Resource recommendation method and device, equipment and storage medium

Abstract

The present disclosure provides a resource recommendation method, device and equipment and storage medium, relates to the technical field of artificial intelligence, in particular to the technical field of information flow. The resource recommendation method comprises: determining the similarity between the candidate resource and the benchmark resource; determining the resource posterior conversion rate of the candidate resource; constructing the recommendation order of the plurality of candidate resources relative to the benchmark resource according to the similarity and the resource posterior conversion rate; in response to the benchmark resource being accessed, recommending the candidate resource according to the recommendation order. The resource recommendation method of the present disclosure constructs the recommendation order of the candidate resource through the similarity between the candidate resource and the benchmark resource and the resource posterior conversion rate of the candidate resource, takes into account the point display ratio and conversion rate of the candidate resource, ensures the consistency of the funnel in the recall stage and the sorting stage of the candidate resource, and improves the recommendation accuracy of the candidate resource.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular, to the technical field of artificial intelligence, and especially to a resource recommendation method and device, equipment, storage medium and computer program product. BACKGROUND

[0002] In an information flow recommendation system, resources usually go through two stages of recall and sorting. Among them, recall is to select several thousand resources from the provided tens of millions of resources, and sorting is to select dozens of resources from the several thousand resources recalled. In the sorting stage, both click-through rate and conversion rate are estimated, and then resources with high click-through rate and conversion rate are selected. However, in the recall stage, most of the recalls do not consider the conversion rate, and the selected resources are often only high in click-through rate, but not necessarily high in conversion rate, resulting in inconsistent funnel between recall and sorting, and thus reducing the recommendation accuracy of resources. SUMMARY

[0003] The present disclosure provides a resource recommendation method, device, equipment, storage medium and computer program product, which improves the resource recommendation accuracy.

[0004] According to a first aspect of the present disclosure, a resource recommendation method is provided, comprising:

[0005] determining the similarity between the candidate resource and the reference resource;

[0006] determining the resource posterior conversion rate of the candidate resource;

[0007] constructing the recommendation order of the plurality of candidate resources relative to the reference resource according to the similarity and the resource posterior conversion rate;

[0008] in response to the reference resource being accessed, recommending the candidate resource according to the recommendation order.

[0009] According to a second aspect of the present disclosure, a resource recommendation device is provided, comprising:

[0010] a first determination module configured to determine the similarity between the candidate resource and the reference resource;

[0011] a second determination module configured to determine the posterior conversion rate of the candidate resource;

[0012] a construction module configured to construct the recommendation order of the plurality of candidate resources relative to the reference resource according to the similarity and the resource posterior conversion rate;

[0013] a recommendation module configured to, in response to the reference resource being accessed, recommend the candidate resource according to the recommendation order.

[0014] According to a third aspect of the present disclosure, an electronic device is provided, comprising:

[0015] at least one processor; and

[0016] a memory communicatively connected with the at least one processor; wherein

[0017] 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 provided in the first aspect.

[0018] According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method provided in the first aspect.

[0019] 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 provided in the first aspect.

[0020] It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0021] The accompanying drawings are used to better understand the present scheme, and do not limit the present disclosure. Among them:

[0022] Figure 1 An exemplary system architecture to which the resource recommendation method of the present disclosure can be applied is shown;

[0023] Figure 2 A flowchart of an embodiment of the resource recommendation method according to the present disclosure is shown;

[0024] Figure 3 A flowchart of another embodiment of the resource recommendation method according to the present disclosure is shown;

[0025] Figure 4 A structural diagram of an embodiment of the resource recommendation device according to the present disclosure is shown;

[0026] Figure 5 A block diagram of an electronic device for implementing the resource recommendation method of the present disclosure is shown. DETAILED DESCRIPTION

[0027] Exemplary embodiments of the present disclosure are described herein with reference to the accompanying drawings, which are cited by way of example only. The present disclosure is therefore not limited to the embodiments described herein, but encompasses all embodiments within the scope of the present disclosure. As such, various changes and modifications can be suggested to one skilled in the art, and it is intended that the present disclosure encompass such changes and modifications as fall within the scope of the appended claims. Also, for the purpose of clarity and a concise description, the detailed description term used in the following description will be described below.

[0028] It should be noted that the embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In an information flow recommendation system, a product-carrying resource is a relatively special resource. The product-carrying resource itself is a carrier of content, and in addition, it inserts a product related to the content at a position in the content, such as a middle or end position, for a user in need to purchase. For example, an article resource explaining badminton training techniques has a purchase link of a certain brand of badminton at the end of the article. After reading the article, the user may directly purchase the badminton through the link at the end of the article, so that the user's deeper experience is satisfied, the creator of the article can obtain a share of the profits, and the platform providing the article can also obtain a share of the profits, which is a win-win mode. Therefore, the product-carrying resource is highly valued by various resource distribution platforms.

