Recommendation method, recommendation device, electronic device and computer-readable storage medium

Abstract

This application provides a recommendation method, recommendation device, electronic device, and computer-readable storage medium. By providing reverse recommendation results that meet the recommendation criteria included in the target user's recommendation request but do not match the target user's preference attribute tags, this application achieves a contrasting comparison of positive and negative attributes in the recommended products, providing users with prominent recommendation resources and improving the success rate of recommendations.

Description

Technical Field

[0001] The present application relates to the field of personalized recommendation technology, and in particular to a recommendation method, a recommendation device, an electronic device, and a computer-readable storage medium. Background Art

[0002] With the rapid development of e-commerce, various e-commerce platforms provide users with more and more choices, which makes it difficult for users to accurately choose the resources they need, especially for some "choice-difficult" users who do not have a clear choice tendency. When faced with a large number of similar resources, it is even more difficult for them to make a satisfactory choice. The resource recommendation method in the existing technology often treats all users equally. In the process of capturing big data, the behavioral characteristics of these "choice-difficult" users who do not have a clear choice tendency are not taken into account, and it is impossible to provide prominent recommended products for such users, which seriously affects the shopping experience of such users and further affects the sales of resources. Summary of the invention

[0003] In view of this, the present application proposes a recommendation method, a recommendation device, an electronic device and a computer-readable storage medium, which improve the success rate of recommendations by comparing the positive and negative attributes of the recommended products.

[0004] This application provides a recommended method, including:

[0005] In response to a recommendation request from a target user, a list of positive recommendation results corresponding to the recommendation request is determined by a pre-trained first recommendation model; wherein the target user is a first type of user who has not adopted the historical recommendation result within a specified time interval from the moment the historical recommendation result is obtained; the first recommendation model is a recommendation model trained using the historical behavior data of a second type of user as training data; the second type of user is a user who has adopted the historical recommendation result within a specified time interval from the moment the historical recommendation result is obtained; the adoption probability of the positive recommendation result in the positive recommendation result list is not less than a preset adoption probability;

[0006] Determine a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preferred attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition included in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preferred attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the preferred attribute of the target user;

[0007] A comprehensive recommendation result list is determined based on the positive recommendation result list and the negative recommendation result list, and the comprehensive recommendation result list is recommended to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result.

[0008] The present application also provides a recommendation device, which includes:

[0009] A user screening module is configured to determine whether the target user has a clear purchasing tendency based on the historical operation information of the target user;

[0010] The first confirmation module is configured to respond to the recommendation request of the target user and determine the positive recommendation result list corresponding to the recommendation request through the pre-trained first recommendation model; wherein the target user is a first type of user who browses or selects the historical recommendation results within a specified time interval from the moment of obtaining the historical recommendation results but does not adopt the historical recommendation results for a number of times exceeding a preset threshold; the first recommendation model is a recommendation model trained with the historical behavior data of the second type of users as training data; the second type of users are users who adopt the historical recommendation results within a specified time interval from the moment of obtaining the historical recommendation results; the adoption probability of the positive recommendation results in the positive recommendation result list is not less than a preset adoption probability;

[0011] The second confirmation module is configured to determine a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preference attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition included in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preference attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the preferred attribute of the target user;

[0012] The recommendation module is configured to determine a comprehensive recommendation result list based on the positive recommendation result list and the negative recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result.

[0013] The present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the recommendation methods described above when executing the computer program.

[0014] The present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute any one of the recommendation methods described above.

[0015] From the above, it can be seen that the recommendation method, recommendation device, electronic device and computer-readable storage medium provided by the present application determine the positive recommendation result list corresponding to the recommendation request through a pre-trained first recommendation model; further, determine the reverse recommendation result list according to the attribute label of the initial reverse recommendation result and the preference attribute label of the target user; finally, determine the comprehensive recommendation result list according to the positive recommendation result list and the reverse recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list contains at least one positive recommendation result and at least one reverse recommendation result. The present application realizes the positive and negative contrast of the attributes of the recommended products through the reverse recommendation results that meet the recommendation conditions contained in the recommendation request of the target user but do not match the preference attribute label of the target user, so as to provide users with prominent recommendation resources and improve the success rate of recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A schematic diagram of an application scenario of the recommended method provided in an embodiment of the present application.

