Recommended methods, devices, equipment and storage media for the entire process

Abstract

The present application discloses a full-process recommendation method, device, equipment and storage medium, the method comprising: when detecting a behavior operation instruction of a target user on a first target channel, obtaining the process state of the target user on the first target channel, and obtaining a first comprehensive user portrait of the target user on multiple channels; determining a first response file that responds to the target user based on the process state, the first target channel and the first comprehensive user portrait; obtaining first feedback information of the target user generating behavior feedback for the first response file, and adjusting the first comprehensive user portrait based on the first feedback information to obtain a second comprehensive user portrait on multiple channels; and making targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on multiple channels. The present application aims to solve the technical problems of low accuracy and poor effect of data recommendation in the prior art.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology in financial technology (Fintech), and in particular to a full-process recommendation method, device, equipment and storage medium. Background Art

[0002] With the continuous development of financial technology, especially Internet technology finance, more and more technologies are applied in the financial field, but the financial industry also has higher requirements for technology, such as higher requirements for data recommendation.

[0003] At present, when making data recommendations, it is usually necessary to separately study or generate recommendation files for different recommendation channels to make data recommendations, or to separately analyze recommendation results. However, for the same data or business, there may be multiple recommendation channels. Separately studying or generating recommendation files for different recommendation channels to make data recommendations, or to separately analyze recommendation results, results in low recommendation efficiency and low recommendation accuracy. Summary of the invention

[0004] The main purpose of this application is to provide a full-process recommendation method, device, equipment and storage medium, aiming to solve the technical problems of low recommendation efficiency and low recommendation accuracy in the prior art.

[0005] To achieve the above objectives, the present application provides a full-process recommendation method, which includes:

[0006] When a behavior operation instruction of a target user on a first target channel is detected, a process state of the target user on the first target channel is obtained, and a first comprehensive user portrait of the target user on multiple channels is obtained;

[0007] Determine a first response file responding to the target user according to the process status, the first target channel and the first comprehensive user portrait;

[0008] Acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user profile according to the first feedback information to obtain a second comprehensive user profile on multiple channels;

[0009] According to the second comprehensive user portrait on multiple channels, targeted recommendations are made to the target user on the multiple channels.

[0010] Optionally, when a behavior operation instruction of a target user is detected, determining whether the target user is a logged-in user;

[0011] If the target user is a logged-in user, determining a target user unique identifier shared by the target user in multiple channels;

[0012] According to the target user unique identifier, the process status of the target user on the first target channel is obtained.

[0013] Optionally, the first response file includes dialogue reply content and guidance content, and the step of determining the first response file responding to the behavioral operation instruction according to the process state, the first target channel and the first comprehensive user portrait includes:

[0014] If the process state is the first preset process state, determining a target dialogue mode for responding to the behavioral operation instruction and a target dialogue reply content corresponding to the target dialogue mode according to the first target channel and the first comprehensive user portrait;

[0015] If the process state is the second preset process state, the target recommended text content that responds to the behavioral operation instruction and the guidance content corresponding to the target recommended text content are determined according to the first target channel and the first comprehensive user portrait.

[0016] Optionally, if the process state is the second preset process state, the step of determining, according to the first target channel and the first comprehensive user portrait, target recommended text content responding to the behavioral operation instruction and guidance content corresponding to the target recommended text content includes:

[0017] If the process state is the second preset process state, determining the target recommended text content that responds to the behavioral operation instruction according to the first target channel and the first comprehensive user portrait, and determining the display mode of the target recommended text content, the display mode including a graphic display mode, a pure text display mode and a pure picture display mode;

[0018] According to the display method, the display layout of the target recommended text content is determined to determine the guidance content corresponding to the target recommended text content.

[0019] Optionally, after the step of determining a first response file responding to the behavioral operation instruction according to the process state, the first target channel and the first comprehensive user portrait, the method further includes:

[0020] Acquire second feedback information fed back by the target user in response to the first response file;

[0021] According to the second feedback information, priority rankings of the multiple channels are obtained to perform channel circulation on the first target channel of the target user to obtain a second target channel;

[0022] According to the second target channel, the first response file for the target user is adjusted to a second response file.

[0023] Optionally, the step of performing targeted recommendations for the target user on the multiple channels according to the second comprehensive user portraits on the multiple channels includes:

[0024] determining a recommended topic for the target user based on the second comprehensive user portrait on multiple channels;

[0025] Based on the recommended topic, target content associated with the recommended topic is selected from preset selected resources of various channels;

[0026] The target content is recommended to the target user.

[0027] Optionally, the step of obtaining a first comprehensive user portrait of the target user on multiple channels includes:

[0028] Obtaining the target user's browsing records and order records on multiple channels;

[0029] Inputting the browsing records and ordering records into a preset recognition model, and performing portrait recognition processing on the browsing records and ordering records based on the preset recognition model to obtain a first comprehensive user portrait;

[0030] The preset recognition model is obtained by iterative training based on a preset basic model.

