Store recommendation methods, servers, display methods, clients and systems

Abstract

This disclosure provides a store recommendation method, server, display method, client, and system. The method includes: responding to a store recommendation request for a target user issued by a client; obtaining a target store to be displayed and multiple combinations of recommendation reason texts for the target store; wherein each combination of recommendation reason texts includes at least two recommendation reason texts in an ordered manner; obtaining a target user feature vector corresponding to the target user, a target store feature vector corresponding to the target store, and a combination feature vector corresponding to each combination of recommendation reason texts; determining a combination of recommendation reason texts that matches the target user and the target store based on the target user feature vector, the target store feature vector, and the combination feature vector, as the target text combination; and pushing the target text combination and the store information of the target store to the client for display by the client.

Description

Technical Field

[0001] This disclosure relates to the field of recommendation technology, and more specifically, to a store recommendation method based on recommendation reasons, a server, a method for displaying store information, a client, and a store recommendation system based on recommendation reasons. Background Technology

[0002] With the development of the information society, people are increasingly accustomed to using the internet to find or select stores. Generally, some reasons for recommendation are displayed to users along with the store information. However, with existing technology, it is possible to pre-edit some text of the reasons for recommendation for use when displaying the information.

[0003] In existing technologies, a backend server can push multiple recommendation reasons for a single store to a client, and the client displays the received recommendation reasons in a preset sorting order. These recommendation reasons are primarily used to highlight the merchant's unique features and assist users in making purchasing decisions.

[0004] However, the total length of these recommendation reasons may exceed the width of the client's interface. At least part of the recommendation reasons that exceed the interface will not be displayed by the client, resulting in blank space in the display area of ​​the recommendation reasons.

[0005] In such Figures 1a to 1c In the example shown, the backend server pushes four recommendation reasons to the client for a single store: Reason 1, Reason 2, Reason 3, and Reason 4. In such a case... Figure 1a In the example shown, the four recommendation reasons are sorted according to sorting scheme 1. Sorting scheme 1 can be recommendation 1, recommendation 2, recommendation 3, and recommendation 4. Therefore, the store card on the client side can only display recommendation 1 and recommendation 2. Figure 1b In the example shown, the four recommendation reasons are sorted according to sorting scheme 2. Sorting scheme 2 can be sentence 1, sentence 2, sentence 4, and sentence 3. Therefore, the store card on the client side can only display sentence 1, sentence 2, and sentence 4. Figure 1c In the example shown, the four recommendation reasons are sorted according to sorting scheme 3. Sorting scheme 3 can be sentence 2, sentence 3, sentence 4 and sentence 1. Therefore, the store card on the client can only display sentence 2, sentence 3 and sentence 4. Summary of the Invention

[0006] One objective of this disclosure is to provide a new technical solution that can improve the presentation of recommendation reason copy.

[0007] According to a first aspect of this disclosure, a method for recommending stores based on store recommendation reasons is provided, the method being used on a server, the method comprising:

[0008] In response to a store recommendation request sent by the client for a target user, the system obtains the target store to be displayed and multiple combinations of recommendation reason texts for the target store; wherein each combination of recommendation reason texts includes at least two recommendation reason texts in an ordered order.

[0009] Obtain the target user feature vector corresponding to the target user, the target store feature vector corresponding to the target store, and the combination feature vector corresponding to each combination of recommendation reason copy.

[0010] Based on the target user feature vector, the target store feature vector, and the combined feature vector, a combination of recommendation reason copywriting that matches the target user and the target store is determined as the target copywriting combination;

[0011] The target copy combination and the target store's store information are pushed to the client for display.

[0012] Optionally, obtain the combined feature vector corresponding to the combination of recommendation reason copy, including:

[0013] Obtain the text feature vector of each recommendation reason text in the combination of recommendation reason texts;

[0014] The text feature vectors of the recommendation reason texts in the combination of recommendation reason texts are concatenated to obtain the combination feature vector corresponding to the combination of recommendation reason texts.

[0015] Optionally, the combination of the multiple recommendation reason texts can be obtained, including:

[0016] Obtain multiple recommendation reasons for the target store;

[0017] Based on the interface size of the client and the length of each recommendation reason text, multiple combinations of recommendation reason texts are obtained.

[0018] Optionally, the step of obtaining multiple combinations of recommendation reason texts based on the client's interface size and the length of each recommendation reason text includes:

[0019] The multiple recommendation reasons are arranged and combined to obtain multiple permutation and combination results;

[0020] Based on the interface size of the client and the length of each recommendation reason text, at least two recommendation reason texts to be displayed on the client's interface in each permutation and combination result are determined as the recommendation reason text combination.

[0021] Optionally, the dimensions of the target user feature vector and the target store feature vector are both the target dimension;

[0022] The step of determining a combination of recommendation reason texts matching the target user and the target store, based on the target user feature vector, the target store feature vector, and the combined feature vector, as the target text combination, includes:

[0023] The combined feature vector is subjected to target dimension transformation to obtain the combined feature vector of the target dimension;

[0024] Perform a dot product operation on the target user feature vector, the target store feature vector, and the combined feature vector of the target dimension to obtain the predicted matching score of the target user, the target store, and each combination of the recommendation reason copy.

[0025] The combination of recommendation reason copy with the highest predicted matching score with the target user and the target store is selected as the target copy combination.

[0026] Optionally, obtaining the user feature vector corresponding to the target user and the store feature vector corresponding to the target store includes:

[0027] Obtain the user identifier of the target user and the store identifier of the target store;

[0028] Based on the user identifier, obtain the user feature vector corresponding to the target user from the user feature vectors of multiple pre-stored users, and use it as the target user feature vector;

[0029] Based on the store identifier, the store feature vector corresponding to the target store is obtained from the store feature vectors of multiple pre-stored stores, and used as the target store feature vector.