[0030] The point-to-show ratio refers to the ratio of the number of clicks to the number of displays of a resource, which is an important measure for various resources. In addition to considering the point-to-show ratio, the conversion rate, i.e., the probability that a product carried in a product-carrying resource is purchased, also needs to be considered for a product-carrying resource.

[0031] In an information flow recommendation system, when performing icf (Item Collaborative Filtering) recall on a product-carrying resource, only the point-to-show ratio is considered, and the conversion rate is not considered. In the sorting stage, both the point-to-show ratio and the conversion rate are considered, which leads to inconsistent funnels in the recall and sorting stages, resulting in low recommendation accuracy of the product-carrying resource.

[0032] The present disclosure provides a resource recommendation method for determining the similarity between a candidate resource and a reference resource and the resource posterior conversion rate of the candidate resource. In the recall stage, a recommendation order of multiple candidate resources is constructed according to the similarity and the resource posterior conversion rate. When the reference resource is accessed, the candidate information is recommended according to the recommendation order. The point-to-show ratio and the resource posterior conversion rate of the candidate resource are considered at the same time, which ensures that the funnels in the recall and sorting stages are consistent, and effectively improves the recommendation accuracy of the candidate resource.

[0033] Figure 1 An exemplary system architecture 100 is shown, which can apply the embodiments of the resource recommendation method or the resource recommendation device of the present disclosure.

[0034] As shown in Figure 1 System architecture 100 can include terminal device 101, network 102 and server 103. Network 102 is used to provide communication link between terminal device 101 and server 103, and can include various connection types, such as wired communication link, wireless communication link or optical cable, etc.

[0035] User can use terminal device 101 to interact with server 103 through network 102 to receive or send information, etc. Various client applications can be installed on terminal device 101.

[0036] Terminal device 101 can be hardware or software. When terminal device 101 is hardware, it can be various electronic devices, including but not limited to smart phone, tablet computer, laptop computer and desktop computer, etc. When terminal device 101 is software, it can be installed in the above-mentioned electronic devices. It can be implemented as multiple software or software modules, or as a single software or software module. No specific limitation is made here.

[0037] Server 103 can be hardware or software. When server 103 is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server. When server 103 is software, it can be implemented as multiple software or software modules (for example, to provide distributed services), or as a single software or software module. No specific limitation is made here.

[0038] The resource recommendation method provided by the embodiments of the present disclosure is generally executed by server 103, and correspondingly, the resource recommendation apparatus is generally provided in server 103.

[0039] It should be noted that Figure 1 The number of terminal devices 101, networks 102 and servers 103 in

[0040] In the embodiments of the present disclosure, the resource recommendation method is executed by server 103, and the recommended candidate resources are sent to terminal device 101 installed with client, for example, the recommended candidate resources are sent to the terminal device where the user views the candidate resources, or the recommended candidate resources are sent to terminal device 101 installed with resource distribution platform.

[0041] Figure 2 Flow 200 of one embodiment of the resource recommendation method according to the present disclosure is shown. Referring to Figure 2 The resource recommendation method includes the following steps:

[0042] Step S201, determine the similarity between the candidate resource and the reference resource.

[0043] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, determines the similarity between the candidate resource and the reference resource.

[0044] In some optional embodiments, the execution subject determines the reference resource and the candidate resource in a resource set.

[0045] For example, a resource set includes several goods-carrying resources, any of which can be used as a reference resource. When one of the goods-carrying resources is determined as a reference resource, the remaining goods-carrying resources are candidate resources.

[0046] Then, the execution subject determines the similarity between the candidate resource and the reference resource according to the determined reference resource and candidate resource.

[0047] In some optional implementations of the embodiments of the present disclosure, determining the similarity between the candidate resource and the reference resource includes: using a pre-trained model to obtain a feature vector of the candidate resource and a feature vector of the reference resource; and determining the cosine similarity between the feature vector of the candidate resource and the feature vector of the reference resource as the similarity between the candidate resource and the reference resource.

[0048] In this embodiment, the execution subject obtains the feature vectors of the candidate resource and the reference resource through the pre-trained model, and calculates the cosine similarity between them according to the feature vectors as the similarity between them, which guarantees the similarity accuracy of the candidate resource and the reference resource and provides a stable basis for improving the recommendation order accuracy of the candidate resource.

[0049] The pre-trained model can be a relationship model between the user and the goods-carrying resource. For example, an initial model is constructed according to the relationship between the user and the goods-carrying resource, and after sample data training, a pre-trained model is generated. For example, the initial model can be a graph model of the relationship between the user and the goods-carrying resource.

[0050] In some optional embodiments, the feature vector of the candidate resource and the feature vector of the reference resource obtained by the execution subject are embedding vectors. For example, the feature vector of the candidate resource and the feature vector of the reference resource can be 32-dimensional feature vectors or 100-dimensional feature vectors.