[0018] Figure 2 A schematic diagram of another application scenario of the recommended method provided in an embodiment of the present application.

[0019] Figure 3 A flowchart of a recommended method provided in an embodiment of the present application.

[0020] Figure 4 A schematic diagram of a recommended device provided in an embodiment of the present application.

[0021] Figure 5 This is a schematic diagram of the hardware structure of the electronic device proposed in this application. DETAILED DESCRIPTION

[0022] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0023] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be the usual meanings understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0024] As described in the background technology section, nowadays in various e-commerce platforms, music platforms or games, different recommendation methods are used to recommend resources such as goods, music, game props or game play to users.

[0025] However, the applicant has found through research that most of the current recommendation methods recommend the items that the backend server considers the user to be "most liked" to the client for display. However, if the user "likes" or "almost likes" the products, music, game props or game play they see, it will be difficult for the user to make a decision. As a result, many users with "choice difficulty" end up not buying anything, which seriously affects the shopping experience of such users and further affects the sales of resources.

[0026] In response to the above problems, the present application provides a recommendation method, a recommendation device, an electronic device and a computer-readable storage medium.

[0027] The technical solution of the present application is further described in detail below through specific embodiments.

[0028] refer to Figure 1 , which is a schematic diagram of an application scenario of the recommended method provided in an embodiment of the present application.

[0029] The application scenario includes a terminal device 101 .

[0030] Terminal devices include but are not limited to desktop computers, mobile phones, mobile computers, tablet computers, media players, smart wearable devices, personal digital assistants (PDAs) or other electronic devices that can achieve the above functions.

[0031] The terminal device 101 is used to provide a search box or a recommendation request receiving channel for the user. The terminal device has certain data storage and data computing capabilities.

[0032] Combine the following Figure 1 The application scenario is used to describe the recommended method according to the exemplary embodiment of the present application. It should be noted that the above application scenario is only shown to facilitate understanding of the spirit and principle of the present application, and the implementation of the present application is not limited in this respect. On the contrary, the implementation of the present application can be applied to any applicable scenario.

[0033] refer to Figure 2 , which is a schematic diagram of another application scenario of the recommended method provided in an embodiment of the present application.

[0034] The application scenario includes a terminal device 101, a server 102 and a data storage system 103. The terminal device 101, the server 102 and the data storage system 103 may be connected via a wired or wireless communication network.

[0035] The terminal device 101 includes but is not limited to a desktop computer, a mobile phone, a mobile computer, a tablet computer, a media player, a smart wearable device, a personal digital assistant (PDA) or other electronic devices capable of realizing the above functions.

[0036] The server 102 is used to provide data computing services for terminal devices.

[0037] The data storage system 103 is used to provide data storage services for terminal devices.

[0038] Among them, the data storage system and server can be independent physical servers, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, as well as big data and artificial intelligence platforms.

[0039] Combine the following Figure 2 The information processing method according to the exemplary embodiment of the present disclosure is described by using the application scenario. It should be noted that the above application scenario is only shown to facilitate understanding of the spirit and principle of the present disclosure, and the embodiments of the present disclosure are not limited in this respect. On the contrary, the embodiments of the present disclosure can be applied to any applicable scenario.

[0040] refer to Figure 3 , which is a flow chart of a recommended method provided in an embodiment of the present application.

[0041] Step S301, in response to a recommendation request from a target user, determining a list of positive recommendation results corresponding to the recommendation request through a pre-trained first recommendation model; wherein the target user is a first type of user who browses or selects historical recommendation results but does not adopt historical recommendation results for a number of times exceeding a preset threshold within a specified time interval from the moment the historical recommendation results are obtained; the first recommendation model is a recommendation model trained using historical behavior data of a second type of user as training data; the second type of user is a user who adopts the historical recommendation results within a specified time interval from the moment the historical recommendation results are obtained; the adoption probability of the positive recommendation results in the positive recommendation result list is not less than a preset adoption probability.

[0042] In specific implementations, recommendation requests can be understood as user search conditions. For example, when a user is browsing and purchasing products on an e-commerce platform and wants to buy a coat, the user can enter search conditions related to "coat" in the search box of the platform, such as: denim jacket, long-sleeved jacket, light-colored jacket, dark-colored jacket, etc.