[0030] Optionally, before obtaining the user feature vector corresponding to the target user and the store feature vector corresponding to the target store, the method further includes:

[0031] Obtain a first feature vector of the target user, wherein the first feature vector includes multiple user features reflecting the target user's preferences for stores and recommendation reason copy;

[0032] The first feature vector is subjected to a first dimension transformation to obtain a second feature vector of the target dimension; the first feature vector is subjected to a second dimension transformation to obtain a third feature vector of the target dimension.

[0033] The target user feature vector is obtained by summing the second feature vector and the third feature vector.

[0034] Obtain the fourth feature vector of the target store, wherein the fourth feature vector includes multiple store features that influence the target user's preference for the target store;

[0035] The fourth feature vector is transformed in the third dimension to obtain the fifth feature vector of the target dimension; the fourth feature vector is transformed in the fourth dimension to obtain the sixth feature vector of the target dimension.

[0036] The target store feature vector is obtained by summing the fifth feature vector and the sixth feature vector.

[0037] Store the target user feature vector and the target store feature vector.

[0038] According to a second aspect of this disclosure, a method for displaying store information is provided, the method being used on a client side, the method comprising:

[0039] In response to a target user's target operation performed through the client, a store recommendation request for the target user is sent to the server. The server then responds to the store recommendation request by determining a combination of recommendation reason texts that matches the target user and the target store, which is then used as the target text combination. The server then pushes the target text combination and the store information of the target store to the client. Each combination of recommendation reason texts includes at least two recommendation reason texts in an ordered manner.

[0040] Receive the target copy combination and the store information of the target store;

[0041] The store card corresponding to the target store displays the store information of the target store and the target copywriting combination.

[0042] Optionally, displaying the target copy combination includes:

[0043] The recommendation reasons included in the target copy combination will be displayed in the recommendation reason display box of the store card.

[0044] According to a third aspect of this disclosure, a server is provided, including a first memory and a first processor, the first memory being configured to store an executable first computer program; the first computer program being configured to control the first processor to perform the method as described in the first aspect of this disclosure.

[0045] According to a fourth aspect of this disclosure, a client is provided, including a second processor and a second memory, the second memory being used to store an executable second computer program; the second computer program being used to control the second processor to perform the method as described in the second aspect of this disclosure.

[0046] According to a fifth aspect of this disclosure, a store recommendation system based on recommendation reasons is provided, including a server as described in a third aspect of this disclosure and a client as described in a fourth aspect of this disclosure.

[0047] Through the embodiments of this disclosure, multiple recommendation reasons can be maximized to occupy the corresponding display space on the client's interface, thereby improving the display effect of the recommendation reasons. Furthermore, when a target store has multiple recommendation reasons, the recommendation reasons can be personalized for each target user, resulting in personalized recommendations of combinations of recommendation reasons.

[0048] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0049] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments of the present disclosure and, together with their description, serve to explain the principles of the present disclosure.

[0050] Figure 1a This is a schematic diagram illustrating the display effect of recommendation text sorted according to sorting scheme 1 in the client in the existing technology;

[0051] Figure 1b This is a schematic diagram illustrating the display effect of recommendation text sorted according to sorting scheme 2 in the client.

[0052] Figure 1c This is a schematic diagram illustrating the display effect of recommendation text sorted according to sorting scheme 3 in the client.

[0053] Figure 2 This is a block diagram showing an example of a hardware configuration for a vehicle system that can be used to implement embodiments of the present disclosure;

[0054] Figure 3 A flowchart illustrating a store recommendation method based on recommendation reasons according to an embodiment of this disclosure is shown;

[0055] Figure 4 A schematic diagram illustrating an example of a store recommendation method based on recommendation reasons according to an embodiment of the present disclosure is shown;

[0056] Figure 5 A schematic block diagram of a server according to an embodiment of the present disclosure is shown;

[0057] Figure 6 A flowchart illustrating a method for displaying store information according to an embodiment of this disclosure is provided.

[0058] Figure 7 A schematic block diagram of a client according to an embodiment of the present disclosure is shown;

[0059] Figure 8 A schematic block diagram of a store recommendation system based on recommendation reasons, according to an embodiment of the present disclosure, is shown. Detailed Implementation

[0060] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0061] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0062] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0063] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0064] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0065] <Hardware Configuration>

[0066] like Figure 2 As shown, the recommendation system 100 includes a feature server 1000, a client 2000, a recommendation server 3000, and a network 4000.

[0067] Feature server 1000 provides service points for processing, databases, and communication facilities. Feature server 1000 can be a monolithic server or a distributed server spanning multiple computers or a computer data center. Feature servers can be of various types, such as, but not limited to, web servers, news servers, mail servers, messaging servers, advertising servers, file servers, application servers, interactive servers, database servers, or proxy servers. In some embodiments, each server may include hardware, software, or embedded logical components or combinations of two or more such components for performing suitable functions supported or implemented by the server. For example, a feature server may be a blade server, a cloud server, or a server cluster consisting of multiple servers, and may include one or more of the types of servers described above. Figure 2 As shown, the feature server 1000 can be server group 1000-1.

[0068] In one example, feature server 1000 can be like... Figure 2 As shown, it includes a processor 1100, a memory 1200, an interface device 1300, a communication device 1400, a display device 1500, and an input device 1600. Although the server may also include speakers, microphones, etc., these components are not relevant to the present disclosure and are therefore omitted here.

[0069] The processor 1100 may be, for example, a central processing unit (CPU) or a microprocessor (MCU). The memory 1200 may include, for example, ROM (Read-Only Memory), RAM (Random Access Memory), or non-volatile memory such as a hard disk. The interface device 1300 may include, for example, a USB interface, a serial interface, or an infrared interface. The communication device 1400 may be capable of wired or wireless communication. The display device 1150 may be, for example, a liquid crystal display (LCD), an LED display, or a touchscreen display. The input device 1160 may include, for example, a touchscreen or a keyboard.

[0070] In this embodiment, the client 2000 is an electronic device with communication and business processing functions. The client 2000 can be a mobile terminal, such as a mobile phone, laptop, tablet computer, PDA, etc. In one example, the client 2000 is a device that performs management operations on the vehicle 3000, for example, a mobile phone with a specified application (APP) installed.

