Intervention method and device applied to recommended scene, electronic equipment and vehicle

Abstract

The invention provides an intervention method and device applied to a recommended scene, electronic equipment and a vehicle. The intervention method comprises the following steps: receiving scene general information and artificial recommendation strategy information; constructing a recommendation scene according to the scene general information and determining a content library; analyzing the content in the content library based on a preset machine learning algorithm to obtain an analysis result; adjusting the analysis result according to the manual recommendation strategy information to obtain to-be-recommended data; when the manual recommendation strategy information comprises intervention information, further performing intervention operation on the to-be-recommended data to obtain final to-be-recommended data; and if the access request is received, displaying the to-be-recommended data of the recommendation scene. The to-be-recommended data is adjusted through the manual recommendation strategy, and the to-be-recommended data is intervened based on the intervention information, so that the final to-be-recommended data better fits the interest of the user, the recommendation accuracy is improved, and the recommended data in a specific scene is met. The recommendation system can quickly respond to market changes and business strategy adjustment through intervention information, and the content popularization efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to an intervention method, device, electronic equipment and vehicle applied to a recommendation scenario. Background Art

[0002] With the rapid development of the Internet and digital media, media recommendation systems have become a key bridge connecting users with massive amounts of information. Traditional recommendation algorithms rely on user behavior data and content features to recommend relevant content to users through automated means. Although this method can meet user needs to a certain extent, its flexibility and adaptability are still limited. These traditional algorithms are usually difficult to adjust in a timely manner to adapt to specific business goals or respond to real-time market dynamics, and there is a problem of inaccurate media recommendations.

[0003] Therefore, how to adjust the recommendation method in a timely manner to improve the accuracy of media resource recommendations is an urgent problem that needs to be solved. Summary of the invention

[0004] In view of this, embodiments of the present invention provide an intervention method, device, electronic device and vehicle applied to a recommendation scenario to solve the problem of inaccurate recommendations in a media resource recommendation algorithm.

[0005] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0006] A first aspect of the present invention discloses an intervention method applied to a recommendation scenario, the method comprising:

[0007] Receive the scenario general information and manual recommendation strategy information entered by the administrator in the new scenario interface of the operation platform;

[0008] Constructing a recommended scenario according to the scenario general information, and determining a content library corresponding to the recommended scenario, wherein the content library stores content collected in advance using a preset content collection method;

[0009] Analyze the content in the content library based on a preset machine learning algorithm to obtain an analysis result;

[0010] Adjusting the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario;

[0011] When the manual recommendation strategy information includes intervention information, performing an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended;

[0012] If an access request is received, determining a recommended scene accessed by the access request;

[0013] The data to be recommended that matches the recommendation scenario accessed by the access request is extracted and displayed.

[0014] Preferably, after extracting and displaying the data to be recommended that matches the recommendation scenario accessed by the access request, the method further includes:

[0015] Collecting user feedback data regarding the data to be recommended;

[0016] Evaluate the intervention effect according to the feedback data to obtain an evaluation result;

[0017] The preset machine learning algorithm is trained based on the evaluation result to obtain an optimized preset machine learning algorithm.

[0018] Preferably, the adjusting the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario includes:

[0019] According to the user portrait included in the manual recommendation strategy information, filtering out content matching the user portrait from the analysis results;

[0020] Based on the time period preference setting information included in the manual recommendation strategy information, a plurality of time period preference contents matching the time period preference setting information are screened out from the contents matching the user portrait;

[0021] The weights of the plurality of time period preference contents are adjusted according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

[0022] Preferably, when the manual recommendation strategy information includes intervention information, performing an intervention operation on the data to be recommended based on the intervention information to obtain the final data to be recommended includes:

[0023] When the manual recommendation strategy information includes intervention information, acquiring target data to be recommended whose type is consistent with the intervention type included in the intervention information from the data to be recommended;

[0024] Based on the intervention content and display order in the intervention information, an intervention operation is performed on the target data to be recommended to obtain final data to be recommended.

[0025] Preferably, after analyzing the content in the content library based on a preset machine learning algorithm and obtaining the analysis result, the method further includes:

[0026] When the manual recommendation strategy information includes intervention parameters, the parameters of the preset machine learning algorithm are adjusted according to the intervention parameters, and the step of analyzing the content in the content library based on the preset machine learning algorithm to obtain the analysis results is returned.

[0027] A second aspect of the present invention discloses an intervention device applied to a recommendation scenario, the device comprising:

[0028] A receiving unit, used to receive the scenario general information and manual recommendation strategy information entered by the administrator in the new scenario interface of the operation platform;

[0029] A construction unit, configured to construct a recommendation scene according to the general scene information, and determine a content library corresponding to the recommendation scene, wherein the content library stores content collected in advance using a preset content collection method;

[0030] An analysis unit, configured to analyze the content in the content library based on a preset machine learning algorithm to obtain an analysis result;

[0031] An adjustment unit, configured to adjust the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario;

[0032] an intervention unit, configured to, when the manual recommendation strategy information includes intervention information, perform an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended;

[0033] a determining unit, configured to, upon receiving an access request, determine a recommended scene accessed by the access request;

[0034] The display unit is used to extract and display the data to be recommended that matches the recommendation scenario accessed by the access request.