[0031] The present application also provides a full-process recommendation device, which includes:

[0032] A detection module, configured to, when detecting a behavior operation instruction of a target user on a first target channel, obtain a process state of the target user on the first target channel, and obtain a first comprehensive user portrait of the target user on multiple channels;

[0033] A determination module, configured to determine a first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait;

[0034] A first acquisition module is used to acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user portrait according to the first feedback information to obtain a second comprehensive user portrait on multiple channels;

[0035] A recommendation module is used to make targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on the multiple channels.

[0036] Optionally, the detection module includes:

[0037] A first determining unit, configured to determine whether the target user is a logged-in user when a behavior operation instruction of the target user is detected;

[0038] A first acquisition unit, configured to determine a target user unique identifier shared by the target user in multiple channels if the target user is a logged-in user;

[0039] The second acquisition unit is used to acquire the process status of the target user on the first target channel according to the unique identifier of the target user.

[0040] Optionally, the determining module includes:

[0041] A second determining unit is configured to determine, if the process state is the first preset process state, a target dialogue mode for responding to the behavioral operation instruction and a target dialogue reply content corresponding to the target dialogue mode according to the first target channel and the first comprehensive user portrait;

[0042] The third determination unit is used to determine the target recommended text content that responds to the behavioral operation instruction and the guidance content corresponding to the target recommended text content based on the first target channel and the first comprehensive user portrait if the process state is the second preset process state.

[0043] Optionally, the third determining unit includes:

[0044] A first determining subunit is used to determine, if the process state is a second preset process state, a target recommended text content that responds to the behavioral operation instruction according to the first target channel and the first comprehensive user portrait, and determine a display mode of the target recommended text content, wherein the display mode includes a graphic display mode, a pure text display mode, and a pure picture display mode;

[0045] The second determining subunit is used to determine the display layout of the target recommended text content according to the display mode, so as to determine the guidance content corresponding to the target recommended text content.

[0046] Optionally, the whole process recommendation device further includes:

[0047] A second acquisition module, configured to acquire second feedback information fed back by the target user in response to the first response file;

[0048] The third acquisition module is used to acquire the priority rankings of the multiple channels according to the second feedback information, so as to perform channel circulation on the first target channel of the target user to obtain a second target channel.

[0049] An adjustment module is used to adjust the first response file for the target user into a second response file according to the second target channel.

[0050] Optionally, the recommendation module includes:

[0051] a fourth determining unit, configured to determine a recommended topic for the target user based on the second comprehensive user portrait on multiple channels;

[0052] An association unit, configured to select target content associated with the recommended topic from preset selected resources of various channels based on the recommended topic;

[0053] A recommendation unit is used to recommend the target content to the target user.

[0054] Optionally, the detection module includes:

[0055] A third acquisition unit is used to acquire the target user's browsing records and order records on multiple channels;

[0056] an identification unit, configured to input the browsing records and order records into a preset identification model, and perform portrait identification processing on the browsing records and order records based on the preset identification model to obtain a first comprehensive user portrait;

[0057] The preset recognition model is obtained by iterative training based on a preset basic model.

[0058] The present application also provides a storage medium, on which is stored a program for implementing the above-mentioned full-process recommendation method. When the program for the full-process recommendation method is executed by a processor, the steps of the full-process recommendation method as described above are implemented.

[0059] This application obtains the process status of the target user on the first target channel and the first comprehensive user portrait of the target user on multiple channels when a behavioral operation instruction of the target user on the first target channel is detected; determines a first response file in response to the target user based on the process status, the first target channel and the first comprehensive user portrait; obtains first feedback information of the target user generating behavioral feedback for the first response file, and adjusts the first comprehensive user portrait based on the first feedback information to obtain a second comprehensive user portrait on multiple channels; and makes targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on multiple channels. Compared with the prior art that studies or generates recommendation files for different recommendation channels separately for data recommendation, this application studies or generates recommendation files for multiple recommendation channels as a whole for targeted data recommendation, thereby improving recommendation efficiency, and in this application,

[0060] Based on the overall research of richer data from multiple channels or the generation of recommendation files for data recommendation, the recommendation accuracy is improved, the recommendation effect is improved, and the technical defect of low recommendation accuracy in the prior art is overcome. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0063] Figure 1 This is a flowchart of the first embodiment of the recommended method for the entire process of this application;

[0064] Figure 2 This is a schematic diagram of the detailed steps of S10 in the recommended method of the whole process of this application;

[0065] Figure 3 This is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present application.