[0071] like Figure 2 As shown, the client 2000 may include a processor 2100, a memory 2200, an interface device 2300, a communication device 2400, a display device 2500, an input device 2600, a speaker 2700, a microphone 2800, etc. The processor 2100 may be a central processing unit (CPU), a microprocessor (MCU), etc. The memory 2200 may include, for example, ROM (Read-Only Memory), RAM (Random Access Memory), or non-volatile memory such as a hard disk. The interface device 2300 may include, for example, a USB interface, a headphone jack, etc. The communication device 2400 may be capable of wired or wireless communication. The display device 2500 may be, for example, an LCD screen, a touch screen, etc. The input device 2600 may include, for example, a touch screen, a keyboard, etc. The user can input / output voice information through the speaker 2700 and the microphone 2800.

[0072] Recommendation server 3000 provides service points for processing, databases, and communication facilities. Recommendation server 3000 can be a monolithic server or a distributed server spanning multiple computers or computer data centers. Recommendation servers can be of various types, such as, but not limited to, web servers, news servers, mail servers, messaging servers, advertising servers, file servers, application servers, interactive servers, database servers, or proxy servers. In some embodiments, each server may include hardware, software, or embedded logical components or combinations of two or more such components for performing suitable functions supported or implemented by the server. For example, recommendation servers may be blade servers, cloud servers, etc., or may be a server cluster consisting of multiple servers, and may include one or more of the types of servers described above, etc. Figure 2 As shown, the recommended server 3000 can be server group 3000-1.

[0073] In one example, a recommended server 3000 could be as follows: Figure 2 As shown, it includes a processor 3100, a memory 3200, an interface device 3300, a communication device 3400, a display device 3500, and an input device 3600. Although the server may also include speakers, microphones, etc., these components are not relevant to the present disclosure and are therefore omitted here.

[0074] The processor 3100 may be, for example, a central processing unit (CPU) or a microprocessor (MCU). The memory 3200 may include, for example, ROM (Read-Only Memory), RAM (Random Access Memory), or non-volatile memory such as a hard disk. The interface device 3300 may include, for example, a USB interface, a serial interface, or an infrared interface. The communication device 3400 may be capable of wired or wireless communication. The display device 3500 may be, for example, a liquid crystal display (LCD), an LED display, or a touchscreen display. The input device 3600 may include, for example, a touchscreen or a keyboard.

[0075] Network 4000 can be a wireless communication network or a wired communication network, and can be a local area network (LAN) or a wide area network (WAN). Figure 2 In the recommendation system shown, recommendation server 3000 and feature server 1000, and client 2000 and recommendation server 3000, can communicate via network 4000. Furthermore, the network 4000 on which recommendation server 3000 and feature server 1000, and client 2000 and recommendation server 3000 communicate, can be the same or different.

[0076] It should be understood that, despite Figure 2Only one feature server 1000, client 2000, and recommendation server 3000 are shown, but this does not mean that the corresponding number is limited. The recommendation system 100 may contain multiple feature servers 1000, multiple clients 2000, and multiple recommendation servers 3000.

[0077] Figure 2 The recommendation system 100 shown is merely illustrative and is by no means intended to limit this disclosure, its application, or its use.

[0078] In embodiments of this disclosure, the memory 1200 of the feature server 1000 is used to store instructions for controlling the processor 1100 to operate in order to execute the method implemented by the feature server provided in embodiments of this disclosure.

[0079] Despite Figure 2 The feature server 1000 is shown with multiple devices, but this disclosure may only relate to some of these devices. For example, the feature server 1000 may only relate to the memory 1200 and the processor 1100.

[0080] In embodiments applied to this disclosure, the memory 2200 of the client 2000 is used to store instructions for controlling the processor 2100 to run the client 2000 to execute the method implemented by the client according to the embodiments of this disclosure.

[0081] Despite Figure 2 The client 2000 shows multiple devices, however, this disclosure may refer only to some of them, for example, the client 2000 may refer only to memory 2200 and processor 2100.

[0082] In embodiments of this disclosure, the memory 3200 of the recommendation server 3000 is used to store instructions for controlling the processor 3100 to operate in order to execute the method implemented by the recommendation server provided in embodiments of this disclosure.

[0083] Despite Figure 2 The present disclosure illustrates multiple devices for the recommendation server 3000; however, this disclosure may refer only to some of these devices, for example, the recommendation server 3000 may refer only to the memory 3200 and the processor 3100.

[0084] In the above description, those skilled in the art can design instructions based on the scheme disclosed in this disclosure. How the instructions control the processor to operate is well known in the art, and therefore will not be described in detail here.

[0085] <First Embodiment>

[0086] <Method>

[0087] This embodiment provides a method for recommending stores based on store recommendation reasons, which can be implemented by a server. In one example, the server may include, for example... Figure 2 The recommended server shown is 3000. In another example, the server may also include, for example, Figure 2 The feature server 1000 and the recommendation server 3000 are shown.

[0088] like Figure 3 As shown, the store recommendation method based on store recommendation reasons includes steps S3100 to S3400.

[0089] Step S3100: In response to the store recommendation request sent by the client for the target user, obtain the target store to be displayed and multiple combinations of recommendation reason texts for the target store; wherein, each combination of recommendation reason texts includes at least two recommendation reason texts with an ordered arrangement.

[0090] In one embodiment, the client may send a store recommendation request to the server in response to a target user's target action performed by the client.

[0091] Among them, the target user can be a user who logs into the target application through this client.

[0092] In this embodiment, the target operation can be either the target user opening the target application or the target user opening the target page of the target application.

[0093] In one embodiment of this disclosure, the feature server may pre-store the feature vectors of multiple stores.

[0094] Furthermore, the store recommendation request for the target user can include the target user's user identifier.

[0095] Therefore, in response to a store recommendation request, obtaining the target store to be displayed can include: sending the target user's user identifier to the feature server and receiving the target store returned by the feature server. Specifically, the feature server can obtain the target user feature vector corresponding to the target user based on the user identifier sent by the recommendation server, and based on the target user feature vector and the store feature vectors of multiple pre-stored stores, obtain the target store to be displayed from multiple stores, and return the target store to the recommendation server.

