Object position determination method and apparatus, electronic device, and storage medium

Abstract

This application provides a method, apparatus, electronic device, and computer storage medium for determining the location of an object. The method is applied to a recommendation system and includes: acquiring historical interaction information of multiple target users; the historical interaction information includes interaction information related to recommendation sections and objects in the system; based on the acquired historical interaction information, determining a first preference score for each target user on each recommendation section in the system, and a second preference score for each object; matching each recommendation section and each object in the system according to the first preference score and the second preference score to obtain at least two matched objects; the matched object represents a recommendation section that successfully matches an object; and determining the location information of each of the at least two matched objects.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and to, but is not limited to, a method, apparatus, electronic device, and computer storage medium for determining the location of an object. Background Technology

[0002] Currently, with the rapid development of information and internet technologies, more and more users are browsing products of interest through internet platforms or applications. However, due to the rapidly increasing amount of information users encounter daily, it is becoming increasingly difficult for users to obtain information about objects of interest directly from the internet. Therefore, identifying the objects of interest to users and determining their location to facilitate interactive operations is a pressing technical problem that needs to be solved. Summary of the Invention

[0003] In view of the problems in related technologies, embodiments of this application provide a method, apparatus, electronic device and computer storage medium for determining the location of an object.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] This application provides a method for determining the location of an object, applied to a recommendation system, including:

[0006] Acquire historical interaction information of multiple target users; the historical interaction information includes interaction information related to the recommended sections and objects in the system.

[0007] Based on the acquired historical interaction information, determine the first preference score of each target user for each recommended section in the system, and the second preference score for each object;

[0008] Based on the first preference score and the second preference score, each recommendation section in the system is matched with each object to obtain at least two matching objects; the matching object represents the recommendation section that is successfully matched with the object.

[0009] Determine the position information of each of the at least two matching objects.

[0010] This application provides an object location determination device, applied to a recommendation system, comprising:

[0011] The acquisition module is used to acquire historical interaction information of multiple target users; the historical interaction information includes interaction information related to the recommended sections and objects in the system.

[0012] The first determining module is used to determine, based on the acquired historical interaction operation information, the first preference score of each target user for each recommendation section in the system, and the second preference score for each object.

[0013] The matching module is used to match each recommendation section and each object in the system according to the first preference score and the second preference score to obtain at least two matching objects; the matching object represents the recommendation section that is successfully matched with the object;

[0014] The second determining module is used to determine the position information of each of the at least two matching objects.

[0015] This application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described method for determining the location of an object.

[0016] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for determining the location of an object.

[0017] The method, apparatus, electronic device, and computer storage medium for determining the location of an object provided in this application embodiment acquire historical interaction operation information of multiple target users; the historical interaction operation information includes interaction operation information related to recommended sections and objects in the system; based on the acquired historical interaction operation information, a first preference score for each target user to each recommended section in the system and a second preference score for each object are determined; according to the first preference score and the second preference score, each recommended section and each object in the system are matched to obtain at least two matched objects; the matched object represents the recommended section that is successfully matched with the object; and the location information of each matched object among the at least two matched objects is determined.

[0018] As can be seen, in this embodiment, two different dimensions of preference scores are determined based on the historical interaction information of multiple target users: the preference score for the recommendation section dimension and the preference score for the object dimension. Compared with determining only the preference score for a single object dimension, determining two different dimensions of preference scores ensures that the subsequent recommendation results are more personalized. Then, based on the preference scores for the recommendation section dimension and the object dimension, each recommendation section and object in the system is matched, so that each target user's preferred recommendation section can be matched with their preferred object. In this way, in an e-commerce scenario, it is equivalent to matching the user's preferred products in the user's preferred recommendation section, which can improve the product's click-through rate and conversion rate. Subsequently, by determining the position of each successfully matched recommendation section, it is easier for the user to perform subsequent interaction operations, further improving the product's click-through rate and conversion rate. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a method for determining the location of an object provided in an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of the structure for determining the behavioral weight coefficients corresponding to each target user, provided in an embodiment of this application.

[0021] Figure 3 This is a schematic diagram illustrating the structure for determining the relationships between various recommendation modules in the system, provided in an embodiment of this application.

[0022] Figure 4 This is a schematic diagram illustrating the matching of recommendation sections and products in an e-commerce scenario, provided in an embodiment of this application.

[0023] Figure 5 This is a schematic diagram of the composition of the object location determination device provided in the embodiments of this application;

[0024] Figure 6 This is a schematic diagram of the composition structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In the following description, references to "some embodiments" refer to a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. The terminology used in the embodiments of this application is for the purpose of describing the embodiments of this application only and is not intended to limit the application.

