CN116757770A

CN116757770A - Method and device for recalling goods-carrying resources

Info

Publication number: CN116757770A
Application number: CN202310653469.4A
Authority: CN
Inventors: 肖涛; 李善涛
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2023-06-02
Filing date: 2023-06-02
Publication date: 2023-09-15

Abstract

The disclosure provides a recall method and device for a cargo resource, relates to the field of artificial intelligence, and particularly relates to the field of intelligent search. The specific implementation scheme is as follows: responding to a received search request of a user, and acquiring the goods resources clicked by at least one similar user of the user; calculating posterior conversion rate of the goods resources clicked by each similar user; acquiring the average unit price of the goods mounted by the goods resources clicked by each similar user; determining the similarity between the user and each similar user as the point display ratio of the goods resources; for each goods-carrying resource clicked by the similar user, calculating thousands of display benefits of the goods-carrying resource according to the product of the posterior conversion rate of the goods-carrying resource, the average unit price of the mounted goods and the point-to-display ratio; a predetermined number of the tape resources are recalled in the order of thousands of presentation benefits from high to low. According to the implementation mode, thousands of display benefits are fully mined during recall, and the recommendation precision of the goods-carrying resources is improved.

Description

Method and device for recalling goods-carrying resources

Technical Field

The disclosure relates to the field of artificial intelligence, in particular to the field of intelligent search, and specifically relates to a recall method and device for a cargo resource.

Background

In the information flow recommendation system, the on-demand resource is a relatively special resource, which is itself a carrier of content, and in addition, it inserts commodities related to the content in the middle or at the end of the content for the purchase of users with needs. For example, an article resource for explaining the training skills of the shuttlecock is obtained, a shuttlecock purchasing link of a certain brand is hung at the tail end of the article, and after the user finishes looking at the article, the user can directly purchase the shuttlecock from the link at the tail end of the article, so that the deeper experience of the user is satisfied, the creator of the article can obtain the score, the platform can also obtain the score, and the method is a three-win mode. Accordingly, the source of the goods is deeply paid attention to each large content distribution platform.

In the information flow recommendation system, the click-to-display ratio (click/display) is an important measurement index of various resources, and the commodity resources are required to consider the conversion rate (i.e. the probability of purchasing the commodity on which the resources are mounted) and the price of the commodity besides the click-to-display ratio.

In a stream recommendation system, resources are typically subjected to two phases, recall and sort. Recall is a thousands of resources selected from tens of thousands of resources, and ranking is a tens of thousands of resources selected from recalls. In the sorting stage, the click-to-display ratio and conversion rate are estimated at the same time, and the price of the commodity is considered, so that resources with high thousands of display benefits (ecpm, effective cost per mille) are selected. However, in the recall stage, most recalls only consider the point-to-display ratio, and the selected resources are always high in point-to-display ratio (but ecpm is not necessarily high), so that recall and sorting funnels are inconsistent, and the recommendation accuracy of the loaded resources is reduced.

Disclosure of Invention

The present disclosure provides a method, apparatus, device, storage medium, and computer program product for recall of a commodity resource.

According to a first aspect of the present disclosure, there is provided a recall method of a stocked resource, comprising: responding to a received search request of a user, and acquiring a commodity resource matched with the search request and clicked by at least one similar user of the user; calculating posterior conversion rates of the at least one similarly user-clicked on the commodity resource; acquiring the average unit price of the goods mounted by the goods resources clicked by the at least one similar user; for each similar user in the at least one similar user, determining the similarity between the user and the similar user as the point-to-display ratio of the goods resources clicked by the similar user; for each goods-carrying resource clicked by the similar user, calculating thousands of display benefits of the goods-carrying resource according to the product of the posterior conversion rate of the goods-carrying resource, the average unit price of the mounted goods and the point-to-display ratio; a predetermined number of the tape resources are recalled in the order of thousands of presentation benefits from high to low.