[0043] In the specific implementation, users can be divided into two categories. One is the second type of users with clear purchasing tendencies. It can be understood that this type of users can accurately and quickly select or purchase the products they need. After initiating the search behavior, they can make a "decisive" choice based on the recommendation list returned by the server. The other is the first type of users who do not have a clear purchasing tendency. It can be understood that this type of users have a certain "choice difficulty syndrome". They cannot make a clear judgment on similar products and "like" or "like almost" many similar products.

[0044] In a specific implementation, whether the user is a first-type user can be determined based on the user's historical behavior. For example, if the number of times that the target user browses or selects historical recommendation results but does not adopt the historical recommendation results within a specified time interval from the time the historical recommendation results are obtained exceeds a preset threshold, then the target user can be determined to be a first-type user. For example, if the preset threshold is 5 times, and the target user's recommendation request is a recommendation request for a "coat", then the positive recommendation results may include "warm-colored coats", "cold-colored coats" or "gradient-colored coats". When the target user adds the "warm-colored coat" in the recommendation results to the shopping cart within 5 minutes, removes it from the shopping cart, and then adds it to the shopping cart again, and repeats the operation for more than 5 times, then the target user can be determined to be a first-type user, and this type of user is a "choice-difficulty" user.

[0045] In a specific implementation, whether the user is a second type of user can be determined based on the user's historical behavior. For example, if the user browses or selects the historical recommendation results and adopts the historical recommendation results within a specified time interval from the time the historical recommendation results are obtained, the user can be determined to be a second type of user. For example, if the user's recommendation request is a recommendation request for "jacket", when the user adds the "denim jacket" in the recommendation result to the shopping cart within 5 minutes and directly pays for the purchase, then the target user can be determined to be a second type of user, which is a user with a clear purchase tendency.

[0046] As an optional embodiment, the user's historical behavior may also be that the user repeatedly searches for a certain category but does not select or purchase any products in the category, or the user clicks back and forth between multiple similar items to browse, or the user repeatedly changes the search terms, but the search terms point to the same category, such as "watermelon", "Qilin watermelon" and "Qilin melon", or the user's browsing time on a certain item exceeds 3 times the browsing time of all users on the item, and this situation has occurred multiple times in the past month, or the user clicks on different items and different customer service inquiries after a search, but does not actually purchase.

[0047] As an optional embodiment, since the first type of users cannot make a clear choice, but the second type of users can make a clear choice, the historical behavior data of the second type of users is used as a training data set, and the first recommendation model is obtained by fitting the training data set.

[0048] In a specific implementation, the type of the first recommendation model may be, but is not limited to, support vector machine (SVM), eXtreme Gradient Boosting, and the like.

[0049] During the training process, the input of the first recommendation model may be, but is not limited to, the user's historical operation information and the user's search behavior information. The user's historical operation information includes, but is not limited to, the user's historical operation records of clicking, browsing, and purchasing products, and the user's search behavior information includes, but is not limited to, the user's searched product categories, game props, music tracks, composers, and music categories. The predicted output of the first recommendation model is each product and its adoption probability. For example, the output format of the first recommendation model can be [(item1, 0.91), (item2, 0.73), (item3, 0.82) ...], (item1, 0.91) indicates that the product code corresponding to the output result is item1, and the probability of the user adopting the product is 91%.

[0050] Furthermore, the recommendation results obtained by the first recommendation model can be used as positive recommendation results, and a positive recommendation list can be composed of several positive recommendation results, wherein the adoption probability of the positive recommendation results in the positive recommendation result list is not less than the preset adoption probability, the purpose of which is to ensure the recommendation success probability of the positive recommendation results.

[0051] Step S302, determining a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preference attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition contained in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preference attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the attribute of the target user's preference.

[0052] In a specific implementation, the target user's preference attribute label can be determined based on the user's recommendation request. The preference attribute label represents the target user's preferred attribute. The recommendation request may include recommendation conditions. For example, if the target user searches for a certain type of product, the target user may choose to purchase the product with high sales volume, bright colors, or free shipping. Or the target user may prefer to listen to a certain type of music and add a certain type of music to the frequently listened list many times. Or the target user may choose to purchase a certain game prop many times, such as a name change card, intimacy gift, experience / coin doubling card, etc. The target user's preference attribute label can be determined based on the user's historical operation information as described above.