[0027] In related technologies, the deployment of recommended products in e-commerce recommendation systems is mainly controlled based on system-level factors such as traffic and business focus of each recommendation section. In business scenarios targeting individual users (To Customer, ToC), this consideration method is relatively simplistic; this implicitly sets an upper limit for improving metrics such as click-through rate and conversion rate of the recommendation system.

[0028] The personalized experience extends beyond the user-product relationship; it also extends to the recommendation sections within a system, where click preferences are segmented. These preferences are difficult to accurately measure and define based solely on traffic and popularity. Just like shopping apps users use daily, each user's preferred sections differ, making this personalized behavioral preference a crucial dimension in recommendation systems. Currently, recommendation systems suggest products users are interested in. By incorporating personalized recommendation sections, users can fill their preferred sections with recommended products. This strong correlation can further improve metrics such as click-through rate and conversion rate. The following example illustrates this concept.

[0029] In some embodiments of this application, the method for determining the location of an object can be implemented using a processor in the object location determination device. The processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, and microprocessor.

[0030] It should be noted that the collection, use, storage, sharing and transfer of user personal information involved in the technical solution of the present invention all comply with the provisions of relevant laws and regulations, and require notification to users and obtaining their consent or authorization. When applicable, user personal information is subjected to de-identification and / or anonymization and / or encryption technical processing.

[0031] Figure 1 A flowchart illustrating a method for determining the location of an object provided in an embodiment of this application is shown below. Figure 1 As shown, the process may include:

[0032] Step S101: Obtain historical interaction information of multiple target users.

[0033] In this embodiment, the method for determining the location of an object can be applied to a recommendation system, which can be simply referred to as the system. Here, the application scenario of the recommendation system is not limited. For example, it can be an e-commerce scenario or other scenarios. The following description uses an e-commerce scenario as an example.

[0034] For example, the system may include multiple objects and multiple recommendation sections; wherein, the recommendation section is also called the recommendation column; in the e-commerce scenario, each object included in the system can be a product under any brand category, and each recommendation section included in the system can be used to display recommendation information for related products; wherein, the recommendation information may include relevant information such as the product name, price and image.

[0035] In this embodiment of the application, historical interaction operation information may include interaction operation information related to recommended sections and objects in the system; for example, it may include interaction operation information of each target user among multiple target users related to recommended sections and objects in the system within a historical time period.

[0036] Here, the length of the historical period can be set according to the actual situation. This application embodiment does not limit this. For example, it can be the last 15 days or the last 7 days, etc.

[0037] For example, for each target user, the historical interaction information may include the interaction information of the target user with several recommendation sections and several objects in the system during a historical period; wherein, several can be one or at least two.

[0038] For example, interactive operation information related to the recommended sections in the system includes, but is not limited to: click operation information, press operation information, etc.; interactive operation information related to objects in the system includes, but is not limited to: click operation information, browse operation information, favorite operation information, and order placement operation information, etc.

[0039] Step S102: Based on the obtained historical interaction operation information, determine the first preference score of each target user for each recommended section in the system, and the second preference score for each object.

[0040] In some embodiments, determining the first preference score of each target user for each recommended section in the system based on the acquired historical interaction information may include: determining the behavioral weight coefficient of each target user and each relevant recommended section, and the first preference score of each target user for each relevant recommended section, based on the acquired historical interaction information; determining the association between various recommended sections in the system based on the behavioral weight coefficient; and determining the first preference score of each target user for each irrelevant recommended section based on the association between various recommended sections; where irrelevant recommended sections are other recommended sections in the system besides relevant recommended sections.

[0041] Here, "related recommendation sections" refers to recommendation sections that have interactive operations with the target user; "irrelevant recommendation sections" refers to recommendation sections that do not have interactive operations with the target user; and recommendation sections can be either related or irrelevant.

[0042] For example, for each relevant recommendation section that interacts with each target user, by analyzing the data of the interaction information between each target user and each relevant recommendation section, the first preference score of each target user for each relevant recommendation section can be obtained.

[0043] For example, for each unrelated recommendation section that has no interaction with each target user, since the historical interaction information obtained does not contain information about the interaction between the two, it is impossible to determine the first preference score of each target user for each unrelated recommendation section through data analysis. To address this issue, the embodiments of this application are implemented in the following way: First, the behavioral weight coefficient between each target user and each related recommendation section is determined using a forward propagation calculation formula. Then, based on the behavioral weight coefficient corresponding to each target user and the number of recommendation sections related to that target user, the association relationship between each recommendation section is determined using a backward propagation calculation formula. Finally, based on the association relationship between each recommendation section, the first preference score of each target user for each unrelated recommendation section is determined. The following is an exemplary description.