According to a second aspect of the present disclosure, there is provided a recall device for a stocked resource, comprising: a request unit configured to obtain, in response to receiving a search request from a user, a pickup resource that matches the search request and that has been clicked by at least one similar user of the user; a computing unit configured to compute a posterior conversion of the at least one similarly user clicked on the commodity resource; an obtaining unit configured to obtain an average unit price of the commodity mounted by the loaded resource clicked by the at least one similar user; a determining unit configured to determine, for each of the at least one similar user, a similarity between the user and the similar user as a point spread ratio of the commodity resource clicked by the similar user; a revenue unit configured to calculate, for each of the similar user clicked on the commodity-carrying resource, thousands of display revenue for the commodity-carrying resource based on a product of a posterior conversion rate of the commodity-carrying resource, an average unit price of the commodity-carrying and a point-to-display ratio; a recall unit configured to recall a predetermined number of the tape resources in order of thousands of presentation benefits from high to low.

According to a third aspect of the present disclosure, there is provided an electronic device for recall of a commodity resource, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the first aspects.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method of any one of the first aspects.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the method of any of the first aspects.

According to the recall method and device for the goods resources, the recall of ucf (Item Collaborative Filtering, collaborative filtering of similar users) of the goods resources is achieved, and through the three factors of average unit price, conversion rate and point-to-display ratio, ecpm signals can be fully mined, so that recall-sorting funnels are consistent, and the problem that recommendation accuracy is reduced due to inconsistent recall-sorting funnels is solved.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is an exemplary system architecture diagram in which an embodiment of the present disclosure may be applied;

FIG. 2 is a flow chart of one embodiment of a method of recall of a commodity resource according to the present disclosure;

FIG. 3 is a flow chart of yet another embodiment of a method of recall of a commodity resource according to the present disclosure;

FIG. 4 is a schematic illustration of one application scenario of the with-resource recall method according to the present disclosure;

FIG. 5 is a schematic structural view of one embodiment of a recall device for a commodity resource according to the present disclosure;

fig. 6 is a schematic diagram of a computer system suitable for use in implementing embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

FIG. 1 illustrates an exemplary system architecture 100 to which embodiments of the present disclosure of a method or apparatus for recall of a commodity resource may be applied.

As shown in fig. 1, a system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used as a medium to provide communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user may interact with the server 105 via the network 104 using the terminal devices 101, 102, 103 to receive or send messages or the like. Various communication client applications, such as a web browser application, a shopping class application, a search class application, an instant messaging tool, a mailbox client, social platform software, etc., may be installed on the terminal devices 101, 102, 103.

The terminal devices 101, 102, 103 may be hardware or software. When the terminal devices 101, 102, 103 are hardware, they may be various electronic devices having a display screen and supporting video playback, including but not limited to smartphones, tablet computers, electronic book readers, MP3 players (Moving Picture Experts Group Audio Layer III, dynamic video expert compression standard audio plane 3), MP4 (Moving Picture Experts Group Audio Layer IV, dynamic video expert compression standard audio plane 4) players, laptop and desktop computers, and the like. When the terminal devices 101, 102, 103 are software, they can be installed in the above-listed electronic devices. Which may be implemented as multiple software or software modules (e.g., to provide distributed services), or as a single software or software module. The present invention is not particularly limited herein.

The server 105 may be a server providing various services, such as a background video server providing support for videos displayed on the terminal devices 101, 102, 103. The background video server may analyze and process the received data such as the video search request, and feed back the processing result (e.g., the video with the goods resources) to the terminal device.

The server may be hardware or software. When the server is hardware, the server may be implemented as a distributed server cluster formed by a plurality of servers, or may be implemented as a single server. When the server is software, it may be implemented as a plurality of software or software modules (e.g., a plurality of software or software modules for providing distributed services), or as a single software or software module. The present invention is not particularly limited herein. The server may also be a server of a distributed system or a server that incorporates a blockchain. The server can also be a cloud server, or an intelligent cloud computing server or an intelligent cloud host with artificial intelligence technology.

It should be noted that, the method for recalling the resources with goods provided by the embodiments of the present disclosure is generally performed by the server 105, and accordingly, the recall device for the resources with goods is generally disposed in the server 105.

It should be understood that the number of terminal devices, networks and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

With continued reference to FIG. 2, a flow 200 of one embodiment of a method of recall of a commodity resource according to the present disclosure is shown. The recall method for the goods resources comprises the following steps:

in response to receiving the search request from the user, the step 201 obtains the tape resources that match the search request and have been clicked by at least one similar user.