[0066] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0067] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0068] The present application embodiment provides a full-process recommendation method. In the first embodiment of the full-process recommendation method of the present application, refer to Figure 1 , the recommended method of the whole process includes:

[0069] Step S10, when a behavior operation instruction of a target user on a first target channel is detected, a process state of the target user on the first target channel is obtained, and a first comprehensive user portrait of the target user on multiple channels is obtained;

[0070] Step S20, determining a first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait;

[0071] Step S30, obtaining first feedback information of the target user generating behavioral feedback for the first response file, and adjusting the first comprehensive user portrait according to the first feedback information to obtain a second comprehensive user portrait on multiple channels;

[0072] Step S40: Perform targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on the multiple channels.

[0073] Step S10, when a behavioral operation instruction of a target user on a first target channel is detected, the process state of the target user on the first target channel is obtained, and a first comprehensive user portrait of the target user on multiple channels is obtained.

[0074] On the whole, in this embodiment, it should be noted that it can be applied to the recommendation system of the whole process. The recommendation system of the whole process is subordinate to the recommendation device of the whole process. For the recommendation system of the whole process, various channels can be embedded, such as various APPs, small programs, etc., specifically, small programs, APPs, H5 are embedded in the form of floating windows, and the public account menu bar is connected, or the recommendation system of the whole process is connected to various recommendation channels for external communication to obtain user-related information of various channels. The user-related information obtained by the recommendation system of the whole process includes the user's reading records on multiple channels, the purchase records or order records on multiple channels, etc. The user's reading records on multiple channels include the length of stay on multiple channels, the process status on multiple channels and other information.

[0075] Specifically, when customers come into contact with the product, there may be the following paths:

[0076] 1. Advertising: You can directly generate actions (purchase, download, etc.), or leave contact information and wait for the account manager to contact you;

[0077] 2. APP, mini-programs, and public accounts: Learn about products, consult customer service, and take actions after consultation. A small number of people may leave contact information.

[0078] 3. Recommendation by people (shared by account managers and friends): seeing product information and first contact; adding WeChat or chatting, which may generate interest and take action later;

[0079] 4: SMS push.

[0080] Based on the above method, user-related information is obtained, including but not limited to clicked questions, consultation records, viewed business cards, added contact information and other information left. User-related information can be managed through a unified entrance of the full-process recommendation system to obtain more data, records, and files to analyze and support recommendation files, thereby significantly improving recommendation efficiency.

[0081] After the full-process recommendation system obtains the user association information of the target user, the first comprehensive user portrait of the user on multiple channels can be obtained based on the user association information of the target user. It should be noted that in this embodiment, the first comprehensive user portrait is obtained based on the comprehensiveness of multiple channels, rather than based on a single channel. Therefore, the first comprehensive user portrait can be obtained more comprehensively and accurately. The step of obtaining the first comprehensive user portrait of the target user on multiple channels includes:

[0082] Step S11, obtaining the target user's browsing records and order records on multiple channels;

[0083] Obtaining the target user's browsing records and order records on multiple channels. Obtaining the target user's browsing records and order records on multiple channels includes the following two methods:

[0084] Method 1: Requesting the target user's browsing records and order records on multiple channels from multiple channels that are connected to the recommendation system of the entire process at preset time intervals;

[0085] Method 2: Obtain the target user's browsing records and order records on multiple channels from databases corresponding to multiple channels embedded in the full-process recommendation system at preset time intervals.

[0086] Step S12, inputting the browsing records and order records into a preset recognition model, performing portrait recognition processing on the browsing records and order records based on the preset recognition model, and obtaining a first comprehensive user portrait;

[0087] The preset recognition model is obtained by iterative training based on a preset basic model.

[0088] It should be noted that, in this embodiment, a preset basic model is pre-stored, and based on the preset basic model and preset training data, the preset basic model is iteratively trained until the preset training completion condition is reached, such as the preset loss function converges, to obtain a preset recognition model, and the reading records and order records are input into the preset recognition model, and the reading records and order records are subjected to portrait recognition processing based on the preset recognition model to obtain a first comprehensive user portrait. The first comprehensive user portrait can be composed of labels such as Sagittarius, anime enthusiast, and milk tea enthusiast.

[0089] In this embodiment, when the target user's behavior operation instruction on the first target channel is detected, the process state of the target user on the first target channel is obtained, and the process state may include the path state of the target user on the first target channel, such as the target user passing through the channel public account, A enterprise, a1 business, entrance, page and other paths, or the target user passing through the channel mini program, enterprise, business, entrance, page and other paths. It should be noted that the process state may also include the behavior state of the target user on each path corresponding to the path state, wherein the behavior state may be based on the dialogue state such as dialogue communication, or text communication, etc., and the behavior state may also be a reading state such as an article reading state or a question preview state.