[0044] For example, a preset weighting formula can be used to calculate the weight of the acquired historical interaction information to obtain the behavioral weight of each relevant recommendation section relative to each target user. Here, the behavioral weight of other irrelevant recommendation sections relative to each target user is 0. Then, a forward propagation method is used to combine the behavioral weight to determine the behavioral weight coefficient between each target user and each relevant recommendation section. Here, the weighting formula is mainly determined based on factors such as the order in which the target user clicks on each relevant recommendation section, the number of clicks, and whether there is a purchase behavior after clicking.

[0045] It should be noted that if, based on the historical interaction information, it is determined that a target user only interacts with one of the recommendation sections in the system (i.e., the target user corresponds to one relevant recommendation section), then there is no need to use the preset weight calculation formula to calculate the weight; the behavioral weight of the relevant recommendation section relative to the target user is directly determined to be 1, and the behavioral weight of other unrelated recommendation sections relative to the target user is 0. However, if, based on the historical interaction information, it is determined that a target user interacts with at least two recommendation sections in the system (i.e., the target user corresponds to at least two relevant recommendation sections), then it is necessary to use the preset weight calculation formula to calculate the behavioral weight of each relevant recommendation section relative to the target user. Similarly, the behavioral weight of other unrelated recommendation sections relative to the target user is 0.

[0046] Furthermore, after obtaining the behavioral weights of each relevant recommendation section relative to each target user, a forward propagation approach can be used to determine the behavioral weight coefficient f between each target user and each relevant recommendation section. t Here, the formula for calculating forward propagation is shown in formula (1):

[0047]

[0048] Where t represents the historical time period, α represents the smoothing factor, which is used to avoid cases where the denominator is 0, and can be set to 0.001, etc.; w t This represents the behavioral weight of each relevant recommendation section relative to each target user; n t This indicates the number of target users associated with each recommendation section.

[0049] Below, based on formula (1), combined with Figure 2 The process of determining the behavioral weight coefficients for each target user and each relevant recommendation section is illustrated with an example. For example... Figure 2As shown, the recommendation system includes three recommendation sections, namely recommendation section p_1, recommendation section p_2 and recommendation section p_3, and four target users, namely target user u_1, target user u_2, target user u_3 and target user u_4.

[0050] See Figure 2 It can be seen that target user u_1 interacts with recommendation section p_1, that is, recommendation section p_1 is a relevant recommendation section for target user u_1, and the behavioral weight of this relevant recommendation section relative to target user u_1 is 1; target user u_2 interacts with recommendation sections p_1, p_2 and p_3, that is, recommendation sections p_1, p_2 and p_3 are all relevant recommendation sections for target user u_2, and the behavioral weights of these three relevant recommendation sections relative to target user u_2 are 0.42, 0.32 and 0.26 respectively. Target user u_3 interacts with recommendation sections p_2 and p_3; that is, recommendation sections p_2 and p_3 are both relevant recommendation sections for target user u_3, and their behavioral weights relative to target user u_3 are 0.6 and 0.4, respectively. Target user u_4 interacts with recommendation sections p_1 and p_3; that is, recommendation sections p_1 and p_3 are both relevant recommendation sections for target user u_4, and their behavioral weights relative to target user u_4 are 0.55 and 0.45, respectively. Figure 2 It can be seen that the number of target users related to the recommendation section p_1 is 3, namely target user u_1, target user u_2 and target user u_4; the number of target users related to the recommendation section p_2 is 2, namely target user u_2 and target user u_3; and the number of target users related to the recommendation section p_3 is 3, namely target user u_2, target user u_3 and target user u_4.

[0051] Furthermore, after obtaining the above information, substituting it into formula (1), we can obtain: the behavioral weight coefficient between target user u_1 and related recommendation section p_1 is 0.33; the behavioral weight coefficients between target user u_2 and related recommendation sections p_1, p_2, and p_3 are 0.14, 0.16, and 0.09, respectively; the behavioral weight coefficients between target user u_3 and related recommendation sections p_2 and p_3 are 0.30 and 0.13, respectively; and the behavioral weight coefficients between target user u_4 and related recommendation sections p_1 and p_3 are 0.18 and 0.15, respectively. It can be seen that, according to the above steps, the behavioral weight coefficient between each target user and each related recommendation section can be obtained.