[0051] The execution subject calculates the cosine similarity between each candidate resource and the reference resource through the feature vector. The higher the dimension of the feature vector, the higher the accuracy of the calculated cosine similarity. The higher the cosine similarity value between a candidate resource and the reference resource, the higher the similarity between the candidate resource and the reference resource, that is, the more similar the candidate resource and the reference resource, which sufficiently guarantees the calculation reliability of the similarity between the candidate resource and the reference resource.

[0052] In step S202, the resource posterior conversion rate of the candidate resource is determined.

[0053] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, determines the resource posterior conversion rate of the candidate resource.

[0054] For example, the resource posterior conversion rate of the candidate resource refers to the actual calculation probability of the purchase of the mounted goods according to the historical conversion amount and the historical distribution amount of the candidate resource within a certain period of time.

[0055] The execution subject can determine the resource posterior conversion rate of the candidate resource according to the actual distribution amount and the actual conversion amount of the candidate resource within a certain period of time.

[0056] In step S203, the recommendation order of the plurality of candidate resources relative to the reference resource is constructed according to the similarity and the resource posterior conversion rate.

[0057] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, constructs the recommendation order of the plurality of candidate resources relative to the reference resource according to the similarity determined in step S201 and the resource posterior conversion rate determined in step S202.

[0058] The execution subject considers the similarity between the candidate resource and the reference resource and the resource posterior conversion rate of the candidate resource when constructing the recommendation order of the plurality of candidate resources, which can improve the reliability of the recommendation order of the plurality of candidate resources.

[0059] In some optional embodiments, the similarity and the posterior conversion rate can be respectively given corresponding weights according to actual conditions, and the recommendation order of the plurality of candidate resources relative to the reference resource is constructed according to the similarity and the weight thereof and the posterior conversion rate and the weight thereof. For example, the plurality of candidate resources can be sorted according to the sum of the product of the similarity and the weight thereof and the product of the posterior conversion rate and the weight thereof in descending order to construct the recommendation order of the plurality of candidate resources.

[0060] In some optional implementation of the embodiments of the present disclosure, the recommendation order of the plurality of candidate resources relative to the benchmark resource is constructed according to the similarity and the resource posterior conversion rate, including: determining the product of the similarity and the resource posterior conversion rate as the recommendation score of the candidate resource; and constructing the recommendation order of the plurality of candidate resources relative to the benchmark resource according to the preset order of the recommendation score.

[0061] In the embodiment, the recommendation order of the plurality of candidate resources relative to the benchmark resource is constructed according to the order of the product of the similarity and the resource posterior conversion rate, and the point display ratio and the conversion rate of the candidate resource are considered, so that the point display ratio and the conversion rate of the candidate resource in the front of the recommendation order are both high, the situation of high point display ratio and low conversion rate is avoided, and the recommendation accuracy of the candidate resource is effectively ensured.

[0062] For example, the benchmark resource is nid1, and the candidate resources are nid2, nid3, nid4, …, if the similarity of the candidate resources nid2, nid3, nid4 and the benchmark resource nid1 is 0.9, 0.8, 0.7 respectively, and the posterior conversion rate of nid2, nid3, nid4 is 0.01, 0.02, 0.03 respectively. If the posterior conversion rate is not considered, the recommendation order of the candidate resources relative to the benchmark resource is nid1[nid2&0.9, nid3&0.8, nid4&0.7, …]; and according to the scheme of the present disclosure, the product of the similarity and the resource posterior conversion rate is used as the recommendation score, so that the recommendation order of the candidate resources relative to the benchmark resource is nid1[nid4&0.021, nid3&0.016, nid2&0.009, …]. It can be seen that after considering the resource posterior conversion rate, the order of the candidate resources nid2 and nid4 is changed, and the reliability of the recommendation order is higher compared with the scheme of considering only the similarity.

[0063] In step S204, in response to the benchmark resource being accessed, the candidate resource is recommended according to the recommendation order.

[0064] In the embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, in response to the benchmark resource being accessed, recommends the candidate resource according to the recommendation order constructed in step S203.

[0065] That is, in the embodiment, when the benchmark resource is accessed, the candidate resource is recommended according to the recommendation score order determined relative to the benchmark resource.

[0066] In some optional implementations of the embodiments of the present disclosure, the candidate resources are recommended according to the recommendation order, including: determining a recommendation score threshold and a recommendation quantity; in response to the recommendation score of the candidate resource being greater than or equal to the recommendation score threshold and the ranking of the candidate resource in the recommendation order being within the recommendation quantity, determining the candidate resource as a target resource; and recommending the target resource according to the recommendation order.

[0067] In the present embodiment, the execution subject determines the target resource in the plurality of candidate resources according to the recommendation score threshold and the recommendation quantity, and then recommends the target resource according to the recommendation order, thereby further ensuring the similarity between the recommended candidate resource and the benchmark resource and improving the recommendation accuracy of the candidate resource.