[0090] Step S20, determining a first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait;

[0091] In this embodiment, after obtaining the process status and the first comprehensive user portrait, a first response file in response to the target user is determined according to the process status, the first target channel and the first comprehensive user portrait. The first response file may be outputting the corresponding first type of article for the target user, or outputting the corresponding first type of question and the first type of answer for the target user, or a first type of interaction method determined for the target user, etc. Whether the corresponding first type of article is outputted for the target user, or the corresponding first type of question and the first type of answer is outputted for the target user, or a first type of interaction method determined for the target user, etc., depends on the specific behavior state of the current user.

[0092] The first response file includes dialogue reply content and guidance content, and the step of determining the first response file responding to the behavioral operation instruction according to the process status, the first target channel and the first comprehensive user portrait includes:

[0093] Step S21, if the process state is the first preset process state, determining a target dialogue mode for responding to the behavior operation instruction and a target dialogue reply content corresponding to the target dialogue mode according to the first target channel and the first comprehensive user portrait;

[0094] In this embodiment, if the process state is the first preset process state, the target dialogue mode for responding to the behavior operation instruction and the target dialogue reply content corresponding to the target dialogue mode are determined according to the first target channel and the first comprehensive user portrait, wherein the first preset process state refers to a dialogue state such as a text dialogue state or a voice dialogue state, and the first response file is the target dialogue mode and the target dialogue reply content corresponding to the target dialogue mode. Specifically, the target dialogue reply content corresponding to the target dialogue mode includes at least the following three types:

[0095] 1: If the target conversation mode is manual conversation, a conversation manager is assigned to the target user, wherein the conversation information of the user's history is collected to determine whether the target user has a designated conversation manager. If so, the designated conversation manager is assigned. If no designated conversation manager exists, the conversation manager is assigned based on the first comprehensive user portrait. The reply content of the target conversation is determined by the conversation manager.

[0096] 2. If the target conversation mode is non-human conversation, a conversation robot is assigned to the target user. After the conversation robot is determined, the initial conversation content of the target user is received. Based on the initial conversation content of the target user, the reply content corresponding to the initial conversation content is obtained from the preset configuration knowledge base, and the reply content is output.

[0097] Three: The target dialogue mode is to convert the machine dialogue into a manual dialogue. Specifically, the initial dialogue content of the target user is received, and the reply content corresponding to the initial dialogue content is obtained from the preset configuration knowledge base. Based on the initial dialogue content and the corresponding reply content, it is determined whether it needs to be converted to a manual dialogue. If it is determined that it needs to be converted to a manual dialogue based on the initial dialogue content and the corresponding reply content, the designated dialogue manager for the manual dialogue is determined based on the first comprehensive user portrait and process status of the target user, and the manual target dialogue reply content is determined based on the designated dialogue manager.

[0098] Based on the initial conversation content and the corresponding reply content, whether to switch to manual conversation can be determined in the following ways:

[0099] Method 1: If the initial conversation content and the corresponding reply content contain preset sensitive words such as order or purchase, it is determined that the conversation needs to be transferred to manual mode;

[0100] Method 2: If the initial conversation content and the corresponding reply content correspond to multiple preset times of back-and-forth reply cycles, it is determined that a manual conversation needs to be switched.

[0101] It should be noted that, in this embodiment, the conversion of machine dialogue into human dialogue can be determined by the target user or by the recommendation system of the entire process.

[0102] Step S22, if the process state is the second preset process state, determine the target recommended text content that responds to the behavioral operation instruction and the guidance content corresponding to the target recommended text content according to the first target channel and the first comprehensive user portrait.

[0103] If the process state is the second preset process state, the target recommended text content in response to the behavioral operation instruction and the guidance content corresponding to the target recommended text content are determined according to the first target channel and the first comprehensive user portrait, wherein the second preset process state is a reading state instead of a dialogue state, and after the reading state is determined, the target recommended text content in response to the behavioral operation instruction and the guidance content corresponding to the target recommended text content are determined. Specifically, for example, if the first comprehensive user portrait is a literary youth, then the target recommended text content of the behavioral operation instruction may be the target recommended text content of flower art, pottery, etc. in the current business; if the first comprehensive user portrait is an anime enthusiast, then the target recommended text content of the behavioral operation instruction may be the target recommended text content of the latest anime exhibition in the current business.

[0104] Step S30, obtaining first feedback information of the target user generating behavioral feedback for the first response file, and adjusting the first comprehensive user portrait according to the first feedback information to obtain a second comprehensive user portrait on multiple channels;

[0105] In this embodiment, after obtaining the first response file, it is determined whether the target user generates behavioral feedback in response to the first response file. If behavioral feedback is generated, first feedback information of the target user generating behavioral feedback in response to the first response file is obtained, and the first feedback information includes purchase behavior, click behavior, order behavior, conversation behavior, etc., and the first comprehensive user portrait is adjusted according to the first feedback information to obtain a second comprehensive user portrait on multiple channels. Specifically, the first feedback information, the target user's reading records and order records on multiple channels are input into a preset recognition model to obtain a second comprehensive user portrait on multiple channels.