[0052] In this embodiment of the application, after obtaining the behavioral weight coefficients of each target user and each relevant recommendation section according to the above steps, the backpropagation method is further used to determine the association relationship between each recommendation section in the system. t Here, the formula for calculating backpropagation is shown in formula (2):

[0053]

[0054] Here, w t ' represents the behavioral weight of each target user relative to each relevant recommendation section, and its value can be 1; n t 'Indicates the number of recommendation sections relevant to each target user; it can be seen that the behavioral weight coefficient f t This is the result of forward propagation, and here the result of forward propagation is used as the input for the backpropagation calculation of the association relationship.

[0055] For example, in formula (2) and Figure 2 Based on, combined Figure 3 The process of determining the relationships between various recommendation sections in the system is illustrated with an example. For example... Figure 3 As shown, the number of recommendation sections related to target user u_1 is 1, namely recommendation section p_1; the number of recommendation sections related to target user u_2 is 3, namely recommendation section p_1, recommendation section p_2 and recommendation section p_3; the number of recommendation sections related to target user u_3 is 2, namely recommendation section p_2 and recommendation section p_3; and the number of recommendation sections related to target user u_4 is 2, namely recommendation section p_1 and recommendation section p_3.

[0056] Furthermore, after obtaining the behavioral weight coefficients of each target user (target user u_1, target user u_2, target user u_3 or target user u_4) and each related recommendation section, as well as the number of recommendation sections related to each target user, and substituting them into formula (2), the association relationships between recommendation sections p_1, p_2 and p_3 can be obtained, respectively: p_1=0.46*p_1+0.05*p_2+0.10*p_3; p_2=0.04*p_1+0.20*p_2+0.09*p_3; p_3=0.13*p_1+0.20*p_2+0.16*p_3; which can be represented by Table 1 as follows:

[0057] Relationship p_1 p_2 p_3 p_1 0.46 0.05 0.10 p_2 0.04 0.20 0.09 p_3 0.13 0.20 0.16

[0058] Table 1

[0059] Here, the association relationship represents the weight between the various recommendation sections. For example, in the table, p_1 = 0.46*p_1 + 0.05*p_2 + 0.10*p_3, representing the weight of recommendation section p_1 in relation to the other recommendation sections after backpropagation. It can be seen that, for recommendation section p_1, the weight of recommendation section p_3 is higher than that of recommendation section p_2; for recommendation section p_2, the weight of recommendation section p_3 is higher than that of recommendation section p_1; and for recommendation section p_3, the weight of recommendation section p_2 is higher than that of recommendation section p_1. Understandably, in the actual recommendation process of the recommendation system, this differentiation between recommendation sections will make the recommendation results more personalized.

[0060] In some embodiments, determining the first preference score of each target user for each unrelated recommendation section based on the correlation between the various recommendation sections may include: determining the first preference score of each target user for each unrelated recommendation section based on the correlation between the various recommendation sections and the first preference score of each target user for each related recommendation section.

[0061] For example, assuming, as shown in Table 2, target user u_1's first preference score for related recommendation section p_1 is 0.69; target user u_2's first preference scores for related recommendation sections p_1, p_2, and p_3 are 0.83, 0.21, and 0.39, respectively; target user u_3's first preference scores for related recommendation sections p_2 and p_3 are 0.53 and 0.91, respectively; and target user u_4's first preference scores for related recommendation sections p_1 and p_3 are 0.30 and 0.38, respectively. Based on the relationships between the recommendation sections and each target user's first preference score for each related recommendation section, we can calculate each target user's first preference score for each unrelated recommendation section that has no interactive operation. For example, the first preference score of target user u_3 for the irrelevant recommendation section p_1 is: (u_3,p_1)=0.05*0.53+0.10*0.91=0.11. Similarly, the first preference scores of other target users for each irrelevant recommendation section can be obtained, ultimately yielding the first preference score of each target user for each recommendation section.

[0062] First preference score P_1 P_2 P_3 U_1 0.69 0.00 0.00 U_2 0.83 0.21 0.39 U_3 0.00 0.53 0.91 U_4 0.30 0.00 0.38

[0063] Table 2

[0064] In some embodiments, after determining the first preference score of each target user for each recommended section in the system, the above method may further include: obtaining traffic information for each recommended section in the system; and correcting the first preference score corresponding to each recommended section based on the traffic information to obtain a corrected first preference score.