[0068] In some example embodiments, the execution subject can determine the average of the recommendation scores of the plurality of candidate resources as the recommendation score threshold. When the benchmark resource is accessed, the candidate resource with a recommendation score greater than or equal to the recommendation score threshold is determined as the target resource, and then the target resource is recommended according to the recommendation order.

[0069] The recommendation quantity can be selected and set according to the actual situation (such as the running situation, etc.) of the execution subject. For example, the recommendation quantity can be 30-200, such as 30 or 50 or 80 or 120, etc.

[0070] For example, if the recommendation quantity is set to 50 and there are 1000 candidate resources, the candidate resources ranked in the top 50 in the recommendation order are determined as the target resources, and then the target resources are recommended according to the recommendation order.

[0071] In some optional embodiments, if the number of candidate resources with a recommendation score greater than or equal to the recommendation score threshold is 230 in 1000 candidate resources and the recommendation quantity is 50, the candidate resources ranked in the top 50 in the recommendation order are determined as the target resources.

[0072] In some optional embodiments, if the number of candidate resources with a recommendation score greater than or equal to the recommendation score threshold is 60 in 800 candidate resources and the recommendation quantity is 80, the candidate resources ranked in the top 60 in the recommendation order are determined as the target resources.

[0073] In the resource recommendation method provided by the embodiments of the present disclosure, the recommendation order of the candidate resources is constructed by the similarity between the candidate resources and the benchmark resource and the resource posterior conversion rate of the candidate resources, the click-through rate and the conversion rate of the candidate resources are considered, the funnel consistency of the candidate resources in the recall stage and the ranking stage is ensured, and the recommendation accuracy of the candidate resources is improved.

[0074] Figure 3Flow 300 shows an embodiment of the resource recommendation method according to the present disclosure. Refer to Figure 3 As shown, the resource recommendation method comprises the following steps:

[0075] Step S301, determine the similarity between the candidate resource and the reference resource.

[0076] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, determines the similarity between the candidate resource and the reference resource.

[0077] Step S301 is basically consistent with Step S201 of the embodiment shown above, and the specific implementation manner can refer to the foregoing description of Step S201, which will not be described here again. Figure 2 Step S201 of the embodiment shown above, and the specific implementation manner can refer to the foregoing description of Step S201, which will not be described here again.

[0078] Step S302, obtain the resource distribution amount and the resource conversion amount of the candidate resource.

[0079] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, obtains the resource distribution amount and the resource conversion amount of the candidate resource.

[0080] In some optional embodiments, the execution subject described above obtains the resource distribution amount and the resource conversion amount of each candidate resource within a preset time length. The resource distribution amount is the number of the candidate resource distributed within the preset time length; and the resource conversion amount is the number of the goods mounted by the candidate resource purchased within the preset time length.

[0081] For example, if a candidate resource mounts multiple goods, the resource conversion amount of the candidate resource is the sum of the number of the multiple goods mounted by the candidate resource purchased within the preset time length.

[0082] Step S303, determine the confidence distribution amount of the candidate resource.

[0083] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, determines the confidence distribution amount of the candidate resource.

[0084] The confidence distribution amount is subsequently used as the measurement standard of the resource posterior conversion rate of the candidate resource to ensure the reliability of the resource posterior conversion rate.

[0085] In some optional embodiments, the confidence distribution amount can be determined according to the average conversion rate of multiple candidate resources. Because the average conversion rate of multiple candidate resources comprehensively considers the resource distribution amount and the resource conversion amount thereof, the reliability of the confidence distribution amount can be greatly improved by inversely deducing the confidence distribution amount according to the average conversion rate.

[0086] Exemplarily, according to the resource distribution quantity and the resource conversion quantity of the candidate resource within the preset time length, a ratio of the resource conversion quantity to the resource distribution quantity of the candidate resource within the preset time length can be calculated as a resource conversion rate of the candidate resource within the preset time length.

[0087] According to the average value of the resource conversion rates of the plurality of candidate resources within the preset time length, an average conversion rate of the plurality of candidate resources within the preset time length can be obtained.

[0088] Exemplarily, assuming that the average conversion rate of the plurality of candidate resources within the preset time length is 1%, the confidence distribution quantity can be determined as 1000. According to the average conversion rate calculation, when the resource distribution quantity is 1000, the resource conversion quantity is 10.

[0089] In step S304, the resource posterior conversion rate of the candidate resource is determined according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity.

[0090] In the embodiment, the execution subject of the resource recommendation method, for example, a terminal device or a server installed with a resource distribution platform client, determines the resource posterior conversion rate of the candidate resource according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity.

[0091] The execution subject determines the resource posterior conversion rate of the candidate resource according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity in a manner to repeatedly guarantee the reliability of the resource posterior conversion rate.