[0106] Step S40: Perform targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on the multiple channels.

[0107] After obtaining the second comprehensive user portrait of the target user on multiple channels, targeted recommendations are made to the target user on the multiple channels based on the second comprehensive user portrait on multiple channels. Specifically, recommendations can be made to users via text messages, public accounts, H5 pages, WeChat or applications. The recommended content can be specific products, specific articles and the like, and the content can be in the form of pictures, text, pictures and text, voice, or video.

[0108] The step of performing targeted recommendations for the target user on the multiple channels based on the second comprehensive user portraits on the multiple channels includes:

[0109] Step S41, determining a recommended topic for the target user based on the second comprehensive user portrait on multiple channels;

[0110] Determining the recommended topics for the target user based on the second comprehensive user portraits on multiple channels includes:

[0111] Method 1: determining a recommended topic for the target user based on keyword information or information of interest carried in the second comprehensive user portrait on multiple channels;

[0112] For example, if the keyword information carried by the portrait is young people and animation, the recommended theme is the theme of the latest animation.

[0113] Method 2: Based on the second comprehensive user portrait on multiple channels, determine other target users whose portrait similarity with the second comprehensive user portrait is greater than a preset value, obtain the related topics of the other target users, and determine the recommended topic of the target user based on the related topics of the other target users.

[0114] Step S42, based on the recommended theme, selecting target content associated with the recommended theme from preset selected resources of various channels;

[0115] Based on the recommended theme, target content associated with the recommended theme is selected from the preset selected resources of each channel. Specifically, the target channel used by the target user needs to be obtained. If the target user uses or follows the A public account channel, the target content associated with the recommended theme is selected from the corresponding database (preset selected resources) in the A public account channel, such as selecting target content associated with animation from the corresponding database (preset selected resources) in the A public account channel, or selecting target content associated with literature and art from the corresponding database (preset selected resources) in the A public account channel.

[0116] Step S43: recommending the target content to the target user.

[0117] The target content is recommended to the target user. Specifically, the target content is sorted in the order of generation time, and the target content is recommended to the target user in sequence based on the order.

[0118] In this embodiment, it should be noted that the overall technical means of conducting targeted closed-loop recommendation based on the full-process recommendation file with feedback obtained from multiple channels is adopted to conduct targeted recommendation. The technical means of conducting targeted closed-loop recommendation based on the full-process recommendation file with feedback obtained from multiple channels is: when detecting the behavior operation instruction of the target user on the first target channel, obtaining the process state of the target user on the first target channel, and obtaining the first comprehensive user portrait of the target user on multiple channels;

[0119] Determine a first response file responding to the target user according to the process status, the first target channel and the first comprehensive user portrait;

[0120] Acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user profile according to the first feedback information to obtain a second comprehensive user profile on multiple channels;

[0121] According to the second comprehensive user portrait on multiple channels, targeted recommendations are made to the target user on the multiple channels.

[0122] This application obtains the process state of the target user on the first target channel and obtains the first comprehensive user portrait of the target user on multiple channels when detecting the target user's behavior operation instruction on the first target channel; determines the first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait; obtains the first feedback information of the target user generating behavior feedback for the first response file, and adjusts the first comprehensive user portrait according to the first feedback information to obtain the second comprehensive user portrait on multiple channels; and makes targeted recommendations for the target user on the multiple channels according to the second comprehensive user portrait on multiple channels. Compared with the prior art that studies or generates recommendation files for different recommendation channels separately for data recommendation, this application studies or generates recommendation files for multiple recommendation channels as a whole for data recommendation, thereby improving the recommendation efficiency, and in this application, based on the overall study or generation of recommendation files for data recommendation with higher richness of multiple channels, thereby improving the recommendation accuracy and recommendation effect, and overcoming the technical defect of low recommendation accuracy in the prior art.

[0123] Further, based on the first embodiment of the present application, another embodiment of the present application is provided. In this embodiment, referring to Figure 2 , when detecting a behavior operation instruction of a target user on a first target channel, the step of obtaining the process status of the target user on the first target channel comprises:

[0124] Step S13, when the target user's behavior operation instruction is detected, determining whether the target user is a logged-in user;

[0125] When a behavior operation instruction of a target user is detected, it is determined whether the target user is a logged-in user. Specifically, if information that can uniquely identify the target user can be obtained, it is determined that the target user is a logged-in user.

[0126] Step S14, if the target user is a logged-in user, determining a target user unique identifier shared by the target user in multiple channels;

[0127] Step S15: acquiring the process status of the target user on the first target channel according to the unique identifier of the target user.