[0065] Here, the traffic information for each recommendation section in the system can be pre-allocated according to actual needs. Since the traffic information of each recommendation section reflects the popularity of each recommendation section to a certain extent, after obtaining the traffic information of each recommendation section, the first preference score corresponding to each recommendation section can be corrected based on the traffic information to improve the accuracy of the first preference score and ensure the effectiveness of subsequent recommendation results.

[0066] For example, as shown in Table 3, assume that the traffic information of recommendation sections p_1, p_2, and p_3 are 0.36, 0.43, and 0.21, respectively. Taking target user u_3 as an example, after correction based on the traffic information of each recommendation section, the corrected first preference score of target user u_3 for recommendation section p_1 is: (u_3,p_1)=0.11*0.36=0.03; the corrected first preference score of target user u_3 for recommendation section p_2 is: (u_3,p_2)=0.53*0.43=0.22; the corrected first preference score of target user u_3 for recommendation section p_3 is: (u_3,p_3)=0.91*0.21=0.19; and so on, the corrected first preference score of each target user for each recommendation section can be obtained.

[0067] P_1 P_2 P_3 0.36 0.43 0.21

[0068] Table 3

[0069] For example, the calculation of the second preference score of each target user for each object in the system can be carried out using some methods in the prior art, such as methods based on deep learning models, methods using preset rules, etc., and the embodiments of this application limit this.

[0070] Step S103: Based on the first preference score and the second preference score, match each recommended section and each object in the system to obtain at least two matched objects.

[0071] In some embodiments, matching each recommended section and each object in the system based on a first preference score and a second preference score may include: sorting the first preference scores corresponding to each recommended section to obtain a first sorting result; sorting the second preference scores corresponding to each object to obtain a second sorting result; and matching each recommended section and each object in the system based on the first sorting result and the second sorting result.

[0072] In one embodiment, after obtaining the first preference score for each target user on each recommended section, the first preference scores for each recommended section can be sorted in descending order to obtain a first sorting result. Similarly, after obtaining the second preference score for each target user on each object, the second preference scores for each object can be sorted in descending order to obtain a second sorting result. Based on the first and second sorting results, each recommended section and each object in the system can be matched sequentially according to the order of the recommended sections in the first sorting result and the order of the objects in the second sorting result to obtain each matched object. Here, a matched object represents a recommended section that successfully matches an object.

[0073] For example, for target user u_3, the first preference scores corresponding to recommendation sections p_1, p_2, and p_3 are sorted, and the first sorting result is: recommendation section p_3, recommendation section p_2, recommendation section p_1; assuming that the second preference scores of target user u_3 for objects 1, 2, 3, and 4 included in the system are 0.2, 0.5, 0.3, and 0.1 respectively, the second sorting result is: object 2, object 3, object 1, object 4.

[0074] Furthermore, matching each recommended section and each object in the system is equivalent to matching recommended sections with high preference scores with objects with high preference scores; that is, matching recommended section p_3 with object 2 results in recommended section p_3; matching recommended section p_2 with object 3 results in recommended section p_2; and matching recommended section p_1 with object 1 results in recommended section p_1. In other words, the matched objects include recommended sections p_3, p_2, and p_1. Among them, object 4 did not match any recommended section.

[0075] In some embodiments, matching each recommended section and each object in the system based on a first preference score and a second preference score may further include: matching each recommended section and each object in the system based on a modified first preference score and second preference score.

[0076] For example, after obtaining the corrected first preference score for each target user on each recommended section, each recommended section and each object in the system can be matched based on the corrected first preference score and second preference score.

[0077] For example, the modified first preference scores corresponding to each recommended section can be sorted in descending order to obtain a third sorting result; based on the order of each recommended section in the third sorting result and the order of each object in the second sorting result, each recommended section and each object in the system can be matched sequentially to obtain each matched object.

[0078] For example, for target user u_3, the corrected first preference scores of target user u_3 for recommended sections p_1, p_2 and p_3 are 0.03, 0.22 and 0.19 respectively; the third ranking result obtained after sorting is: recommended section p_2, recommended section p_3, recommended section p_1; here, the process of obtaining each corrected matching object is similar to the above steps, and will not be repeated here.

[0079] Step S104: Determine the position information of each of the at least two matching objects.

[0080] In some embodiments, determining the location information of each of the at least two matching objects may include: determining the target object that is successfully matched with each of the at least two matching objects; filling the corresponding matching object with the recommendation information of each target object; and determining the location information of each of the at least two matching objects after filling.