In this embodiment, the execution body of the recall method for the resources on the tape (e.g., the server shown in fig. 1) may receive a search request from a terminal with which a user performs web browsing or video browsing through a wired connection or a wireless connection, where the search request may include a search word, a search image, and the like. The search request may further include a user identifier, which may be a user login ID or user cookie information, so that the behavior of the user may be queried from the log of the server according to the user identifier.

Similar users can be found through the actions of the users (such as clicking, collecting, focusing, praying, etc.), for example, the users who click on the same goods resources, collect the same goods resources, focus on the same goods resources, pray the same goods resources are similar users. The behavior of the user can be matched with the behavior of other users, and the more the same behavior is matched, the higher the similarity of the behaviors is. Users whose similarity of behavior is above a predetermined similarity threshold are determined to be similar users.

And acquiring the carried resource which is clicked by at least one similar user and matched with the search request. The tape resources may include information such as a topic name, a summary, etc. The shipped resources clicked by the similar user can be matched with the search words or search images in the search request, and the shipped resources with the matching degree lower than the preset value are filtered out.

Step 202, calculating posterior conversion rates of at least one similarly user clicked on the commodity resource.

In this embodiment, the log of the server stores a click record and a purchase record of all the tape resources, where the click record may include a user identifier and a click amount for clicking the tape resources, and the purchase record may include a user identifier and a purchase amount for purchasing the commodity on which the tape resources are mounted. The posterior conversion may be derived from the purchase/click-through of the shipped resource that was clicked on by the similar user.

Step 203, obtaining the average unit price of the goods loaded by the loaded resource clicked by at least one similar user.

In this embodiment, each of the goods resources is loaded with at least one commodity, and the average value of the unit prices of the commodities can be obtained. Alternatively, a weight higher than that of the non-singled items may be set for the singled items (click purchase), and a weighted average unit price is calculated.

Step 204, for each of at least one similar user, determining a similarity between the user and the similar user as a point-to-display ratio of the tape-out resource clicked by the similar user.

In this embodiment, the expected user's click-to-display ratio is a percentage of ideal. While it is impossible for similar users to duplicate the user's operations entirely, it is possible to predict that the point spread ratio is 100% for similar users based on the similarity. For example, if the similarity between the user 1 and the user 2 is 0.8, the point spread ratio of the resources with the goods of the user 2 is 0.8.

Step 205, for each of the similar user clicked on the commodity resources, calculating thousands of display benefits of the commodity resources according to the product of the posterior conversion rate of the commodity resources, the average unit price of the mounted commodity and the point-to-display ratio.

In this embodiment, ecpm, effective cost per mille, is first referred to as the revenue available per thousand displays. For a commodity resource, that is, a thousand exhibits the total revenue available for selling the commodity.

Assuming n presentations of the resource, and eventually m clicks, a single amount of k, an average price per single of p,

the total revenue obtained gmv=p×k

Ecpm=gmv/n 1000=p k/n 1000=p (k/m) 1000

And (k/m) represents the conversion cvr; (m/n) represents a dot expansion ratio ctr;

thus ecpm=p cvr ctr 1000, i.e. ecpm is equal to the product of average unit price, conversion, point spread ratio, constant 1000. We hope that ecpm is higher, then there is a need to raise the three factors of average unit price, conversion, point spread ratio simultaneously.

Step 206, recalling a predetermined number of the tape resources in the order of thousands of presentation benefits from high to low.

In this embodiment, the predetermined number may be 100, and the 100 with the highest display benefit is selected from the plurality of the tape resources clicked by the similar users, and if there are duplicate tape resources, the duplication is removed. Some similar users have more selected pickup resources, some similar users have less selected pickup resources, and the number of pickup resources selected by each similar user may not be limited. Alternatively, the same number of shipped resources may be selected from among the shipped resources clicked on by each similar user.

The recalled goods resources can be directly pushed to the user according to the sequence of thousands of display benefits from high to low, or can be pushed to the user after being sequenced by a sequencing algorithm.

According to the method provided by the embodiment of the disclosure, in the recall stage, the ecpm signal is fully considered, so that recall resources are guaranteed to be consistent with ordering ratio, conversion rate and order unit price signals, and the recommending precision of the goods resources is improved.