[0128] If the target user is a logged-in user, the target user unique identifier shared by the target user in multiple channels is determined. That is, in this embodiment, the user identity information of multiple channels is unified into a string so that the process status of the target user in the entire full-process system can be queried based on the string segment, rather than querying the process status according to the login name of different channels. Therefore, the leakage of user information on different channels is avoided, and the recommendation security is improved. Moreover, since the query of the process status is unified into a string, the query convenience is improved. The unified user identity information is the unique identifier of the target user. Specifically, the unique identifier of the target user can be obtained through the user information mapping library.

[0129] In this embodiment, when the target user's behavioral operation instructions are detected, it is determined whether the target user is a logged-in user; if the target user is a logged-in user, the target user's unique identifier shared by the target user in multiple channels is determined; based on the target user's unique identifier, the process status of the target user on the first target channel is obtained to improve the recommendation security.

[0130] In another embodiment of the recommendation method of the whole process of the present application, if the process state is the second preset process state, the step of determining the target recommendation text content responding to the behavior operation instruction and the guidance content corresponding to the target recommendation text content according to the first target channel and the first comprehensive user portrait includes:

[0131] Step S221, if the process state is the second preset process state, determining the target recommended text content that responds to the behavior operation instruction according to the first target channel and the first comprehensive user portrait, and determining the display mode of the target recommended text content, the display mode including a graphic display mode, a pure text display mode and a pure picture display mode;

[0132] In this embodiment, if the process state is the second preset process state, the display type of the target user is determined according to the first target channel and the first comprehensive user portrait (including age portrait), and the display method of the target recommended text content is also determined according to the display type, and the display method includes a graphic display method, a pure text display method and a pure picture display method.

[0133] Step S222: determining the display layout of the target recommended text content according to the display mode, so as to determine the guidance content corresponding to the target recommended text content.

[0134] According to the display method, the corresponding graphics and texts to be displayed and the preset display layout of the graphics and texts to be displayed are obtained, and then the display layout of the target recommended text content is determined to further determine the guidance content corresponding to the target recommended text content. Specifically, for example, the current process of the target recommended text content, such as the page process, is obtained, and the operation process (guidance content) of the page process is determined to answer common user questions.

[0135] In this embodiment, if the process state is the second preset process state, the target recommended text content that responds to the behavioral operation instruction is determined according to the first target channel and the first comprehensive user portrait, and the display method of the target recommended text content is determined, and the display method includes a graphic display method, a pure text display method and a pure picture display method; according to the display method, the display layout of the target recommended text content is determined to determine the guidance content corresponding to the target recommended text content to improve the user experience. In this embodiment, guidance is also provided to achieve an improved user experience.

[0136] In another embodiment of the recommendation method of the whole process of the present application, after the step of determining the first response file responding to the behavioral operation instruction according to the process state, the first target channel and the first comprehensive user portrait, the method further includes:

[0137] Step A1, obtaining second feedback information fed back by the target user in response to the first response file; Step A2, obtaining the usage priority rankings of multiple channels of the target user according to the second feedback information, so as to perform channel circulation on the first target channel of the target user to obtain a second target channel;

[0138] Step A3: adjusting the first response file for the target user into a second response file according to the second target channel.

[0139] In this embodiment, second feedback information of the target user in response to the first response file is also obtained. According to the second feedback information, the first response file for the target user is adjusted to a second response file. Specifically, the second response file may be a targeted recommendation based only on the first comprehensive user portrait, or the second response file may be a recommendation based on the currently hottest search items.

[0140] In this embodiment, second feedback information of the target user in response to the first response file is obtained, and the usage priority ranking of multiple channels of the target user is obtained based on the second feedback information, so as to perform channel circulation on the first target channel of the target user and obtain the second target channel. That is, in this embodiment, all channels can circulate with each other, that is, when one channel cannot properly meet the needs of the target user, the usage priority ranking of multiple channels of the target user is obtained based on the second feedback information, so as to perform channel circulation on the first target channel of the target user and obtain the second target channel. Specifically, for example, when the second feedback information is detected on the official account, the usage priority ranking of multiple channels of the target user is obtained. If the target user uses the mini program the most, the mini program can be used as the second target channel.

[0141] According to the second target channel, the first response file for the target user is adjusted to the second response file, that is, in this embodiment, the customer is first retained through other channels, and then the first response file for the target user is adjusted to the second response file. For example, after the first response file, the most popular search item is selected to serve the target user.

[0142] In this embodiment, the second feedback information fed back by the target user for the first response file is obtained; according to the second feedback information, the priority rankings of the multiple channels are obtained to transfer the first target channel of the target user to obtain the second target channel; according to the second target channel, the first response file for the target user is adjusted to the second response file. In this embodiment, the response file is adjusted to avoid user disgust and improve user experience.

[0143] Reference Figure 3 , Figure 3 It is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present application.