[0081] For example, the target object can be any of the multiple objects included in the system; as described above, the matching object represents the recommended section that is successfully matched with the object; after obtaining each matching object, the target object that is successfully matched with each matching object can be obtained accordingly; for example, for the target user u_3, before correction based on traffic information, the matching objects include: recommended section p_3, recommended section p_2, and recommended section p_1; in this case, the target objects include object 2 that is successfully matched with recommended section p_3, object 3 that is successfully matched with recommended section p_2, and object 1 that is successfully matched with recommended section p_1.

[0082] Furthermore, after obtaining each target object that successfully matches each matching object, the recommendation information of each target object is filled into the corresponding matching object, and the position information of each matching object in at least two matching objects after filling is determined; here, the recommendation information may include information describing the target object. For example, if the target object is a product, the recommendation information may include the product's name, price, and image, etc.

[0083] In some embodiments, each matched object after being filled has a corresponding display area; determining the position information of each matched object among the at least two filled matched objects may include: determining the click-through rate of each display area based on the obtained historical interaction operation information; arranging the display areas of the at least two filled matched objects in descending order of click-through rate to obtain an arrangement result; and determining the position information of each matched object among the at least two filled matched objects based on the arrangement result.

[0084] For example, the historical interaction information obtained above may also include interaction information related to the display area. By analyzing the interaction information related to each display area, the click-through rate of each display area can be obtained. Then, the display areas of each matched object after filling are arranged in descending order of click-through rate to obtain the arrangement result. According to the sorting result, the display area corresponding to each matched object after filling is adjusted, and the position information of the adjusted display area is determined as the position information of each matched object after filling.

[0085] For example, assuming the click-through rate (CTR) of display area 3 corresponding to the filled match 3 (corresponding to the above recommended section p_3) is 0.4, the CTR of display area 2 corresponding to the filled match 2 (corresponding to the above recommended section p_2) is 0.5, and the CTR of display area 1 corresponding to the filled match 1 (corresponding to the above recommended section p_1) is 0.2, then the sorted result is: display area 2, display area 3, display area 1. At this time, the display area 3 corresponding to the filled match 3 can be adjusted to display area 2, the display area 2 corresponding to the filled match 2 can be adjusted to display area 3, and the display area 1 corresponding to the filled recommended section p_1 remains unchanged.

[0086] As can be seen, this application embodiment sorts the display areas of each matched object after filling them according to the click rate of each display area, and adjusts the display areas of each matched object after filling them according to the sorting results, so that users can click on the object they want to click on in the area they like to click, which is more in line with user habits and improves the recommendation effect.

[0087] The method for determining the location of an object provided in this application involves acquiring historical interaction information of multiple target users; the historical interaction information includes interaction information related to recommended sections and objects in the system; based on the acquired historical interaction information, determining a first preference score for each target user on each recommended section in the system, and a second preference score for each object; matching each recommended section and each object in the system according to the first preference score and the second preference score, obtaining at least two matched objects; the matched object represents a recommended section that successfully matches an object; and determining the location information of each of the at least two matched objects. As can be seen, in this embodiment, two different dimensions of preference scores are determined based on the historical interaction information of multiple target users: the preference score for the recommendation section dimension and the preference score for the object dimension. Compared with determining only the preference score for a single object dimension, determining two different dimensions of preference scores ensures that the subsequent recommendation results are more personalized. Then, based on the preference scores for the recommendation section dimension and the object dimension, each recommendation section and object in the system is matched, so that each target user's preferred recommendation section can be matched with their preferred object. In this way, in an e-commerce scenario, it is equivalent to matching the user's preferred products in the user's preferred recommendation section, which can improve the product's click-through rate and conversion rate. Subsequently, by determining the position of each successfully matched recommendation section, it is easier for the user to perform subsequent interaction operations, further improving the product's click-through rate and conversion rate.

[0088] To better illustrate the purpose of this application, further explanation will be provided based on the above embodiments and in the context of e-commerce scenarios.

[0089] See Figure 4In an e-commerce scenario, suppose a target user's first preference scores for recommendation sections 1 through 4 in the system are 0.4, 0.5, 0.6, and 0.3, respectively; and their second preference scores for products 1 through 4 in the system are 0.8, 0.7, 0.6, and 0.5, respectively. The list above represents the product filling status in each recommendation section before matching; the list below represents the product filling status in each recommendation section after matching. Accordingly, recommendation section 1 is filled with recommendation information for product 3, recommendation section 2 is filled with recommendation information for product 2, and so on. Recommendation section 3 (highest first preference score) is populated with recommendation information for product 1 (highest second preference score), and recommendation section 4 (lowest first preference score) is populated with recommendation information for product 4 (lowest second preference score). It can be seen that after obtaining the target user's first preference score for each recommendation section and second preference score for each item, matching each recommendation section with its corresponding product based on the sorted first and second preference scores allows for the sequential combination of recommendation sections and products that the target user is interested in. This can improve metrics such as click-through rate and conversion rate.