In some optional implementations of this embodiment, the obtaining the on-demand resource that matches the search request that is clicked by the at least one similar user of the user includes: acquiring user portraits of all users; calculating the similarity between the user portrait of the user and the user portraits of other users; and querying the tape resources clicked by at least one similar user with the similarity higher than a preset similarity threshold value from the historical click records. User portraits are used to describe the identity and behavioral characteristics of a user, such as gender, age, occupation, click, attention, collection, browsing, etc. The user representation may be represented by vectors and then the similarity between the vectors (including but not limited to cosine similarity, hamming distance, etc.) is calculated. The similarity is higher than a threshold, for example, more than 0.6, and the user clicks on the commodity resource to make reference sense. The effective candidate recommended resources can be quickly and accurately queried.

In some optional implementations of the present embodiment, the calculating the posterior conversion rate of the at least one similar user clicked on the commodity resource includes: counting the distribution amount and conversion amount of each goods-carrying resource clicked by the at least one similar user; for each of the commodity resources, the ratio of the conversion of the commodity resource to the distribution of the commodity resource is taken as the posterior conversion of the commodity resource. The distribution amount may be a weighted sum of the number of forwarding, downloading, clicking, etc. operations. The conversion may be the number of orders converted from the mounted merchandise. Posterior conversion was = conversion/distribution.

In some optional implementations of the present embodiment, the calculating the posterior conversion rate of the at least one similar user clicked on the commodity resource includes: counting the distribution amount and conversion amount of each goods-carrying resource clicked by the at least one similar user; and for each goods-carrying resource, if the distribution amount of the goods-carrying resource is smaller than the preset confidence distribution amount, acquiring the distribution amount and conversion amount of the goods carried by the goods-carrying resource, and taking the ratio of the conversion amount of the goods carried to the distribution amount of the goods carried as the posterior conversion rate of the goods-carrying resource. This prevents the calculated posterior conversion from being excessively large when the distribution amount is excessively small, contrary to the fact.

In some optional implementations of the present embodiment, the calculating the posterior conversion rate of the at least one similar user clicked on the commodity resource includes: for each of the shipped resources, if the dispense amount of the shipped resource is greater than the predetermined dispense amount threshold and the posterior conversion is less than the predetermined conversion threshold, the posterior conversion of the shipped resource is set to 0. High distribution and ultra-low conversion resources are punished. For example, a 10w dispense, but only 10 conversions (at 1% average conversion, which should be 1000 conversions) such resources are generally considered unsuitable for shipment, and therefore the posterior conversion is set directly to 0.

In some optional implementations of this embodiment, the obtaining the average unit price of the commodity with the cargo resource mount clicked by the at least one similar user includes: and for each goods with the goods clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods, and calculating the weighted average unit price of all goods mounted by the goods with the goods, wherein the weight of the goods which are already singly made is higher than that of the goods which are not singly made. This increases the impact of the already-singulated products on the tape resources, and thus tends to recommend tape resources that can be singulated, increasing conversion and thus ecpm.

In some optional implementations of this embodiment, the obtaining the average unit price of the commodity with the cargo resource mount clicked by the at least one similar user includes: and for each goods with the goods resources clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods resources, modifying the unit price of the goods with the unit price exceeding the upper limit price to the upper limit price, and then calculating the weighted average unit price of all the goods mounted by the goods with the goods resources, wherein the weight of the goods with the goods being made into a single item is higher than that of the goods without the goods with the goods being made into a single item. Considering that the unit price of a part of goods can be very high (such as an automobile, etc.), when the unit price is too high, an upper limit price (such as 2000) can be taken to prevent the calculated unit price of the resource from being far higher than other resources, so that the recommended diversity is poor. The sorting stage also limits the upper limit on the average unit price, which sorting funnel remains consistent.

In some optional implementations of the present embodiment, recalling a predetermined number of the tape resources in a thousand presentation benefits from high to low order includes: for each of the similar user clicked on the shipping resources, a predetermined number of the shipping resources are selected in order of thousands of display benefits from high to low. In this way, each similar user can have the same number of the resources with goods selected, so as to avoid that users with highest similarity are pushed mainly, and the diversity of the recommendation is poor.

With further reference to FIG. 3, a flow 300 of yet another embodiment of a method of recall of a commodity resource is shown. The process 300 of the recall method for the resources in the shipment comprises the following steps:

step 301, constructing a graph model of the user-resource based on the history click record, and obtaining an embedded vector of each user by training the graph model.