[0144] like Figure 3As shown, the recommended device of the whole process may include: a processor 1001, such as a CPU, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to realize the connection and communication between the processor 1001 and the memory 1005. The memory 1005 can be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0145] Optionally, the full-process recommendation device may also include a rectangular user interface, a network interface, a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, etc. The rectangular user interface may include a display screen (Display), an input submodule such as a keyboard (Keyboard), and the optional rectangular user interface may also include a standard wired interface and a wireless interface. The network interface may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).

[0146] Those skilled in the art will understand that Figure 3 The recommended equipment structure of the entire process shown in does not constitute a limitation on the recommended equipment of the entire process, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0147] like Figure 3 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, and a full-process recommendation program. The operating system is a program that manages and controls the hardware and software resources of the full-process recommendation device, and supports the operation of the full-process recommendation program and other software and / or programs. The network communication module is used to realize the communication between the components inside the memory 1005, and the communication with other hardware and software in the full-process recommendation system.

[0148] exist Figure 3 In the full-process recommendation device shown, the processor 1001 is used to execute the full-process recommendation program stored in the memory 1005 to implement the steps of any of the full-process recommendation methods described above.

[0149] The specific implementation methods of the recommended equipment for the entire process of this application are basically the same as the embodiments of the recommended methods for the entire process described above, and will not be repeated here.

[0150] The present application also provides a full-process recommendation device, which includes:

[0151] A detection module, configured to, when detecting a behavior operation instruction of a target user on a first target channel, obtain a process state of the target user on the first target channel, and obtain a first comprehensive user portrait of the target user on multiple channels;

[0152] A determination module, configured to determine a first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait;

[0153] A first acquisition module is used to acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user portrait according to the first feedback information to obtain a second comprehensive user portrait on multiple channels;

[0154] A recommendation module is used to make targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on the multiple channels.

[0155] Optionally, the detection module includes:

[0156] A first determining unit, configured to determine whether the target user is a logged-in user when a behavior operation instruction of the target user is detected;

[0157] A first acquisition unit, configured to determine a target user unique identifier shared by the target user in multiple channels if the target user is a logged-in user;

[0158] The second acquisition unit is used to acquire the process status of the target user on the first target channel according to the unique identifier of the target user.

[0159] Optionally, the determining module includes:

[0160] A second determining unit is configured to determine, if the process state is the first preset process state, a target dialogue mode for responding to the behavioral operation instruction and a target dialogue reply content corresponding to the target dialogue mode according to the first target channel and the first comprehensive user portrait;

[0161] The third determination unit is used to determine the target recommended text content that responds to the behavioral operation instruction and the guidance content corresponding to the target recommended text content based on the first target channel and the first comprehensive user portrait if the process state is the second preset process state.

[0162] Optionally, the third determining unit includes:

[0163] A first determining subunit is used to determine, if the process state is a second preset process state, a target recommended text content that responds to the behavioral operation instruction according to the first target channel and the first comprehensive user portrait, and determine a display mode of the target recommended text content, wherein the display mode includes a graphic display mode, a pure text display mode, and a pure picture display mode;

[0164] The second determining subunit is used to determine the display layout of the target recommended text content according to the display mode, so as to determine the guidance content corresponding to the target recommended text content.

[0165] Optionally, the whole process recommendation device further includes:

[0166] A second acquisition module, configured to acquire second feedback information fed back by the target user in response to the first response file;

[0167] The third acquisition module is used to acquire the priority rankings of the multiple channels according to the second feedback information, so as to perform channel circulation on the first target channel of the target user to obtain a second target channel.

[0168] An adjustment module is used to adjust the first response file for the target user into a second response file according to the second target channel.

[0169] Optionally, the recommendation module includes:

[0170] a fourth determining unit, configured to determine a recommended topic for the target user based on the second comprehensive user portrait on multiple channels;

[0171] An association unit, configured to select target content associated with the recommended topic from preset selected resources of various channels based on the recommended topic;

[0172] A recommendation unit is used to recommend the target content to the target user.

[0173] Optionally, the detection module includes:

[0174] A third acquisition unit is used to acquire the target user's browsing records and order records on multiple channels;

[0175] an identification unit, configured to input the browsing records and order records into a preset identification model, and perform portrait identification processing on the browsing records and order records based on the preset identification model to obtain a first comprehensive user portrait;

[0176] The preset recognition model is obtained by iterative training based on a preset basic model.

[0177] The specific implementation of the recommendation device of the full process of the present application is basically the same as the various embodiments of the recommendation method of the full process described above, and will not be repeated here.

[0178] An embodiment of the present application provides a storage medium, and the storage medium stores one or more programs, and the one or more programs can also be executed by one or more processors to implement the steps of the full-process recommendation method described in any one of the above items.

[0179] The specific implementation method of the storage medium of the present application is basically the same as the embodiments of the recommended method of the above-mentioned full process, and will not be repeated here.