[0096] Alternatively, the recommendation server can pre-store the feature vectors of multiple stores. In response to a store recommendation request, the recommendation server obtains the target user feature vector corresponding to the target user, and selects the target store to be displayed from the multiple stores based on the target user feature vector and the pre-stored store feature vectors of multiple stores.

[0097] In this embodiment, a preset store recommendation model can be used to obtain the store preference score of the target user for each store pair based on the target user feature vector and the store feature vectors of multiple stores stored in advance; stores whose descending order of store preference scores conforms to a preset sorting range are selected as target stores.

[0098] The target user's store preference score reflects the degree of preference the target user has for the store. Stores whose store preference scores are sorted in descending order within a preset sorting range are selected as target stores to be displayed. This ensures that the target stores are those with a high degree of preference from the target users, so as to more accurately and effectively meet the user's store preferences and actually satisfy the user's need to find stores.

[0099] The preset sorting range can be set according to specific application scenarios or application requirements. For example, the preset sorting range can be set to 1-10.

[0100] In another embodiment of this disclosure, in response to a store recommendation request, obtaining the target store to be displayed may include at least one of the following:

[0101] Select the first number of stores that are closest to the target user from among multiple stores, and use them as the target stores;

[0102] Select the second number of stores with the highest revenue in the first historical period from multiple stores as target stores;

[0103] Select the third number of stores with the highest click-through rate in the second historical period from multiple stores as target stores.

[0104] The first, second, and third quantities can be pre-defined based on the application scenario or specific requirements. These quantities can be equal or unequal; no specific restrictions apply. For example, the first, second, and third quantities can all be 10.

[0105] In this embodiment, the first and second historical time periods can be pre-defined based on the application scenario or specific needs. The first and second historical time periods can be the same or different. For example, the first historical time period can be the past week, and the second historical time period can be the past 3 days.

[0106] In this embodiment, multiple combinations of recommendation reason texts for each store may be pre-stored. When a target store is obtained, multiple combinations of recommendation reason texts for that target store can be obtained. Each combination of recommendation reason texts includes at least two recommendation reason texts arranged in a specific order.

[0107] In one embodiment of this disclosure, obtaining multiple combinations of recommendation reason texts for a target store may include: obtaining multiple recommendation reason texts for the target store; and obtaining multiple combinations of recommendation reason texts based on the client's interface size and the length of each recommendation reason text.

[0108] In this embodiment, all the recommendation reasons in each combination of recommendation reasons for the target store can be displayed on the client.

[0109] In one example, the length of the recommendation text can be determined based on the word count. Specifically, this can be achieved by pre-setting mapping data that reflects the relationship between the word count and the length of the recommendation text; the length of the recommendation text can be obtained based on this mapping data and the word count. This mapping data can be a lookup table or a mapping function.

[0110] In one embodiment, obtaining multiple combinations of recommendation reason texts based on the client's interface size and the length of each recommendation reason text may include: arranging and combining multiple recommendation reason texts to obtain multiple permutation and combination results; and determining at least two recommendation reason texts to be displayed on the client's interface from each permutation and combination result, based on the client's interface size and the length of each recommendation reason text, as a combination of recommendation reason texts.

[0111] If there are n recommendation reasons for the target store, then n! permutations and combinations can be obtained.

[0112] In this embodiment, the width of the display area for the recommendation text on the client's interface can be obtained based on the client's interface size. Specifically, the width of the display area for the recommendation text can be pre-stored as a percentage of the client's interface width. Based on the client's interface size and this percentage, the width of the display area for the recommendation text on the client's interface can be obtained.

[0113] In this embodiment, each recommendation reason in the combination of recommendation reasons can be fully displayed on the client's interface. That is, for each permutation, the first m recommendation reasons can be fully displayed on the client's interface, while the mth to nth recommendation reasons cannot. Therefore, the sum of the lengths of the first m recommendation reasons is less than or equal to the width of the display area for recommendation reasons on the client's interface, and the sum of the lengths of the first m+1 recommendation reasons is greater than the width of the display area for recommendation reasons on the client's interface.

[0114] For example, if the recommendation text for the target store includes text 1, text 2, text 3, and text 4, the permutation and combination results could include: text 1, text 2, text 3, and text 4; text 1, text 3, text 2, and text 4; text 1, text 2, text 4, and text 3; text 1, text 4, text 3, and text 2; text 1, text 4, text 2, and text 3; text 2, text 1, text 3, and text 4; text 2, text 1, text 4, and text 3; text 2, text 3, text 1, and text 4; text 2, text 3, text 4, and text 1; text 2, text 4, text 1, and text 3; Copywriting 2, Copywriting 3, Copywriting 3 and Copywriting 1; Copywriting 3, Copywriting 1, Copywriting 2 and Copywriting 4; Copywriting 3, Copywriting 1, Copywriting 4 and Copywriting 2; Copywriting 3, Copywriting 2, Copywriting 4 and Copywriting 1; Copywriting 3, Copywriting 2, Copywriting 1 and Copywriting 4; Copywriting 3, Copywriting 4, Copywriting 2 and Copywriting 1; Copywriting 3, Copywriting 4, Copywriting 1 and Copywriting 2; Copywriting 4, Copywriting 2, Copywriting 3 and Copywriting 1; Copywriting 4, Copywriting 2, Copywriting 1 and Copywriting 3; Copywriting 4, Copywriting 1, Copywriting 2 and Copywriting 3; Copywriting 4, Copywriting 1, Copywriting 3 and Copywriting 2.

[0115] The permutation and combination result is as follows: Figure 1a Given the examples 1, 2, 3, and 4 shown, the resulting combination of recommendation reasoning texts can include texts 1 and 2. The permutation and combination results are as follows: Figure 1b Given the examples 1, 2, 4, and 3 shown, the resulting combination of recommendation reasoning texts can include 1, 2, and 4. The permutation and combination results are as follows: Figure 1c Given the examples 2, 3, 4, and 1, the resulting combination of recommendation reasoning texts can include 2, 3, and 4.