In this embodiment, the graph model exists widely in daily life, such as social networks, communication networks, protein structures, and knowledge maps. The graph is made up of a series of nodes and edges. Wherein a node is an entity we define, which constitutes a set of nodes; an edge is a line connected between two nodes and is divided into a directed edge and a non-directed edge. We can additionally define some labels and attribute features in the figure. The label can be an edge label or a node label. For the attribute of the node, if the node is defined as a user, the attribute may be defined as the sex of the user, the age of the user, the occupation of the user, etc.; if the node is a user and a resource, the interaction behavior generated by the user and the resource, such as clicking, sharing and the like, or the residence time length and the time stamp of the user, can be used as the attribute of the edge.

According to the historical clicking behaviors of all users, each user is a node, the resource is a node, the user a points the resource b, and an edge is formed between the edge a and the edge b. The nodes and edges form a graph model, and then the graph model is trained by random walk to obtain the embedded vector of each node.

Step 302, calculating cosine similarity between embedded vectors of any two users, and constructing indexes between the tape resources of each user and other users according to the cosine similarity.

In this embodiment, the cosine similarity is calculated by the vector between the users, and the larger this value, the more similar the two are. And finally, constructing an index for online recall. The index content is as follows:

user1[user2&0.9,user3&0.8,user4&0.7,…]

this means that the head-like users of user1 are user2, user3, user4, with similarities of 0.9, 0.8, 0.7, respectively.

Step 303, in response to receiving the search request of the user, querying an index of at least one similar user having cosine similarity to the user higher than a predetermined similarity threshold, and acquiring the tape resources according to the index of the at least one similar user.

In this embodiment, according to the index constructed in the user identification query step 302, users with low similarity are filtered, and the index of the similar user is obtained to obtain the tape resource.

Step 304, a posterior conversion rate of the at least one similarly user clicked on the commodity resource is calculated.

Step 305, obtaining the average unit price of the goods loaded by the loaded resource clicked by at least one similar user.

Step 306, for each of the at least one similar user, determining the similarity between the user and the similar user as the point-to-display ratio of the tape-out resource clicked by the similar user.

Step 307, for each of the similar user clicked on the commodity resources, calculating thousands of display benefits of the commodity resources according to the product of the posterior conversion rate of the commodity resources, the average unit price of the mounted commodity and the point-to-display ratio.

Step 308 recalls a predetermined number of the tape resources in a thousand show returns from high to low order.

Steps 304-308 are substantially identical to steps 202-206 and are therefore not described in detail.

The solution described in this embodiment may construct a graph model of the user-resource, which defines a series of nodes and edges by constructing graph attributes, so that the loss during training may be conducted to the neighbor nodes of the target node. For example, in the market of relatives, a traditional machine learning task judges whether a man is a good-quality relatives or not according to characteristics of the man, such as occupation, salary, academic, etc., and a graph model can introduce related nodes, such as relatives, friends, etc., to jointly determine the characteristics of the man. The above examples demonstrate that it is more reliable to introduce graph models in the information flow recall service.

With continued reference to fig. 4, fig. 4 is a schematic diagram of an application scenario of the recall method of the tape-out resource according to the present embodiment. In the application scenario of fig. 4, the specific process flow is as follows:

1, constructing a user-resource graph model, and obtaining an EMBeding vector of a user through training. The user then calculates cosine similarity between each pair of vectors, the larger this value, the more similar the two are. And finally, constructing an index for online recall. The index content is as follows:

user1[user2&0.9,user3&0.8,user4&0.7,…]

2, calculating a posterior conversion rate for each cargo resource. The calculation logic is as follows:

2.1, counting the distribution amount and conversion amount of each commodity-carrying resource, and the distribution amount and conversion amount of commodities carried by the commodity-carrying resource, wherein different commodity-carrying resources can be used for carrying the same commodity, and the distribution amount and conversion amount of the commodity are the sum of the distribution amount and conversion amount of all commodity-carrying resources carried by the commodity.

2.2, setting the confidence distribution amount (ensuring that the posterior conversion is sufficiently confidence), assuming that the average conversion is around 1%, the confidence distribution amount may be set to 1000, and 1000 distributions, on average, have 10 conversions, calculated as the average conversion.