[0092] In some optional implementation manners of the embodiments of the present disclosure, determining the resource posterior conversion rate of the candidate resource according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity includes: in response to the resource distribution quantity being greater than or equal to the confidence distribution quantity, determining a ratio of the resource conversion quantity to the resource distribution quantity as the resource posterior conversion rate; and in response to the resource distribution quantity being less than the confidence distribution quantity, determining a commodity posterior conversion rate of a commodity mounted in the candidate resource as the resource posterior conversion rate.

[0093] In the embodiment, the execution subject determines the resource posterior conversion rate of the candidate resource according to the resource distribution quantity and the confidence distribution quantity. When the resource distribution quantity is greater than or equal to the confidence distribution quantity, the resource conversion rate is directly taken as the resource posterior conversion rate, that is, the resource posterior conversion rate is a ratio of the resource conversion quantity to the resource distribution quantity; and when the resource distribution quantity is less than the confidence distribution quantity, that is, when the resource distribution quantity is small, the commodity posterior conversion rate of a commodity mounted in the candidate resource is taken as the resource posterior conversion rate.

[0094] In the embodiment, the confidence distribution amount is taken as the determination criterion, the resource conversion rate or the commodity posterior conversion rate is determined as the resource posterior conversion rate according to different resource distribution amounts by comparing the resource distribution amount with the confidence distribution amount, and the reliability of the resource posterior conversion rate is effectively improved.

[0095] In some optional implementation of the embodiment of the present disclosure, the commodity posterior conversion rate of the commodity mounted in the candidate resource is determined as the resource posterior conversion rate, comprising: obtaining the commodity distribution amount and the commodity conversion amount of the commodity mounted in the candidate resource; and determining the ratio of the commodity conversion amount to the commodity distribution amount as the resource posterior conversion rate.

[0096] In the embodiment, the execution subject obtains the commodity distribution amount and the commodity conversion amount of the commodity mounted in the candidate resource, and then determines the ratio of the commodity conversion amount to the commodity distribution amount as the commodity posterior conversion rate. In response to the resource distribution amount of the candidate resource being less than the confidence distribution amount, the execution subject determines the commodity posterior conversion rate as the resource posterior conversion rate.

[0097] For example, the execution subject obtains the commodity distribution amount and the commodity conversion amount of the commodity mounted in the candidate resource within a preset time length. The commodity distribution amount is the sum of the distribution amounts of all candidate resources mounting the commodity, and the commodity conversion amount is the sum of the conversion amounts of the commodity in all candidate resources mounting the commodity.

[0098] The execution subject calculates the ratio of the commodity conversion amount to the commodity distribution amount of the commodity mounted in the candidate resource within the preset time length as the commodity posterior conversion rate of the commodity mounted in the candidate resource within the preset time length.

[0099] When it is determined that the resource distribution amount of the candidate resource is less than the confidence distribution amount, the execution subject determines the commodity posterior conversion rate as the resource posterior conversion rate of the candidate resource, thereby improving the reliability of the resource posterior conversion rate.

[0100] In some optional implementation of the embodiment of the present disclosure, the resource posterior conversion rate of the candidate resource is determined according to the confidence distribution amount, the resource distribution amount and the resource conversion amount, and further comprising: determining the average posterior conversion rate of the plurality of candidate resources according to the resource posterior conversion rates of the plurality of candidate resources; and in response to the resource distribution amount of one candidate resource being greater than or equal to the first preset multiple of the confidence distribution amount and the average posterior conversion rate being greater than or equal to the second preset multiple of the resource posterior conversion rate of the candidate resource, setting the posterior conversion rate of the candidate resource to 0.

[0101] In the embodiment, after the execution subject determines the resource posterior conversion rate of each candidate resource, the average posterior conversion rate of the plurality of candidate resources is determined, and each candidate resource is evaluated based on the average posterior conversion rate, thereby further improving the reliability of each candidate resource.

[0102] In some optional embodiments, the execution subject punishes the candidate resource with ultra-high distribution volume and ultra-low conversion rate.

[0103] For example, if the resource distribution volume of a candidate resource is much larger than the confidence distribution volume, and the resource posterior conversion rate of the candidate resource is much smaller than the average posterior conversion rate, it can be considered that the candidate resource is not suitable for mounting goods, and therefore, to avoid wasting resources, the posterior conversion rate of the candidate resource can be directly set to 0.

[0104] For example, in response to the resource distribution volume of a candidate resource being greater than or equal to a first preset multiple of the confidence distribution volume, and the average posterior conversion rate being greater than or equal to a second preset multiple of the resource posterior conversion rate of the candidate resource, the posterior conversion rate of the candidate resource is set to 0. The first preset multiple is greater than or equal to 10 times, and the second preset multiple is greater than or equal to 10 times. The first preset multiple and the second preset multiple can be the same or different.