[0090] Figure 5 This is a schematic diagram of the composition of the object location determination device provided in the embodiments of this application, as shown below. Figure 5 As shown, the object location determination device 400 includes: an acquisition module 401, a first determination module 402, a matching module 403, and a second determination module 404, wherein:

[0091] The acquisition module 401 is used to acquire historical interaction operation information of multiple target users; the historical interaction operation information includes interaction operation information related to the recommended sections and objects in the system.

[0092] The first determining module 402 is used to determine, based on the acquired historical interaction operation information, the first preference score of each target user for each recommended section in the system, and the second preference score for each object.

[0093] Matching module 403 is used to match each recommendation section and each object in the system according to the first preference score and the second preference score to obtain at least two matching objects; the matching object represents the recommendation section that is successfully matched with the object;

[0094] The second determining module 404 is used to determine the position information of each of the at least two matching objects.

[0095] In some embodiments, the first determining module 402 is further configured to:

[0096] Based on the acquired historical interaction information, the behavioral weight coefficient of each target user and each related recommendation section is determined, as well as the first preference score of each target user for each related recommendation section.

[0097] Based on the behavioral weight coefficients, the relationships between the various recommendation sections in the system are determined;

[0098] Based on the relationships between the various recommendation sections, a first preference score for each target user on each irrelevant recommendation section is determined; the irrelevant recommendation section refers to other recommendation sections in the system besides the relevant recommendation sections, and the recommendation section is either the relevant recommendation section or the irrelevant recommendation section.

[0099] In some embodiments, the matching module 403 is further configured to:

[0100] The first preference scores corresponding to each recommended section are sorted to obtain the first sorting result;

[0101] The second preference scores corresponding to each object are sorted to obtain a second sorting result;

[0102] Based on the first ranking result and the second ranking result, each recommended section and each object in the system are matched.

[0103] In some embodiments, the second determining module 404 is further configured to:

[0104] Identify the target object that successfully matches each of the at least two matching objects;

[0105] Populate the recommendation information for each target object into the corresponding matching object;

[0106] Determine the position information of each of the at least two matched objects after filling.

[0107] In some embodiments, each matched object after being filled has a corresponding display area; the historical interaction operation information also includes interaction operation information related to the display area, and the second determining module 404 is further configured to:

[0108] Based on the acquired historical interaction information, the click-through rate of each display area is determined;

[0109] Arrange the display areas of at least two matched objects after filling them in descending order of click-through rate to obtain the arrangement result;

[0110] Based on the arrangement result, determine the position information of each matching object in at least two matched objects after filling.

[0111] In some embodiments, after determining the first preference score of each target user for each recommended section in the system, the first determining module 402 is further configured to:

[0112] Obtain traffic information for each recommendation section in the system;

[0113] Based on the traffic information, the first preference score corresponding to each recommendation section is corrected to obtain the corrected first preference score.

[0114] In some embodiments, the matching module 403 is further configured to:

[0115] Based on the corrected first preference score and second preference score, each recommendation section in the system is matched with each object.

[0116] In practical applications, the above-mentioned acquisition module 401, first determination module 402, matching module 403 and second determination module 404 can all be implemented by a processor located in an electronic device. The processor can be at least one of ASIC, DSP, DSPD, PLD, FPGA, CPU, controller, microcontroller and microprocessor.

[0117] Furthermore, in this embodiment, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0118] If the integrated unit is implemented as a software functional module and is not sold or used as an independent object, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software object. This computer software object is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] Specifically, the computer program instructions corresponding to the object location determination method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the computer program instructions corresponding to the object location determination method in the storage media are read or executed by an electronic device, any of the object location determination methods in the aforementioned embodiments can be implemented.

[0120] It should be noted that the description of the apparatus in this application embodiment is similar to the description of the method embodiment described above, and has similar beneficial effects as the method embodiment; therefore, it will not be repeated. For technical details not disclosed in this apparatus embodiment, please refer to the description of the method embodiment of this application for understanding.