2.3, for a tape resource with a dispense volume greater than the confidence dispense volume, the posterior conversion rate is =conversion volume/dispense volume.

2.4, for the tape resources with the distribution amount smaller than the confidence distribution amount, we use the posterior conversion rate of the mounted commodity as the posterior conversion rate of the resources.

2.5, in addition, penalizing resources with high distribution and ultra-low conversion rate. For example, a 10w dispense, but only 10 conversions (at 1% average conversion, which should be 1000 conversions) such resources are generally considered unsuitable for shipment, and therefore the posterior conversion is set directly to 0.

3, obtaining the average unit price (generally a fixed value) of the resource-mounted commodity, considering that the unit price of part of the commodity can be very high (such as an automobile, etc.), when the unit price is too high, an upper limit price (for example 2000) can be obtained when the unit price is calculated, so that the calculated unit price of the resource is not far higher than other resources, and the recommended diversity is poor. The sorting stage also limits the upper limit on the average unit price, which sorting funnel remains consistent.

4, in the first step, the cosine similarity between users is mainly estimated by the point spread ratio ctr; the second step calculates the posterior conversion ctr; the third step achieves an average unit price. Using these three values, ecpm can be calculated. During recall, recall ecpm-high resources are optimized, so that recall and ordered funnels are kept consistent, and the efficiency of the whole recommendation system is improved.

With further reference to fig. 5, as an implementation of the method shown in the foregoing figures, the present disclosure provides an embodiment of a recall device for a stocked resource, which corresponds to the method embodiment shown in fig. 2, and which is particularly applicable to various electronic devices.

As shown in fig. 5, the recall device 500 for a cargo resource of the present embodiment includes: a request unit 501, a calculation unit 502, an acquisition unit 503, a determination unit 504, a profit unit 505 and a recall unit 506. Wherein, the request unit 501 is configured to respond to receiving a search request of a user, and acquire a commodity resource which is clicked by at least one similar user of the user and matches with the search request; a computing unit 502 configured to compute a posterior conversion of the at least one similar user clicked on the commodity resource; an obtaining unit 503 configured to obtain an average unit price of the commodity mounted by the loaded resource clicked by the at least one similar user; a determining unit 504 configured to determine, for each of the at least one similar user, a similarity between the user and the similar user as a click-to-display ratio of the tape-out resource clicked by the similar user; a revenue unit 505 configured to calculate, for each of the similar user clicked on the commodity resources, thousands of display revenue for the commodity resources based on the product of the posterior conversion rate of the commodity resources, the average unit price of the mounted commodity, and the point-to-display ratio; recall unit 506 configured to recall a predetermined number of the tape resources in a thousand presentation revenue from high to low order.

In this embodiment, the specific processing of the request unit 501, the calculation unit 502, the acquisition unit 503, the determination unit 504, the profit unit 505, and the recall unit 506 of the tape-out resource recall device 500 may refer to steps 201-206 in the corresponding embodiment of fig. 2.

In some optional implementations of the present embodiment, the requesting unit 501 is further configured to: acquiring user portraits of all users; calculating the similarity between the user portrait of the user and the user portraits of other users; and querying the tape resources clicked by at least one similar user with the similarity higher than a preset similarity threshold value from the historical click records.

In some optional implementations of the present embodiment, the apparatus 500 further includes a construction unit (not shown in the drawings) configured to: before receiving a search request of a user, constructing a graph model of the user-resource based on the historical click record; obtaining an embedded vector of each user by training the graph model; calculating cosine similarity between embedded vectors of any two users, and constructing indexes between the goods resources of each user and other users according to the cosine similarity; and the requesting unit 501 is further configured to: querying indexes of at least one similar user with cosine similarity with the user higher than a preset similarity threshold; and acquiring the cargo resources according to the index of the at least one similar user.

In some optional implementations of the present embodiment, the computing unit 502 is further configured to: counting the distribution amount and conversion amount of each goods-carrying resource clicked by the at least one similar user; for each of the commodity resources, the ratio of the conversion of the commodity resource to the distribution of the commodity resource is taken as the posterior conversion of the commodity resource.