[0105] For example, when the confidence distribution volume is 1000 and the average posterior conversion rate is 1%, if the resource distribution volume of a candidate resource is 10w, but its resource conversion volume is only 10. If calculated according to the average posterior conversion rate, the theoretical value of the resource conversion volume should be 1000. It can be seen that the resource distribution volume of the candidate resource is 100 times the confidence distribution volume, and the average posterior conversion rate is 100 times the resource posterior conversion rate of the candidate resource. The candidate resource is a candidate resource with high distribution volume and low conversion rate, and it can be determined that the candidate resource is not suitable for mounting goods, and therefore, the resource posterior conversion rate of the candidate resource can be directly set to 0.

[0106] Step S305, according to the similarity and the resource posterior conversion rate, constructing a recommended order of the plurality of candidate resources relative to the reference resource.

[0107] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device installed with a resource distribution platform client, constructs a recommended order of a plurality of candidate resources relative to a reference resource according to the similarity determined in step S301 and the resource posterior conversion rate determined in step S304.

[0108] Step S305 is basically the same as step S203 of the embodiment shown in Figure 2 The specific implementation manner can refer to the foregoing description of step S203, and will not be described here.

[0109] Step S306, in response to the reference resource being accessed, recommending the candidate resource according to the recommended order.

[0110] In this embodiment, the execution subject of the resource recommendation method, for example, a terminal device installed with a resource distribution platform client, recommends candidate resources according to the recommendation order determined in step S305 in response to the reference resource being accessed.

[0111] Step S306 is basically the same as step S204 of the embodiment shown in Figure 2 The specific implementation manner can refer to the foregoing description of step S204, and details are not described herein.

[0112] In the resource recommendation method provided in this embodiment of the present disclosure, the confidence distribution quantity is determined, the resource posterior conversion rate is determined based on the confidence distribution quantity, the reliability of the resource posterior conversion rate is improved, the reliability of the recommendation order of candidate resources is further ensured, and the recommendation accuracy of candidate resources is further improved.

[0113] It should be noted that the collection, storage, use, processing, transmission, provision and disclosure of user-related information involved in the technical solutions of the present disclosure comply with relevant laws and regulations and do not violate public order and good customs. The user-related information obtained in this embodiment is not directed to a specific user, and cannot reflect the personal information of a specific user.

[0114] As an implementation of the method shown in each of the foregoing figures, Figure 4 An embodiment of a resource recommendation device according to the present disclosure is shown. The resource recommendation device is used in Figure 2 The device can be applied to various electronic devices corresponding to the method embodiment shown in

[0115] Referring to Figure 4 The resource recommendation device 400 provided in this embodiment of the present disclosure includes a first determination module 401, a second determination module 402, a construction module 403 and a recommendation module 404. The first determination module 401 is configured to determine the similarity between candidate resources and a reference resource. The second determination module 402 is configured to determine the posterior conversion rate of the candidate resources. The construction module 403 is configured to construct the recommendation order of the plurality of candidate resources relative to the reference resource according to the similarity and the resource posterior conversion rate. The recommendation module 404 is configured to recommend candidate resources according to the recommendation order in response to the reference resource being accessed.

[0116] In this embodiment, the specific processing of the first determination module 401, the second determination module 402, the construction module 403 and the recommendation module 404 in the resource recommendation device 400 and the technical effects brought by the specific processing can be respectively referred to the foregoing descriptions of steps S201-S204 in the corresponding embodiments, and details are not described herein. Figure 2 The specific processing of the first determination module 401, the second determination module 402, the construction module 403 and the recommendation module 404 in the resource recommendation device 400 and the technical effects brought by the specific processing can be respectively referred to the foregoing descriptions of steps S201-S204 in the corresponding embodiments, and details are not described herein.

[0117] In some optional implementation of the embodiment of the present disclosure, the second determining module 402 comprises an obtaining sub-module, a first determining sub-module and a second determining sub-module. The obtaining sub-module is configured to obtain the resource distribution quantity and the resource conversion quantity of the candidate resource; the first determining sub-module is configured to determine the confidence distribution quantity of the candidate resource; and the second determining sub-module is configured to determine the resource posterior conversion rate of the candidate resource according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity.

[0118] In the embodiment, the specific process of the obtaining sub-module, the first determining sub-module and the second determining sub-module in the resource recommendation device 400 and the technical effects brought by the specific process can be referred to the related description of steps S302-S304 in the corresponding embodiment respectively. Figure 3 The related description of steps S302-S304 in the corresponding embodiment is not repeated here.

[0119] In some optional implementation of the embodiment of the present disclosure, the second determining sub-module comprises a first determining unit and a second determining unit. The first determining unit is configured to determine the ratio of the resource conversion quantity to the resource distribution quantity as the resource posterior conversion rate in response to the resource distribution quantity being greater than or equal to the confidence distribution quantity; and the second determining unit is configured to determine the product of the resource posterior conversion rate of the candidate resource and the average posterior conversion rate of the multiple candidate resources as the recommendation score of the candidate resource.