[0116] In another embodiment, multiple combinations of recommendation reason texts are obtained based on the client's interface size and the length of each recommendation reason text. This may further include: randomly selecting at least two recommendation reason texts from the multiple recommendation reason texts whose total length is less than or equal to the width of the display area of ​​the recommendation reason text in the client's interface to obtain a combination of recommendation reason texts.

[0117] In another embodiment of this disclosure, each combination of recommendation reason texts may include all recommendation reason texts for the target store. Therefore, multiple recommendation reason texts can be arranged and combined to obtain multiple permutation and combination results, and each permutation and combination result can be considered as a single recommendation reason text combination.

[0118] Step S3200: Obtain the target user feature vector corresponding to the target user, the target store feature vector corresponding to the target store, and the combination feature vector corresponding to each combination of recommendation reason copy.

[0119] In one embodiment of this disclosure, multiple user feature vectors (including target users) and multiple store feature vectors (including target stores) may be pre-stored in a feature server or recommendation server. Each user feature vector has a unique user identifier corresponding to the user, and each store feature vector has a unique store identifier corresponding to the store, so as to facilitate finding the user feature vector of the corresponding user based on the user identifier and finding the store feature vector of the corresponding store based on the store identifier.

[0120] Therefore, obtaining the target user feature vector corresponding to the target user can include: obtaining the user identifier of the target user, and searching for the user feature vector corresponding to the target user from the feature server based on the user identifier, as the target user feature vector. Similarly, obtaining the target store feature vector corresponding to the target store can include: obtaining the store identifier of the target store, and searching for the store feature vector corresponding to the target store from the feature server based on the store identifier, as the target store feature vector.

[0121] In this embodiment, both the target user feature vector and the target store feature vector have the target dimension. The target dimension can be pre-defined based on the application scenario or specific requirements; for example, the target dimension can be 32-dimensional.

[0122] In an embodiment where the recommendation server stores user feature vectors of multiple users and store feature vectors of multiple stores, before executing step S3200 to obtain the target user feature vector corresponding to the target user and the target store feature vector corresponding to the target store, the method may further include the step of generating and storing the target user feature vector and the target store feature vector, specifically including the following steps S4110 to S4190:

[0123] Step S4110: Obtain the first feature vector of the target user. The first feature vector includes multiple user features reflecting the target user's preferences for the store and the recommendation text.

[0124] The user characteristics in this embodiment can be pre-defined according to the application scenario or specific needs. For example, they may include user age, user gender, user region, the first set number of store types most frequently selected by the user, and the second set number of product types most frequently selected by the user.

[0125] The first feature vector X1 includes user features x1 that reflect the target user's preferences for the store and the copywriting of the recommendation reasons. j The value of j is a natural number from 1 to n, where n represents the total number of user features that the first feature vector X1 has.

[0126] In one example, the feature vector X1 can have 80 features, i.e., n = 80. In this case, the first feature vector X1 can be represented as X1 = (x11, x12, x13, ..., x1...). 79 ,x1 80 ).

[0127] Step S4120: Perform a first-dimensional transformation on the first feature vector to obtain a second feature vector of the target dimension. For example... Figure 4 As shown.

[0128] In this embodiment, a first feature vector can be transformed into a second feature vector of the target dimension based on a pre-trained first machine learning model. The pre-trained first machine learning model can be used to transform an n-dimensional first feature vector into a second feature vector of the target dimension. For example, the pre-trained first machine learning model can be a Transformer model.

[0129] Step S4130: Perform a second-dimensional transformation on the first feature vector to obtain a third feature vector of the target dimension. For example... Figure 4 As shown.

[0130] In this embodiment, a second-dimensional transformation can be performed on the first feature vector based on a pre-trained second machine learning model to obtain a third feature vector of the target dimension. The pre-trained second machine learning model can be used to transform the n-dimensional first feature vector into a third feature vector of the target dimension. The pre-trained second machine learning model is not the same as the pre-trained first machine learning model. For example, the pre-trained second machine learning model could be a DeepFM model.

[0131] Step S4140: Summate the second and third feature vectors to obtain the target user feature vector. For example... Figure 4 As shown.

[0132] Step S4150: Obtain the fourth feature vector of the target store, wherein the fourth feature vector includes multiple store features that influence the target user's preference for the target store.

[0133] The store features in this embodiment can be pre-defined according to the application scenario or specific needs. For example, they may include the store type, the product type of the third set number of products with the highest sales volume in the store, the store location, and the product type of the fourth set number of products with the highest user ratings in the store.

[0134] The fourth feature vector X4 includes store features x4 that influence the target user's preference for the target store. i The value of i is a natural number from 1 to m, where m represents the total number of store features possessed by the fourth feature vector X4.

[0135] In one example, the feature vector X4 can have 30 features, i.e., m = 30. In this case, the fourth feature vector X4 can be represented as X4 = (x41, x42, x43, ..., x4...). 29 x4 30 ).

[0136] Step S4160: Perform a third-dimensional transformation on the fourth feature vector to obtain the fifth feature vector of the target dimension. For example... Figure 4 As shown.

[0137] In this embodiment, a third-dimensional transformation can be performed on the fourth feature vector based on a pre-trained third machine learning model to obtain a fifth feature vector of the target dimension. The pre-trained third machine learning model can be a Transformer model that transforms the m-dimensional fourth feature vector into the fifth feature vector of the target dimension.

[0138] Step S4170: Based on the trained fourth machine learning model, perform a fourth-dimensional transformation on the fourth feature vector to obtain the sixth feature vector of the target dimension. For example... Figure 4 As shown.

[0139] In this embodiment, the fourth feature vector can be transformed into a sixth feature vector of the target dimension based on a pre-trained fourth machine learning model. The pre-trained fourth machine learning model can transform the m-dimensional fourth feature vector into the sixth feature vector of the target dimension. The pre-trained fourth machine learning model is not the same as the pre-trained third machine learning model. For example, the pre-trained fourth machine learning model could be a DeepFM model.