[0053] As an optional embodiment, if the user's recommendation request is "warm-colored coat", it can be determined that the user's recommendation condition is "coat" and the user's preference attribute label is "warm color".

[0054] In specific implementations, since the target user is a user with "difficulty in making choices", even if resources are pushed to the user according to the user's recommendation request and recommendation conditions, the user may not be able to make a clear choice directly. For example, when the resources recommended to the user include "yellow jacket" and "beige jacket", since "yellow" and "beige" are both warm colors and there is no difference between the two, the target user may not be able to make a choice between the two.

[0055] Therefore, it is necessary to provide the user with a reverse recommendation result list, which is determined by the attribute label of the initial reverse recommendation result and the preference attribute label of the target user, wherein the initial reverse recommendation result is a recommendation result that meets the recommendation conditions contained in the recommendation request, and the attribute label of the reverse recommendation result in the reverse recommendation result list does not match the preference attribute label of the target user, and the attribute label represents the attribute of the reverse recommendation result.

[0056] As an optional embodiment, the initial reverse recommendation result is an adopted result in an adopted result list that meets the recommendation condition included in the recommendation request; the adopted result list includes recommendation results adopted within a specified time interval with the current time as the end time.

[0057] For example, if the user's preferred attribute label is "warm color", then the attribute label of the reverse recommendation result is "cold color". "Cold color" and "warm color" are two completely opposite attributes, and there will be a certain contrast between the two. For example, white and black, according to the attribute of "cold color", reverse recommendation results such as "gray coat", "black coat" or "dark blue coat" can be obtained. A reverse recommendation result list can be obtained based on several reverse recommendation results.

[0058] As an optional embodiment, if the preference attribute label is determined to be sales volume, it indicates that the target user is more inclined to purchase target products with higher sales volume. In order to create differentiation, the reverse recommendation result list includes the target products with lower sales volume.

[0059] As an optional embodiment, the adoption results in the adoption result list whose sales volume is not higher than a preset first sales volume threshold may be determined as the initial reverse recommendation results.

[0060] As an optional embodiment, if the preference index is determined to be a free shipping index, it indicates that the user is more inclined to purchase the target product with free shipping. In order to create differentiation, the reverse recommendation result list includes the target products without free shipping.

[0061] As an optional embodiment, if the preference indicator is determined to be soft music, it indicates that the user prefers soft-style music. In order to create a difference, the reverse recommendation result list contains fast-tempo music.

[0062] Step S303, determine a comprehensive recommendation result list based on the positive recommendation result list and the negative recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result positive recommendation result list.

[0063] As an optional embodiment, if the preference attribute tag includes sales volume, the positive recommendation results in the positive recommendation result list are sorted from large to small according to sales volume, and the sorted positive recommendation results are used as the positive recommendation result list;

[0064] sorting the initial reverse recommendation results from small to large according to sales volume, and using the sorted reverse recommendation results as the reverse recommendation result list; or

[0065] The initial reverse recommendation results are sorted from small to large according to sales volume, and the reverse recommendation results whose sales volume is not higher than a preset second sales volume threshold are used as the reverse recommendation result list; the second sales volume threshold is less than or equal to the first sales volume threshold.

[0066] As an optional embodiment, the positive recommendation result is obtained by sorting the sales volume from large to small according to the user's sales volume preference index. For example, the positive recommendation result list good_list = [item66, item52, item43, item42, item3, item31...], wherein item66 indicates that the sales volume of the product is the highest;

[0067] As an optional embodiment, the initial reverse recommendation result is obtained by sorting the sales volume from small to large based on the user's sales volume preference index. For example, the reverse recommendation result list bad_list = [item31, item3, item42, item43, item52, item66...]. Among them, item31 indicates that the sales volume of the product is the lowest;

[0068] Alternatively, a second sales threshold is set, and the second sales threshold is used to limit the attributes of the initial reverse recommendation result. For example, if the second sales threshold is 50%, the sales of the initial reverse recommendation result is not higher than 50%.