[0120] In some optional implementation of the embodiment of the present disclosure, the second determining unit is configured to obtain the product of the resource posterior conversion rate of the candidate resource and the average posterior conversion rate of the multiple candidate resources as the recommendation score of the candidate resource.

[0121] In some optional implementation of the embodiment of the present disclosure, the second determining sub-module further comprises a determining unit and a configuring unit. The determining unit is configured to determine the average posterior conversion rate of the multiple candidate resources according to the resource posterior conversion rates of the multiple candidate resources; and the configuring unit is configured to set the posterior conversion rate of a candidate resource to 0 in response to the resource distribution quantity of the candidate resource being greater than or equal to a first preset multiple of the confidence distribution quantity and the average posterior conversion rate being greater than or equal to a second preset multiple of the resource posterior conversion rate of the candidate resource.

[0122] In some optional implementation of the embodiment of the present disclosure, the constructing module is configured to determine the product of the similarity and the resource posterior conversion rate as the recommendation score of the candidate resource; and construct the recommendation order of the multiple candidate resources relative to the benchmark resource in a preset order of the recommendation score.

[0123] In some optional implementation of the embodiments of the present disclosure, the recommendation module is configured to: determine a recommendation score threshold and a recommendation quantity; in response to the recommendation score of the candidate resource being greater than or equal to the recommendation score threshold and the ranking of the candidate resource in the recommendation sequence being within the recommendation quantity, determine the candidate resource as a target resource; and recommend the target resource according to the recommendation sequence.

[0124] In some optional implementation of the embodiments of the present disclosure, the first determination module is configured to: obtain the feature vector of the candidate resource and the feature vector of the reference resource by using the pre-trained model; and determine the cosine similarity between the feature vector of the candidate resource and the feature vector of the reference resource as the similarity between the candidate resource and the reference resource.

[0125] According to the embodiments of the present disclosure, the present disclosure further provides an electronic device, a non-transitory computer-readable storage medium storing computer instructions, and a computer program product.

[0126] The electronic device includes at least one processor, and a memory connected with the at least one processor in communication; 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 resource recommendation method.

[0127] In some embodiments, a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the resource recommendation method.

[0128] In some embodiments, a computer program product includes a computer program that, when executed by a processor, implements the resource recommendation method.

[0129] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present disclosure described and / or claimed in this document.

[0130] As Figure 5As shown, the device 500 includes a computing unit 501 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0131] A plurality of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0132] The computing unit 501 can be various general and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 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 appropriate processor, controller, microcontroller, etc. The computing unit 501 performs various methods and processes described above, such as the backbone network generation method or the image processing method. For example, in some embodiments, the file processing method can be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the backbone network generation method or the image processing method described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the resource recommendation method by any other appropriate means, such as by means of firmware.

[0133] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0134] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or the block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, or entirely on a remote machine or server.

[0135] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can 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 the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0136] 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 the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; 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 acoustic, speech, or tactile input.

[0137] The systems and techniques described here can be implemented in a computing system that includes a back end 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 back end, 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 a local area network (LAN), a wide area network (WAN), and the Internet.

[0138] The computer system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server is generally established by computer programs running on the respective computers and having a client-server relationship to each other. The servers can be cloud servers, servers of a distributed system, or servers combined with a blockchain.

[0139] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in a different order, without departing from the desired results of the technology disclosed in the present disclosure, which are not limited herein.

[0140] The specific embodiments described above are not intended to limit the scope of the present disclosure. Those skilled in the art will understand that various modifications, combinations, sub-combinations and alternatives can be made to the specific embodiments without departing from the spirit and principles of the present disclosure. Any further modifications, equivalent substitutions, improvements, combinations and the like not described above are also intended to be encompassed within the scope of the present disclosure.

Claims

1. A resource recommendation method, comprising: determining a similarity between a candidate resource and a benchmark resource; determining a resource posterior conversion rate of the candidate resource, comprising: obtaining a resource distribution quantity and a resource conversion quantity of the candidate resource; determining a confidence distribution quantity of the candidate resource; in response to the resource distribution quantity being greater than or equal to the confidence distribution quantity, determining a ratio of the resource conversion quantity to the resource distribution quantity as the resource posterior conversion rate; in response to the resource distribution quantity being less than the confidence distribution quantity, determining a product posterior conversion rate of a product mounted in the candidate resource as the resource posterior conversion rate; constructing a recommendation order of a plurality of the candidate resources relative to the benchmark resource according to the similarity and the resource posterior conversion rate; in response to the benchmark resource being accessed, recommending the candidate resources according to the recommendation order.

2. The resource recommendation method of claim 1, wherein, The determination of the product posterior conversion rate of the product mounted in the candidate resource as the resource posterior conversion rate comprises: obtaining a product distribution quantity and a product conversion quantity of the product mounted in the candidate resource; determining a ratio of the product conversion quantity to the product distribution quantity as the resource posterior conversion rate.