[0140] Furthermore, the trained third machine learning model is not the same model as the trained first machine learning model, and the trained fourth machine learning model is not the same model as the trained second machine learning model.

[0141] Step S4180: Summate the fifth and sixth feature vectors to obtain the target store feature vector. For example... Figure 4 As shown.

[0142] Step S4190: Store the target user feature vector and the target store feature vector.

[0143] In this embodiment, the first feature vector is subjected to a first-dimensional transformation and a second-dimensional transformation to obtain the corresponding second and third feature vectors of the target dimension. Then, the second and third feature vectors of the target dimension are summed to obtain the target user feature vector of the target dimension. The fourth feature vector is subjected to a third-dimensional transformation and a fourth-dimensional transformation to obtain the corresponding fifth and sixth feature vectors of the target dimension. Then, the fifth and sixth feature vectors of the target dimension are summed to obtain the target store feature vector of the target dimension. This can make the target copy combination selected based on the target user feature vector and the target store feature vector more effective and improve the accuracy of the recommendation copy pushed to the client.

[0144] Furthermore, this embodiment pre-generates and stores target user feature vectors and target store vectors, which can be directly invoked during step S3200. This reduces the amount of recommendation calculations, speeds up the recommendation process, and also reduces the computational pressure on the server.

[0145] In one embodiment of this disclosure, before performing step S3200 to obtain the target user feature vector corresponding to the target user and the target store feature vector corresponding to the target store, the method may further include the following steps S4210 to S4280:

[0146] Step S4210: Obtain multiple training samples. Each training sample includes a first feature vector of a sample user, a fourth feature vector of a sample store, and a combined feature vector of a combination of recommendation reason copy of the sample store.

[0147] Step S4220: Obtain the actual matching score for each training sample, where the actual matching score is the click-through rate of the sample user's recommendation reason copy combination for the sample store, or the revenue of the sample store per thousand impressions of the recommendation reason copy combination.

[0148] In an embodiment where the actual matching score is the click-through rate (CTR) of a sample user's recommendation reason copy combination for a sample store, the following can be implemented: First, sample user browsing and click log data can be pre-acquired. Based on this data, the number of times the sample store's recommendation reason copy combination is displayed to the sample user, and the number of times the sample user clicks on the combination, can be determined. Then, the quotient obtained by dividing the number of clicks by the number of displays is calculated, which represents the CTR of the sample user's recommendation reason copy combination for the sample store. The number of times the sample user clicks on the sample store's recommendation reason copy combination can be the sum of the number of times the sample user clicks on all the recommendation reason copy elements within the combination.

[0149] Step S4230: Based on the initial first machine learning model, process the first feature vector of the training sample to obtain the first processing result.

[0150] In this embodiment, the first undetermined coefficients of the initial first machine learning model can be used as variables, and the processing expression of the training sample can be determined based on the first feature vector of the training sample, which is then used as the first processing result.

[0151] Step S4240: Based on the initial second machine learning model, process the first feature vector of the training sample to obtain the second processing result.

[0152] In this embodiment, the second undetermined coefficients of the initial second machine learning model can be used as variables, and the processing expression of the training sample can be determined based on the first feature vector of the training sample, which is then used as the second processing result.

[0153] Step S4250: Based on the initial third machine learning model, process the fourth feature vector of the training samples to obtain the third processing result.

[0154] In this embodiment, the third undetermined coefficient of the initial third machine learning model can be used as a variable, and the processing expression of the training sample can be determined according to the fourth feature vector of the training sample, which is then used as the third processing result.

[0155] Step S4260: Based on the initial fourth machine learning model, process the fourth feature vector of the training samples to obtain the fourth processing result.

[0156] In this embodiment, the fourth undetermined coefficient of the initial fourth machine learning model can be used as a variable, and the processing expression of the training sample can be determined based on the fourth feature vector of the training sample, which is then used as the fourth processing result.

[0157] Step S4270: Summing the first and second processing results to obtain the fifth processing result; summing the third and fourth processing results to obtain the sixth processing result.

[0158] Step S4280: Based on the fifth and sixth processing results of the training samples, the combined feature vector, and the actual matching score, train the initial first machine learning model, the initial second machine learning model, the initial third machine learning model, and the initial fourth machine learning model to obtain the trained first machine learning model, the trained second machine learning model, the trained third machine learning model, and the trained fourth machine learning model.

[0159] In this embodiment, the combined feature vector of the training sample may be transformed to the target dimension to obtain the combined feature vector of the target dimension. Then, the fifth processing result, the sixth processing result, and the combined feature vector of the target dimension of the training sample are multiplied by a dot product to obtain the matching score expression of the training sample. Based on the matching score expression and the actual matching score of the training sample, a loss function is constructed. The loss function is solved to determine the first undetermined coefficient, the second undetermined coefficient, the third undetermined coefficient, and the fourth undetermined coefficient, thereby obtaining the trained first machine learning model, the trained second machine learning model, the trained third machine learning model, and the trained fourth machine learning model.

[0160] In one embodiment of this disclosure, each recommendation reason copy of the target store can have a corresponding feature vector. The combined feature vector corresponding to each combination of recommendation reason copy can be obtained by concatenating the feature vectors corresponding to the recommendation reason copy included in the corresponding combination of recommendation reason copy according to the arrangement order of each recommendation reason copy in the combination of recommendation reason copy.

[0161] For example, the combination of recommendation reason copy includes copy 1, copy 2 and copy 4. The feature vector corresponding to copy 1 is X71, the feature vector corresponding to copy 2 is X72 and the feature vector corresponding to copy 4 is X74. Then, the combined feature vector corresponding to this combination of recommendation reason copy can be represented as {X71,X72,X74}.

[0162] Step S3300: Based on the target user feature vector, the target store feature vector, and the combined feature vector, determine the combination of recommendation reason copy that matches the target user and the target store, and use it as the target copy combination.

[0163] The target copy combination selected in this embodiment can best match the target user's preference for the target store's recommendation copy, and can also maximize the filling of the corresponding display position on the client interface, thereby improving the display effect of the target copy combination.