[0069] In a specific implementation, a comprehensive recommendation result list is determined based on the positive recommendation result list and the negative recommendation result list.

[0070] As an optional embodiment, at least one positive recommendation result is selected in the positive recommendation result list in turn, and the positive recommendation result is added to the comprehensive recommendation result list in sequence, and at least one negative recommendation result is selected in the negative recommendation result list in turn, and the negative recommendation result is added to the comprehensive recommendation result list in sequence to determine the comprehensive recommendation result list.

[0071] As an optional embodiment, the positive recommendation result list good_list and the negative recommendation result list bad_list can be intertwined and merged, and at least one positive recommendation result in the positive recommendation result list is selected in turn and added to the comprehensive recommendation result list, and at least one negative recommendation result in the negative recommendation result list is selected in turn and added to the last positive recommendation result in the comprehensive recommendation result list, until all the products in the positive recommendation result list and the negative recommendation result list are selected, so as to obtain a comprehensive recommendation product list.

[0072] In detail, the number of commodities selected from the positive recommendation result list, the positive recommendation result list and the negative recommendation result list in sequence each time can be set according to actual conditions.

[0073] As an optional embodiment, the comprehensive recommendation result list can also be generated by taking one of the positive recommendation result list and the second recommended product list as a basic list and the other as an insertion list, selecting at least one product from the insertion list in turn, and inserting one or two products in the basic list into the basic list to form a comprehensive recommendation product list.

[0074] Furthermore, product recommendations are made to the users who do not have a clear purchasing intention according to the comprehensive recommendation result list.

[0075] As an optional embodiment, if it is determined in step S301 that the user has a clear purchasing tendency, the positive recommendation result list obtained in step S302 is used as a direct recommendation product list, and products are recommended to the user with a clear purchasing tendency according to the direct recommendation product list.

[0076] According to the above, in the comprehensive recommendation result list, the products in the positive recommendation result list obtained according to the predicted purchase probability results are intertwined with the products sorted from small to large according to sales volume to form a contrasting comprehensive recommendation product list, so that a contrast is formed between at least every two products, which is more conducive to selection for users who do not have clear purchasing power.

[0077] From the above, it can be seen that the recommendation method, recommendation device, electronic device and computer-readable storage medium provided by the present application determine the positive recommendation result list corresponding to the recommendation request through a pre-trained first recommendation model; further, determine the reverse recommendation result list according to the attribute label of the initial reverse recommendation result and the preference attribute label of the target user; finally, determine the comprehensive recommendation result list according to the positive recommendation result list and the reverse recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list contains at least one positive recommendation result and at least one reverse recommendation result. The present application realizes the positive and negative contrast of the attributes of the recommended products through the reverse recommendation results that meet the recommendation conditions contained in the recommendation request of the target user but do not match the preference attribute label of the target user, so as to provide users with prominent recommendation resources and improve the success rate of recommendations.

[0078] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of the embodiment of the present application can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0079] It should be noted that the above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] Based on the same concept, the present application also provides a recommendation device.

[0081] refer to Figure 4 , which is a schematic diagram of a recommended device provided in an embodiment of the present application.

[0082] The device includes:

[0083] The first confirmation module 401 is configured to respond to the recommendation request of the target user and determine the positive recommendation result list corresponding to the recommendation request through the pre-trained first recommendation model; wherein the target user is a first type of user who browses or selects the historical recommendation result within a specified time interval from the moment of obtaining the historical recommendation result but does not adopt the historical recommendation result for a number of times exceeding a preset threshold; the first recommendation model is a recommendation model trained with the historical behavior data of the second type of user as training data; the second type of user is a user who adopts the historical recommendation result within a specified time interval from the moment of obtaining the historical recommendation result; the adoption probability of the positive recommendation result in the positive recommendation result list is not less than the preset adoption probability;

[0084] The second confirmation module 402 is configured to determine a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preference attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition included in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preference attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the attribute of the target user's preference;

[0085] The recommendation module 403 is configured to determine a comprehensive recommendation result list based on the positive recommendation result list and the negative recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result.