3. The resource recommendation method of claim 1, wherein, The determination of the resource posterior conversion rate of the candidate resource according to the confidence distribution quantity, the resource distribution quantity and the resource conversion quantity further comprises: determining an average posterior conversion rate of a plurality of the candidate resources according to resource posterior conversion rates of the plurality of the candidate resources; in response to the resource distribution quantity of one of the candidate resources being greater than or equal to a first preset multiple of the confidence distribution quantity, and the average posterior conversion rate being greater than or equal to a second preset multiple of the resource posterior conversion rate of the candidate resource, setting the posterior conversion rate of the candidate resource to 0.

4. The resource recommendation method according to any one of claims 1 to 3, wherein, The construction of the recommendation order of the plurality of the candidate resources relative to the benchmark resource according to the similarity and the resource posterior conversion rate comprises: determining a product of the similarity and the resource posterior conversion rate as a recommendation score of the candidate resource; constructing the recommendation order of the plurality of the candidate resources relative to the benchmark resource in a preset order of the recommendation score.

5. The resource recommendation method of claim 4, wherein, The recommendation of the candidate resources according to the recommendation order comprises: determining a recommendation score threshold and a recommendation quantity; in response to the recommendation score of the candidate resource being greater than or equal to the recommendation score threshold, and an order of the candidate resource in the recommendation order being within the recommendation quantity, determining the candidate resource as a target resource; recommending the target resource according to the recommendation order.

6. The resource recommendation method according to any one of claims 1 to 3, wherein, The determination of the similarity between the candidate resource and the benchmark resource comprises: obtaining a feature vector of the candidate resource and a feature vector of the benchmark resource by using a pre-trained model; determining a cosine similarity between the feature vector of the candidate resource and the feature vector of the benchmark resource as the similarity between the candidate resource and the benchmark resource.

7. A resource recommendation apparatus, comprising: a first determination module configured to determine a similarity between a candidate resource and a benchmark resource; The second determining module is configured to determine the posterior conversion rate of the candidate resource, comprising: an obtaining submodule configured to obtain a resource distribution quantity and a resource conversion quantity of the candidate resource; a first determining submodule configured to determine a confidence distribution quantity of the candidate resource; and a second determining submodule, comprising: a first determining unit configured to, in response to the resource distribution quantity being greater than or equal to the confidence distribution quantity, determine a ratio of the resource conversion quantity to the resource distribution quantity as the resource posterior conversion rate; and a second determining unit configured to, in response to the resource distribution quantity being less than the confidence distribution quantity, determine a product posterior conversion rate of a product mounted in the candidate resource as the resource posterior conversion rate; The constructing module is configured to construct a recommendation order of the candidate resources relative to the benchmark resource according to the similarity and the resource posterior conversion rate; The recommendation module is configured to, in response to the benchmark resource being accessed, recommend the candidate resources according to the recommendation order.

8. The resource recommendation apparatus according to claim 7, wherein The second determining unit is configured to: obtain a product distribution quantity and a product conversion quantity of a product mounted in the candidate resource; determine a ratio of the product conversion quantity to the product distribution quantity as the resource posterior conversion rate.

9. The resource recommendation apparatus according to claim 7, wherein The second determining submodule further comprises: a determining unit configured to determine an average posterior conversion rate of the candidate resources according to resource posterior conversion rates of the candidate resources; and a configuring unit configured to, in response to the resource distribution quantity of one of the candidate resources being greater than or equal to a first preset multiple of the confidence distribution quantity and the average posterior conversion rate being greater than or equal to a second preset multiple of the resource posterior conversion rate of the candidate resource, set the posterior conversion rate of the candidate resource to 0.

10. The resource recommendation apparatus according to any one of claims 7 to 9, wherein The constructing module is configured to: determine a product of the similarity and the resource posterior conversion rate as a recommendation score of the candidate resource; and construct a recommendation order of the candidate resources relative to the benchmark resource according to a preset order of the recommendation scores.

11. The resource recommendation apparatus of any of claims 10, wherein, The recommendation module is configured to: determine a recommendation score threshold and a recommendation quantity; in response to the recommendation score of the candidate resource being greater than or equal to the recommendation score threshold and an order of the candidate resource in the recommendation order being within the recommendation quantity, determine the candidate resource as a target resource; and recommend the target resource according to the recommendation order.

12. The resource recommendation apparatus according to any one of claims 7 to 9, wherein The first determining module is configured to: obtain a feature vector of the candidate resource and a feature vector of the benchmark resource by using a pre-trained model; and determine a cosine similarity between the feature vector of the candidate resource and the feature vector of the benchmark resource as the similarity between the candidate resource and the benchmark resource.

13. An electronic device, comprising: at least one processor; and a memory connected with the at least one processor in communication; 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 to 6.

14. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 6.

15. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images