[0164] In one embodiment of this disclosure, determining a combination of recommendation reason texts matching the target user and the target store, based on the target user feature vector, the target store feature vector, and the combined feature vector, as the target text combination, may include the following steps S3310 to S3330:

[0165] Step S3310 involves transforming the combined feature vector to the target dimension, resulting in a combined feature vector with the target dimension. For example... Figure 4 As shown.

[0166] In one embodiment of this disclosure, the combined feature vector may be mapped to the space of the target dimension to obtain the combined feature vector of the target dimension.

[0167] Step S3320 involves performing a dot product operation on the target user feature vector, the target store feature vector, and the combined feature vector of the target dimension to obtain the predicted matching score for the target user, the target store, and each combination of recommendation reason copy. For example... Figure 4 As shown.

[0168] Step S3330: Select the combination of recommendation reasons with the highest predicted matching score with the target user and the target store as the target combination of recommendation reasons.

[0169] For example, in such Figure 1a The predicted match score for the combination of recommendation reasons shown can be 0.6, such as... Figure 1b The predicted match score for the combination of recommendation reasons shown can be 0.7, such as... Figure 1c The predicted match score for the combination of recommendation reasons shown is 0.65. Therefore, it could be as follows: Figure 1b The combination of recommendation reasons shown is used as the target combination of recommendation reasons.

[0170] The method in this embodiment can reduce the computational load of selecting target copy combinations and improve the computational speed of selecting target copy combinations.

[0171] Step S3400: Push the target copy combination and the target store's store information to the client for display.

[0172] In this embodiment, the recommendation server pushes the target copy combination and the target store's information to the client. The client can display the target store's store card on the interface. The target store's store card can display the target store's information and all the recommendation reasons in the target copy combination, so as to provide decision-making information for the target user and highlight the store's characteristics.

[0173] Through the embodiments of this disclosure, multiple recommendation reasons can be maximized to occupy the corresponding display space on the client's interface, thereby improving the display effect of the recommendation reasons. Furthermore, when a target store has multiple recommendation reasons, the recommendation reasons can be personalized for each target user, resulting in personalized recommendations of combinations of recommendation reasons.

[0174] <Server>

[0175] In this embodiment, a server 5000 is also provided, such as Figure 5 As shown, it includes a first memory 5100 and a first processor 5200.

[0176] The first memory 5100 is used to store an executable first computer program; the first computer program is used to control the first processor 5200 to execute any of the store recommendation methods based on recommendation reasons provided in this embodiment.

[0177] <Second Embodiment>

[0178] <Method>

[0179] This embodiment provides a method for displaying store information, which can be implemented by a client. In one example, the client may include, for example... Figure 2 The client shown is 2000.

[0180] like Figure 6 As shown, the method for displaying store information includes steps S6100 to S6300.

[0181] Step S6100: In response to the target user's target operation performed through the client, a store recommendation request for the target user is sent to the server. The server responds to the store recommendation request, determines the combination of recommendation reason texts that matches the target user and the target store, and pushes the target text combination and the store information of the target store to the client. Each combination of recommendation reason texts includes at least two recommendation reason texts with an ordered sequence.

[0182] Among them, the target user can be a user who logs into the target application through this client.

[0183] In this embodiment, the target operation can be either the target user opening the target application or the target user opening the target page of the target application.

[0184] When the server receives a store recommendation request for a target user, it may execute the store recommendation method based on store recommendation reasons described in the first embodiment. In response to the store recommendation request, it obtains the target store to be displayed and multiple combinations of recommendation reason texts for the target store. It also obtains the target user feature vector, the target store feature vector, and the combination feature vector corresponding to each combination of recommendation reason texts. Based on the target user feature vector, the target store feature vector, and the combination feature vector, it determines the combination of recommendation reason texts that matches the target user and the target store, and uses this as the target text combination. The server then pushes the target text combination and the store information of the target store to the client. Each combination of recommendation reason texts includes at least two recommendation reason texts with a specific order. For details, please refer to the first embodiment; further elaboration is omitted here.

[0185] Step S6200: Receive the target copy combination and the target store's store information.

[0186] Step S6300: Display the store information and target copy combination of the target store in the store card corresponding to the target store.

[0187] The client can display a store card for the target store on the interface. The store card can display the store information of the target store, as well as all the recommendation reasons in the target copywriting combination, to provide decision-making information for the target users and highlight the store's characteristics.

[0188] Through the embodiments of this disclosure, multiple recommendation reasons can be maximized to occupy the corresponding display space on the client's interface, improving the display effect of the recommendation reasons. Furthermore, when a target store has multiple recommendation reasons, the recommendation reasons can be personalized for the target user, resulting in personalized recommendations of combinations of recommendation reasons. This allows the store information in the store card and all the recommendation reasons in the target recommendation combination to better highlight the store's unique characteristics, thus providing decision-making information for the target user.

[0189] In one embodiment of this disclosure, displaying the target copy combination may include: displaying the recommendation reason copy included in the target copy combination in a full frame within the recommendation reason display box of the store card.

[0190] Displaying the recommendation reasons included in the target copy combination within the full-frame recommendation reason display box of the store card ensures that the recommendation reasons from the target copy combination fill the entire display box. This recommendation reason display box can be a frame used to display the recommendation reasons included in the target copy combination; the client can choose to display the border of this recommendation reason display box in any color, or it can be hidden.

[0191] This can improve the presentation of the recommendation text.

[0192] <Client>

[0193] In this embodiment, a client 7000 is also provided, such as Figure 7 As shown, it includes a second memory 7100 and a second processor 7200.

[0194] The second memory 7100 is used to store an executable second computer program; the second computer program is used to control the second processor 7200 to execute any of the store information display methods provided in this embodiment.

[0195] <Third Embodiment>

[0196] <System>

[0197] In this embodiment, a store recommendation system 8000 based on recommendation reasons is also provided, such as... Figure 8 As shown, it includes the server 5000 described in the first embodiment and the client 7000 described in the second embodiment.