[0086] For the convenience of description, the above devices are described in terms of functions and modules. Of course, when implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0087] Optionally, the initial reverse recommendation result is an adopted result in an adopted result list that meets the recommendation condition included in the recommendation request; the adopted result list includes recommendation results adopted within a specified time interval with the current time as the end time.

[0088] Optionally, the second confirmation module 402 is further configured to determine the adoption result in the adoption result list whose sales volume is not higher than a preset first sales volume threshold as the initial reverse recommendation result.

[0089] Optionally, the preference attribute tag includes sales volume;

[0090] The second confirmation module 402 is further configured to sort the positive recommendation results in the positive recommendation result list according to sales volume from large to small, and use the sorted positive recommendation results as the positive recommendation result list.

[0091] The second confirmation module 402 is further configured to sort the initial reverse recommendation results from small to large according to sales volume, and use the sorted reverse recommendation results as the reverse recommendation result list; or

[0092] The initial reverse recommendation results are sorted from small to large according to sales volume, and the reverse recommendation results whose sales volume is not higher than a preset second sales volume threshold are used as the reverse recommendation result list; the second sales volume threshold is less than or equal to the first sales volume threshold.

[0093] Optionally, the first confirmation module 401 is further configured to obtain historical behavior data of the second type of users; wherein the historical behavior data at least includes historical recommendation requests of the second type of users and historical adoption results that meet the recommendation conditions included in the historical recommendation requests;

[0094] The historical behavior data of the second type of users is used as a training data set, and the first recommendation model is obtained by training according to the training data set.

[0095] Optionally, the recommendation module 403 is also configured to select at least one positive recommendation result in the positive recommendation result list in sequence, and add the positive recommendation result to the comprehensive recommendation result list in sequence, and to select at least one negative recommendation result in the negative recommendation result list in sequence, and add the negative recommendation result to the comprehensive recommendation result list in sequence, so as to determine the comprehensive recommendation result list.

[0096] Optionally, the recommendation module 403 is also configured to randomly select at least one positive recommendation result from the positive recommendation result list, and add the positive recommendation result to the comprehensive recommendation result list in sequence, and randomly select at least one negative recommendation result from the negative recommendation result list, and add the negative recommendation result to the comprehensive recommendation result list in sequence, so as to determine the comprehensive recommendation result list.

[0097] Optionally, the first confirmation module 401 is further configured to obtain the number of times the target user browses or selects the historical recommendation result but does not adopt the historical recommendation result within a specified time interval from the moment the historical recommendation result is obtained;

[0098] In response to the target user browsing or selecting historical recommendation results but not adopting the historical recommendation results for a number of times exceeding a preset threshold, the target user is determined to be a first type of user.

[0099] The devices in the above embodiments are used to implement the corresponding recommended methods in the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be described in detail here.

[0100] Based on the same concept, corresponding to any of the above-mentioned embodiments and methods, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the recommended method described in any of the above embodiments is implemented.

[0101] Figure 5 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0102] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0103] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0104] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0105] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0106] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0107] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0108] The electronic device of the above embodiment is used to implement the corresponding recommended method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0109] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the recommendation method described in any of the above embodiments.

[0110] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0111] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the recommendation method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0112] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0113] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0114] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0115] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.

Claims

1. A recommendation method, characterized in that: include: In response to a recommendation request from a target user, a list of positive recommendation results corresponding to the recommendation request is determined by a pre-trained first recommendation model; wherein the target user is a first type of user who has browsed or selected historical recommendation results but has not adopted historical recommendation results for a number of times exceeding a preset threshold within a specified time interval from the moment when the historical recommendation results are obtained; the first recommendation model is a recommendation model trained using historical behavior data of second type of users as training data; the second type of user is a user who has adopted the historical recommendation results within a specified time interval from the moment when the historical recommendation results are obtained; the adoption probability of the positive recommendation results in the positive recommendation result list is not less than a preset adoption probability; Determine a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preferred attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition included in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preferred attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the preferred attribute of the target user; A comprehensive recommendation result list is determined based on the positive recommendation result list and the negative recommendation result list, and the comprehensive recommendation result list is recommended to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result.

2. The method according to claim 1, characterized in that The initial reverse recommendation result is an adopted result in the adopted result list that meets the recommendation condition included in the recommendation request; the adopted result list includes the recommendation results adopted within a specified time interval with the current time as the end time.