In some optional implementations of the present embodiment, the computing unit 502 is further configured to: counting the distribution amount and conversion amount of each goods-carrying resource clicked by the at least one similar user; and for each goods-carrying resource, if the distribution amount of the goods-carrying resource is smaller than the preset confidence distribution amount, acquiring the distribution amount and conversion amount of the goods carried by the goods-carrying resource, and taking the ratio of the conversion amount of the goods carried to the distribution amount of the goods carried as the posterior conversion rate of the goods-carrying resource.

In some optional implementations of the present embodiment, the computing unit 502 is further configured to: for each of the shipped resources, if the dispense amount of the shipped resource is greater than the predetermined dispense amount threshold and the posterior conversion is less than the predetermined conversion threshold, the posterior conversion of the shipped resource is set to 0.

In some optional implementations of the present embodiment, the obtaining unit 503 is further configured to: and for each goods with the goods clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods, and calculating the weighted average unit price of all goods mounted by the goods with the goods, wherein the weight of the goods which are already singly made is higher than that of the goods which are not singly made.

In some optional implementations of the present embodiment, the obtaining unit 503 is further configured to: and for each goods with the goods resources clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods resources, modifying the unit price of the goods with the unit price exceeding the upper limit price to the upper limit price, and then calculating the weighted average unit price of all the goods mounted by the goods with the goods resources, wherein the weight of the goods with the goods being made into a single item is higher than that of the goods without the goods with the goods being made into a single item.

In some optional implementations of the present embodiment, the recall unit 506 is further configured to: for each of the similar user clicked on the shipping resources, a predetermined number of the shipping resources are selected in order of thousands of display benefits from high to low.

In the technical scheme of the disclosure, the related processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user accord with the regulations of related laws and regulations, and the public order colloquial is not violated.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of flow 200 or 300.

A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method of flow 200 or 300.

A computer program product comprising a computer program that when executed by a processor implements the method of flow 200 or 300.

Fig. 6 illustrates a schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 6, the apparatus 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 602 or a computer program loaded from a storage unit 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 may also be stored. The computing unit 601, ROM 602, and RAM 603 are connected to each other by a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

Various components in the device 600 are connected to the I/O interface 605, including: an input unit 606 such as a keyboard, mouse, etc.; an output unit 607 such as various types of displays, speakers, and the like; a storage unit 608, such as a magnetic disk, optical disk, or the like; and a communication unit 609 such as a network card, modem, wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

The computing unit 601 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 601 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized 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 601 performs the various methods and processes described above, such as the tape resource recall method. For example, in some embodiments, the method of recall of a commodity resource may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 600 via the ROM 602 and/or the communication unit 609. When the computer program is loaded into RAM 603 and executed by computing unit 601, one or more steps of the method of recall of a commodity resource described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the tape resource recall method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These 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 the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

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

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A method of recall of a commodity resource, comprising:

responding to a received search request of a user, and acquiring a commodity resource matched with the search request and clicked by at least one similar user of the user;

calculating posterior conversion rates of the at least one similarly user-clicked on the commodity resource;

acquiring the average unit price of the goods mounted by the goods resources clicked by the at least one similar user;

For each similar user in the at least one similar user, determining the similarity between the user and the similar user as the point-to-display ratio of the goods resources clicked by the similar user;

for each goods-carrying resource clicked by the similar user, calculating thousands of display benefits of the goods-carrying resource according to the product of the posterior conversion rate of the goods-carrying resource, the average unit price of the mounted goods and the point-to-display ratio;

a predetermined number of the tape resources are recalled in the order of thousands of presentation benefits from high to low.

2. The method of claim 1, wherein the obtaining the hosted resource that matches the search request that was clicked by at least one similar user of the user comprises:

acquiring user portraits of all users;

calculating the similarity between the user portrait of the user and the user portraits of other users;

and querying the tape resources clicked by at least one similar user with the similarity higher than a preset similarity threshold value from the historical click records.

3. The method of claim 1, wherein prior to receiving the user's search request, the method further comprises:

constructing a user-resource graph model based on the history click record;

Obtaining an embedded vector of each user by training the graph model;

calculating cosine similarity between embedded vectors of any two users, and constructing indexes between the goods resources of each user and other users according to the cosine similarity;

and

the acquiring the on-demand resources which are clicked by at least one similar user of the user and matched with the search request comprises the following steps:

querying indexes of at least one similar user with cosine similarity with the user higher than a preset similarity threshold;

and acquiring the cargo resources according to the index of the at least one similar user.

