffic of stores of the same star level under the same commodity category on the current day according to the click traffic data; a second regulation processing module configured to determine a store traffic control coefficient according to the click traffic of the store to which the to-be-recommended commodity belongs on the current day and the average click traffic of the stores of the same star level under the same commodity category on the current day; a third regulation processing module configured to determine a traffic quality coefficient of an industry to which the to-be-recommended commodity belongs according to the search traffic conversion data and the recommendation traffic conversion data; a fourth regulation processing module configured to determine the platform traffic regulation coefficient according to the store traffic control coefficient and the traffic quality coefficient of the industry to which the to-be-recommended commodity belongs.

13. The apparatus of claim 12, wherein, The platform traffic regulation coefficient determined by the second submodule satisfies the following conditions: in response to the click traffic of the store to which the to-be-recommended commodity belongs on the current day being greater than the average click traffic of the stores of the same star level under the same commodity category on the current day, the platform traffic regulation coefficient is a negative value, in response to the click traffic of the store to which the to-be-recommended commodity belongs on the current day being less than the average click traffic of the stores of the same star level under the same commodity category on the current day, the platform traffic regulation coefficient is a positive value, in response to the click traffic of the store to which the to-be-recommended commodity belongs on the current day being equal to the average click traffic of the stores of the same star level under the same commodity category on the current day, the platform traffic regulation coefficient is 0.

14. The apparatus of claim 12 or 13, wherein, The platform traffic regulation coefficient determined by the second submodule further satisfies the following conditions: The platform traffic regulation coefficient is positively correlated with a traffic quality coefficient of an industry to which the to-be-recommended commodity belongs. The traffic quality coefficient of the industry to which the to-be-recommended commodity belongs is a ratio of a conversion coefficient of search traffic of the industry to which the to-be-recommended commodity belongs to a conversion coefficient of recommended traffic of the industry to which the to-be-recommended commodity belongs.

15. The apparatus of claim 9, wherein, The second operation module includes a second operation module configured to determine the conversion value, and the second operation module includes: a fourth submodule configured to determine a conversion value coefficient according to the average distribution traffic of the market and the average conversion traffic of the market, a fifth submodule configured to determine a commodity type coefficient according to whether the commodity type of the to-be-recommended commodity is a preset commodity type, a sixth submodule configured to determine the conversion value according to the conversion value coefficient and the commodity type coefficient.

16. The apparatus of claim 9, wherein, The individual recall condition in the second recall strategy includes that the to-be-recommended commodity belongs to a commercial domain material.

17. An electronic device, comprising: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.

18. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the method of any one of claims 1-8.

19. A computer program product comprising a computer program, wherein, The computer program, when executed by the processor, implements the method of any one of claims 1-8. The computer program, when executed by the processor, implements the method of any one of claims 1-8.

Images