3. The method according to claim 2, characterized in that Before determining the reverse recommendation result list according to the attribute label of the initial reverse recommendation result and the preferred attribute label of the target user, the method further includes: The adoption results in the adoption result list whose sales volume is not higher than a preset first sales volume threshold are determined as the initial reverse recommendation results.

4. The method according to claim 3, characterized in that The preference attribute tag includes sales volume, and the method further includes: sorting the positive recommendation results in the positive recommendation result list according to sales volume from large to small, and using the sorted positive recommendation results as the positive recommendation result list; Determining a reverse recommendation result list according to the attribute label of the initial reverse recommendation result and the preferred attribute label of the target user includes: sorting the initial reverse recommendation results from small to large according to sales volume, and using the sorted reverse recommendation results as the reverse recommendation result list; or The initial reverse recommendation results are sorted from small to large according to sales volume, and the reverse recommendation results whose sales volume is not higher than a preset second sales volume threshold are used as the reverse recommendation result list; the second sales volume threshold is less than or equal to the first sales volume threshold.

5. The method according to claim 1, characterized in that The method further includes training the first recommendation model by the following method: Acquire historical behavior data of the second type of users; wherein the historical behavior data at least includes historical recommendation requests of the second type of users and historical adoption results that meet the recommendation conditions included in the historical recommendation requests; The historical behavior data of the second type of users is used as a training data set, and the first recommendation model is obtained by training according to the training data set.

6. The method according to claim 1 or 5, characterized in that: Determining a comprehensive recommendation result list according to the positive recommendation result list and the negative recommendation result list includes: Select at least one positive recommendation result in the positive recommendation result list in turn, and add the positive recommendation result to the comprehensive recommendation result list in turn; select at least one negative recommendation result in the negative recommendation result list in turn, and add the negative recommendation result to the comprehensive recommendation result list in turn, so as to determine the comprehensive recommendation result list.

7. The method according to claim 1, characterized in that Determining a comprehensive recommendation result list according to the positive recommendation result list and the negative recommendation result list includes: At least one positive recommendation result is randomly selected from the positive recommendation result list, and the positive recommendation result is sequentially added to the comprehensive recommendation result list; at least one negative recommendation result is randomly selected from the negative recommendation result list, and the negative recommendation result is sequentially added to the comprehensive recommendation result list to determine the comprehensive recommendation result list.

8. The method according to claim 1, characterized in that The first type of user is determined by the following method, including: Obtaining the number of times the target user browses or selects the historical recommendation results but does not adopt the historical recommendation results within a specified time interval from the moment the historical recommendation results are obtained; In response to the target user browsing or selecting historical recommendation results but not adopting the historical recommendation results for a number of times exceeding a preset threshold, the target user is determined to be a first type of user.

9. A recommendation device, characterized in that: include: The first confirmation module is configured to respond to the recommendation request of the target user and determine the positive recommendation result list corresponding to the recommendation request through the pre-trained first recommendation model; wherein the target user is a first type of user who browses or selects the historical recommendation results within a specified time interval from the moment of obtaining the historical recommendation results but does not adopt the historical recommendation results for a number of times exceeding a preset threshold; the first recommendation model is a recommendation model trained with the historical behavior data of the second type of users as training data; the second type of users are users who adopt the historical recommendation results within a specified time interval from the moment of obtaining the historical recommendation results; the adoption probability of the positive recommendation results in the positive recommendation result list is not less than a preset adoption probability; The second confirmation module is configured to determine a reverse recommendation result list according to the attribute tag of the initial reverse recommendation result and the preference attribute tag of the target user; wherein the initial reverse recommendation result is a recommendation result that meets the recommendation condition included in the recommendation request, and the attribute tag of the reverse recommendation result in the reverse recommendation result list does not match the preference attribute tag of the target user; the attribute tag represents the attribute of the reverse recommendation result, and the preference attribute tag represents the preferred attribute of the target user; The recommendation module is configured to determine a comprehensive recommendation result list based on the positive recommendation result list and the negative recommendation result list, and recommend the comprehensive recommendation result list to the target user; the comprehensive recommendation result list includes at least one positive recommendation result and at least one negative recommendation result.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 8.

Images