[0121] Correspondingly, this application provides an electronic device. Figure 6 This is a schematic diagram of the composition structure of the electronic device provided in the embodiments of this application, such as... Figure 6 As shown, the electronic device 500 includes at least a processor 501 and a computer-readable storage medium 502 configured to store executable instructions, wherein the processor 501 generally controls the overall operation of the electronic device. The computer-readable storage medium 502 is configured to store instructions and applications executable by the processor 501, and may also cache data to be processed or processed by various modules in the processor 501 and the electronic device 500, which may be implemented by flash memory or RAM.

[0122] This application provides a storage medium storing executable instructions. When these executable instructions are executed by a processor, they cause the processor to perform the method provided in this application, for example... Figure 1 The method shown.

[0123] In some embodiments, the storage medium may be a computer-readable storage medium, such as ferromagnetic random access memory (FRAM), ROM, programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic surface memory, optical disk, or compact disk-read-only memory (CD-ROM); or it may be a device that includes one or any combination of the above-mentioned memories.

[0124] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0125] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts within a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files storing one or more modules, subroutines, or code sections). As an example, executable instructions may be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0126] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

[0127] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0128] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not performed.

[0129] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for determining the location of an object, applied to a recommendation system, the method comprising: Obtain historical interaction information from multiple target users; The historical interaction information includes interaction information related to the recommended sections and objects in the system. Based on the acquired historical interaction information, determine the first preference score of each target user for each recommended section in the system, and the second preference score for each object; Based on the first preference score and the second preference score, each recommendation section and each object in the system are dynamically matched. The recommendation section with the higher first preference score is matched with the object with the higher second preference score to obtain at least two matching objects. The matching object represents the recommendation section that is successfully matched with the object. Each recommended section corresponds to one object; Determine the position information of each of the at least two matching objects.

2. The method according to claim 1, wherein determining the first preference score of each target user for each recommended section in the system based on the acquired historical interaction information includes: Based on the acquired historical interaction information, the behavioral weight coefficient of each target user and each related recommendation section is determined, as well as the first preference score of each target user for each related recommendation section. Based on the behavioral weight coefficients, the relationships between the various recommendation sections in the system are determined; Based on the relationships between the various recommendation sections, a first preference score for each target user on each irrelevant recommendation section is determined; the irrelevant recommendation section refers to other recommendation sections in the system besides the relevant recommendation sections, and the recommendation section is either the relevant recommendation section or the irrelevant recommendation section.

3. The method according to claim 1 or 2, wherein matching each recommendation section and each object in the system based on the first preference score and the second preference score includes: The first preference scores corresponding to each recommended section are sorted to obtain the first sorting result; The second preference scores corresponding to each object are sorted to obtain a second sorting result; Based on the first ranking result and the second ranking result, each recommended section and each object in the system are matched.

4. The method according to claim 1 or 2, wherein determining the position information of each of the at least two matching objects comprises: Identify the target object that successfully matches each of the at least two matching objects; Populate the recommendation information for each target object into the corresponding matching object; Determine the position information of each of the at least two matched objects after filling.

5. The method according to claim 4, wherein each matched object after filling has a corresponding display area; the historical interaction operation information further includes interaction operation information related to the display area, and determining the position information of each matched object among at least two matched objects after filling includes: Based on the acquired historical interaction information, the click-through rate of each display area is determined; Arrange the display areas of at least two matched objects after filling them in descending order of click-through rate to obtain the arrangement result; Based on the arrangement result, determine the position information of each matching object in at least two matched objects after filling.

6. The method according to claim 1 or 2, after determining the first preference score of each target user for each recommended section in the system, the method further includes: Obtain traffic information for each recommendation section in the system; Based on the traffic information, the first preference score corresponding to each recommendation section is corrected to obtain the corrected first preference score.

7. The method according to claim 6, wherein matching each recommendation section and each object in the system based on the first preference score and the second preference score comprises: Based on the corrected first preference score and second preference score, each recommendation section in the system is matched with each object.

8. An object location determination device, applied in a recommendation system, the device comprising: The acquisition module is used to acquire historical interaction information of multiple target users; The historical interaction information includes interaction information related to the recommended sections and objects in the system. The first determining module is used to determine, based on the acquired historical interaction operation information, the first preference score of each target user for each recommendation section in the system, and the second preference score for each object. The matching module is used to dynamically match each recommendation section and each object in the system based on the first preference score and the second preference score, and match the recommendation section with the high first preference score with the object with the high second preference score to obtain at least two matching objects; The matching object representation is associated with a recommendation section where the object is successfully matched. Each recommended section corresponds to one object; The second determining module is used to determine the position information of each of the at least two matching objects.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method of any one of claims 1 to 7.

10. A computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the method of any one of claims 1 to 7.

Images