[0180] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0181] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0182] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0183] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A full-process recommendation method, characterized in that: The recommended approach for the full process includes: When a behavior operation instruction of a target user on a first target channel is detected, a process state of the target user on the first target channel is obtained, and a first comprehensive user portrait of the target user on multiple channels is obtained; Determine a first response file responding to the target user according to the process status, the first target channel and the first comprehensive user portrait; Acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user profile according to the first feedback information to obtain a second comprehensive user profile on multiple channels; According to the second comprehensive user portrait on multiple channels, targeted recommendations are made to the target user on the multiple channels.

2. The full-process recommendation method according to claim 1, characterized in that: The step of acquiring the process status of the target user on the first target channel when detecting the target user's behavior operation instruction on the first target channel includes: When detecting a behavior operation instruction of a target user, determining whether the target user is a logged-in user; If the target user is a logged-in user, determining a target user unique identifier shared by the target user in multiple channels; According to the target user unique identifier, the process status of the target user on the first target channel is obtained.

3. The full-process recommendation method according to claim 1, characterized in that: The first response file includes a dialogue reply content or a guide content, and the step of determining the first response file responding to the behavioral operation instruction according to the process state, the first target channel and the first comprehensive user portrait includes: If the process state is the first preset process state, determining a target dialogue mode for responding to the behavioral operation instruction and a target dialogue reply content corresponding to the target dialogue mode according to the first target channel and the first comprehensive user portrait; If the process state is the second preset process state, the target recommended text content that responds to the behavioral operation instruction and the guidance content corresponding to the target recommended text content are determined according to the first target channel and the first comprehensive user portrait.

4. The full-process recommendation method as claimed in claim 3, characterized in that: If the process state is the second preset process state, the step of determining the target recommended text content responding to the behavior operation instruction and the guidance content corresponding to the target recommended text content according to the first target channel and the first comprehensive user portrait includes: If the process state is the second preset process state, determining the target recommended text content that responds to the behavioral operation instruction according to the first target channel and the first comprehensive user portrait, and determining the display mode of the target recommended text content, the display mode including a graphic display mode, a pure text display mode and a pure picture display mode; According to the display method, the display layout of the target recommended text content is determined to determine the guidance content corresponding to the target recommended text content.

5. The full-process recommendation method according to claim 1, characterized in that: After the step of determining a first response file responding to the behavioral operation instruction according to the process state, the first target channel and the first comprehensive user portrait, the method further includes: Acquire second feedback information fed back by the target user in response to the first response file; According to the second feedback information, obtaining the usage priority rankings of the multiple channels of the target user, so as to perform channel circulation on the first target channel of the target user to obtain a second target channel; According to the second target channel, the first response file for the target user is adjusted to a second response file.

6. The full-process recommendation method according to claim 1, characterized in that: The step of performing targeted recommendations for the target user on the multiple channels according to the second comprehensive user portraits on the multiple channels includes: determining a recommended topic for the target user based on the second comprehensive user portrait on multiple channels; Based on the recommended topic, target content associated with the recommended topic is selected from preset selected resources of various channels; The target content is recommended to the target user.

7. The full-process recommendation method according to any one of claims 1 to 6, characterized in that: The step of obtaining a first comprehensive user portrait of the target user on multiple channels includes: Obtaining the target user's browsing records and order records on multiple channels; Inputting the browsing records and ordering records into a preset recognition model, and performing portrait recognition processing on the browsing records and ordering records based on the preset recognition model to obtain a first comprehensive user portrait; The preset recognition model is obtained by iterative training based on a preset basic model.

8. A full-process recommendation device, characterized in that: The recommended devices for the whole process include: A detection module, configured to, when detecting a behavior operation instruction of a target user on a first target channel, obtain a process state of the target user on the first target channel, and obtain a first comprehensive user portrait of the target user on multiple channels; A determination module, configured to determine a first response file responding to the target user according to the process state, the first target channel and the first comprehensive user portrait; A first acquisition module is used to acquire first feedback information of the target user generating behavioral feedback for the first response file, and adjust the first comprehensive user portrait according to the first feedback information to obtain a second comprehensive user portrait on multiple channels; A recommendation module is used to make targeted recommendations for the target user on the multiple channels based on the second comprehensive user portrait on the multiple channels.

9. A full-process recommendation device, characterized in that: The full-process recommendation device includes: a memory, a processor, and a program stored in the memory for implementing the full-process recommendation method. The memory is used to store a program for implementing the recommended method of the whole process; The processor is used to execute a program for implementing the recommendation method of the full process, so as to implement the steps of the recommendation method of the full process as claimed in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a program for implementing the full-process recommendation method, and the program for implementing the full-process recommendation method is executed by a processor to implement the steps of the full-process recommendation method as claimed in any one of claims 1 to 7.

Images