4. The method of claim 1, wherein the calculating the posterior conversion of the at least one similar user clicked on the commodity resource comprises:

counting the distribution amount and conversion amount of each goods-carrying resource clicked by the at least one similar user;

for each of the commodity resources, the ratio of the conversion of the commodity resource to the distribution of the commodity resource is taken as the posterior conversion of the commodity resource.

5. The method of claim 1, wherein the calculating the posterior conversion of the at least one similar user clicked on the commodity resource comprises:

and for each goods-carrying resource, if the distribution amount of the goods-carrying resource is smaller than the preset confidence distribution amount, acquiring the distribution amount and conversion amount of the goods carried by the goods-carrying resource, and taking the ratio of the conversion amount of the goods carried to the distribution amount of the goods carried as the posterior conversion rate of the goods-carrying resource.

6. The method of claim 4, wherein the calculating the posterior conversion of the at least one similarly user clicked on the commodity resource comprises:

for each of the shipped resources, if the dispense amount of the shipped resource is greater than the predetermined dispense amount threshold and the posterior conversion is less than the predetermined conversion threshold, the posterior conversion of the shipped resource is set to 0.

7. The method of claim 1, wherein the obtaining the average unit price of the commodity on the loaded resource clicked by the at least one similar user comprises:

and for each goods with the goods clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods, and calculating the weighted average unit price of all goods mounted by the goods with the goods, wherein the weight of the goods which are already singly made is higher than that of the goods which are not singly made.

8. The method of claim 1, wherein the obtaining the average unit price of the commodity on the loaded resource clicked by the at least one similar user comprises:

and for each goods with the goods resources clicked by the at least one similar user, acquiring the unit price of each goods mounted by the goods with the goods resources, modifying the unit price of the goods with the unit price exceeding the upper limit price to the upper limit price, and then calculating the weighted average unit price of all the goods mounted by the goods with the goods resources, wherein the weight of the goods with the goods being made into a single item is higher than that of the goods without the goods with the goods being made into a single item.

9. The method of claim 1, wherein the recalling a predetermined number of the tape resources in a thousand presentation revenue from high to low order comprises:

for each of the similar user clicked on the shipping resources, a predetermined number of the shipping resources are selected in order of thousands of display benefits from high to low.

10. A recall device for a commodity resource, comprising:

a request unit configured to obtain, in response to receiving a search request from a user, a pickup resource that matches the search request and that has been clicked by at least one similar user of the user;

a computing unit configured to compute a posterior conversion of the at least one similarly user clicked on the commodity resource;

An obtaining unit configured to obtain an average unit price of the commodity mounted by the loaded resource clicked by the at least one similar user;

a determining unit configured to determine, for each of the at least one similar user, a similarity between the user and the similar user as a point spread ratio of the commodity resource clicked by the similar user;

a revenue unit configured to calculate, for each of the similar user clicked on the commodity-carrying resource, thousands of display revenue for the commodity-carrying resource based on a product of a posterior conversion rate of the commodity-carrying resource, an average unit price of the commodity-carrying and a point-to-display ratio;

a recall unit configured to recall a predetermined number of the tape resources in order of thousands of presentation benefits from high to low.

11. The apparatus of claim 10, wherein the requesting unit is further configured to:

acquiring user portraits of all users;

12. The apparatus of claim 10, wherein the apparatus further comprises a construction unit configured to:

Before receiving a search request of a user, constructing a graph model of the user-resource based on the historical click record;

obtaining an embedded vector of each user by training the graph model;

and

the requesting unit is further configured to:

13. The apparatus of claim 10, wherein the computing unit is further configured to:

14. The apparatus of claim 10, wherein the computing unit is further configured to:

15. The apparatus of claim 13, wherein the computing unit is further configured to:

16. The apparatus of claim 10, wherein the acquisition unit is further configured to:

17. The apparatus of claim 10, wherein the acquisition unit is further configured to:

18. The apparatus of claim 10, wherein the recall unit is further configured to:

19. An electronic device for recall of a commodity resource, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

20. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-9.

21. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-9.