[0198] Through the embodiments of this disclosure, multiple recommendation reasons can be maximized to occupy the corresponding display space on the client's interface, thereby improving the display effect of the recommendation reasons. Furthermore, when a target store has multiple recommendation reasons, the recommendation reasons can be personalized for each target user, resulting in personalized recommendations of combinations of recommendation reasons.

[0199] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0200] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0201] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0202] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0203] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0204] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0205] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0206] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It will be known to those skilled in the art that implementation in hardware, implementation in software, and implementation in a combination of software and hardware are equivalent.

[0207] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, and are not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of this disclosure is defined by the appended claims.

Claims

1. A method for recommending stores based on store recommendation reasons, characterized in that, The method is used on a server, and the method includes: In response to a store recommendation request sent by the client for a target user, obtain the target store to be displayed and a combination of multiple recommendation reason texts for the target store; Each of the aforementioned combination of recommendation reason texts includes at least two recommendation reason texts arranged in a specific order; Obtain the target user feature vector corresponding to the target user, the target store feature vector corresponding to the target store, and the combination feature vector corresponding to each combination of recommendation reason copy. Based on the target user feature vector, the target store feature vector, and the combined feature vector, a combination of recommendation reason copywriting that matches the target user and the target store is determined as the target copywriting combination; The target copy combination and the target store's store information are pushed to the client for display. Obtain the combination of multiple recommendation reason texts, including: Obtain multiple recommendation reasons for the target store; Based on the interface size of the client and the length of each recommendation reason text, multiple combinations of recommendation reason texts are obtained; The process of obtaining multiple combinations of recommendation reason texts based on the client's interface size and the length of each recommendation reason text includes: The multiple recommendation reasons are arranged and combined to obtain multiple permutation and combination results; Based on the interface size of the client and the length of each recommendation reason text, at least two recommendation reason texts to be displayed on the client's interface in each permutation and combination result are determined as the recommendation reason text combination; The dimensions of both the target user feature vector and the target store feature vector are the target dimensions; The step of determining a combination of recommendation reason texts matching the target user and the target store, based on the target user feature vector, the target store feature vector, and the combined feature vector, as the target text combination, includes: The combined feature vector is subjected to target dimension transformation to obtain the combined feature vector of the target dimension; Perform a dot product operation on the target user feature vector, the target store feature vector, and the combined feature vector of the target dimension to obtain the predicted matching score of the target user, the target store, and each combination of the recommendation reason copy. The combination of recommendation reason copy with the highest predicted matching score with the target user and the target store is selected as the target copy combination.

2. The method according to claim 1, characterized in that, Before obtaining the target user feature vector corresponding to the target user and the target store feature vector corresponding to the target store, the method further includes: Obtain a first feature vector of the target user, wherein the first feature vector includes multiple user features reflecting the target user's preferences for stores and recommendation reason copy; The first feature vector is transformed in the first dimension to obtain the second feature vector in the target dimension. The first feature vector is transformed in the second dimension to obtain the third feature vector in the target dimension; The target user feature vector is obtained by summing the second feature vector and the third feature vector. Obtain the fourth feature vector of the target store, wherein the fourth feature vector includes multiple store features that influence the target user's preference for the target store; The fourth feature vector is transformed in the third dimension to obtain the fifth feature vector in the target dimension; The fourth feature vector is transformed in the fourth dimension to obtain the sixth feature vector in the target dimension; The target store feature vector is obtained by summing the fifth feature vector and the sixth feature vector. Store the target user feature vector and the target store feature vector.

3. A method for displaying store information, characterized in that, The method is used on a client and includes: in response to a target operation performed by a target user through the client, sending a store recommendation request to the server for the target user, so that the server responds to the store recommendation request, determines a combination of recommendation reason text matching the target user and the target store as a target text combination, and pushes the target text combination and the store information of the target store to the client; Each of the aforementioned combination of recommendation reason texts includes at least two recommendation reason texts arranged in a specific order; Receive the target copy combination and the store information of the target store; The store card corresponding to the target store displays the store information of the target store and the target copywriting combination; The method further includes obtaining the combination of the multiple recommendation reason texts, including: Obtain multiple recommendation reasons for the target store; Based on the interface size of the client and the length of each recommendation reason text, multiple combinations of recommendation reason texts are obtained; The process of obtaining multiple combinations of recommendation reason texts based on the client's interface size and the length of each recommendation reason text includes: The multiple recommendation reasons are arranged and combined to obtain multiple permutation and combination results; Based on the interface size of the client and the length of each recommendation reason text, at least two recommendation reason texts to be displayed on the client's interface in each permutation and combination result are determined as the recommendation reason text combination; Obtain the target user feature vector corresponding to the target user, the target store feature vector corresponding to the target store, and the combination feature vector corresponding to each combination of recommendation reason copy. The dimensions of both the target user feature vector and the target store feature vector are the target dimensions; The step of determining a combination of recommendation reason texts matching the target user and the target store, based on the target user feature vector, the target store feature vector, and the combined feature vector, as the target text combination, includes: The combined feature vector is subjected to target dimension transformation to obtain the combined feature vector of the target dimension; Perform a dot product operation on the target user feature vector, the target store feature vector, and the combined feature vector of the target dimension to obtain the predicted matching score of the target user, the target store, and each combination of the recommendation reason copy. The combination of recommendation reason copy with the highest predicted matching score with the target user and the target store is selected as the target copy combination.

4. The method according to claim 3, characterized in that, The target copywriting combination is shown to include: The recommendation reasons included in the target copy combination will be displayed in the recommendation reason display box of the store card.

5. A server, characterized in that, It includes a first processor and a first memory, the first memory being used to store an executable first computer program; The first computer program is used to control the first processor to perform the method as described in any one of claims 1 to 2.

6. A client, characterized in that, It includes a second processor and a second memory, the second memory being used to store an executable second computer program; The second computer program is used to control the second processor to perform the method as described in claim 3 or 4.

7. A store recommendation system based on recommendation reasons, comprising the server as described in claim 5 and the client as described in claim 6.

Images