[0035] Preferably, the device further comprises:

[0036] A collecting unit, used to collect user feedback data on the data to be recommended;

[0037] An evaluation unit, used to evaluate the intervention effect according to the feedback data to obtain an evaluation result;

[0038] An optimization unit is used to train the preset machine learning algorithm based on the evaluation result to obtain an optimized preset machine learning algorithm.

[0039] Preferably, the adjustment unit comprises:

[0040] A first screening module, configured to screen out content matching the user portrait from the analysis results according to the user portrait included in the manual recommendation strategy information;

[0041] A second screening module is used to screen out a plurality of time period preference contents matching the time period preference setting information from the contents matching the user portrait based on the time period preference setting information included in the manual recommendation strategy information;

[0042] The adjustment module is used to adjust the weights of the plurality of time period preference contents according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

[0043] The third aspect of the present invention discloses an electronic device, comprising: a processor and a memory, wherein the processor and the memory are connected via a communication bus; wherein the processor is used to call and execute a program stored in the memory; and the memory is used to store a program, wherein the program is used to implement the intervention method applied to the recommendation scenario disclosed in the first aspect of the present invention.

[0044] A fourth aspect of the present invention discloses a vehicle, wherein the vehicle comprises the electronic device disclosed in the third aspect of the present invention.

[0045] Based on the above-mentioned embodiments of the present invention, an intervention method, device, electronic device and vehicle applied to recommendation scenarios are provided, receiving scenario general information and manual recommendation strategy information; constructing a recommendation scenario and determining a content library according to the scenario general information; analyzing the content in the content library based on a preset machine learning algorithm to obtain an analysis result; adjusting the analysis result according to the manual recommendation strategy information to obtain data to be recommended; when the manual recommendation strategy information includes intervention information, performing intervention operations on the data to be recommended based on the intervention information to obtain the final data to be recommended; if an access request is received, determining the recommended scenario accessed by the access request; and displaying the data to be recommended of the recommendation scenario. The data to be recommended is adjusted through the manual recommendation strategy, and the data to be recommended is intervened based on the intervention information, so that the final data to be recommended is more in line with the user's interests, the accuracy of media asset recommendations is improved, and the recommendation data in a specific scenario is satisfied. Furthermore, the intervention information enables the recommendation system to quickly respond to market changes and business strategy adjustments, thereby improving the efficiency of content promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention 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, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0047] Figure 1 A flowchart of an intervention method applied to a recommendation scenario provided by an embodiment of the present invention;

[0048] Figure 2 A schematic diagram of a new scene interface provided by an embodiment of the present invention;

[0049] Figure 3 A schematic diagram of the modules and data interaction of the intelligent recommendation platform provided in an embodiment of the present invention;

[0050] Figure 4 A structural block diagram of an intervention device applied to a recommendation scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising 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 "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0053] As can be seen from the background technology, although traditional recommendation algorithms can provide personalized content recommendations by analyzing users' historical behaviors, preferences, content features and contextual information, they are usually difficult to adjust in a timely manner to adapt to specific business goals or respond to real-time market dynamics. There is a problem of inaccurate media resource recommendations. In addition, traditional recommendation algorithms tend to recommend content that users are already aware of or popular, which may limit users' opportunities to access new or innovative products.

[0054] Therefore, an embodiment of the present invention provides an intervention method, device, electronic device and vehicle applied to a recommendation scenario, which receives general scenario information and manual recommendation strategy information; constructs a recommendation scenario and determines a content library based on the general scenario information; analyzes the content in the content library based on a preset machine learning algorithm to obtain an analysis result; adjusts the analysis result based on the manual recommendation strategy information to obtain data to be recommended; when the manual recommendation strategy information includes intervention information, performs intervention operations on the data to be recommended based on the intervention information to obtain the final data to be recommended; if an access request is received, displays the data to be recommended for the recommendation scenario. The data to be recommended is adjusted through the manual recommendation strategy, and the data to be recommended is intervened based on the intervention information, so that the final data to be recommended is more in line with the user's interests, the accuracy of media asset recommendations is improved, and the recommendation data in a specific scenario is satisfied. The intervention information enables the recommendation system to quickly respond to market changes and business strategy adjustments, thereby improving the efficiency of content promotion.

[0055] See also Figure 1 , showing a flow chart of an intervention method applied to a recommendation scenario provided by an embodiment of the present invention. It should be noted that the information (including but not limited to user device information, user personal information) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0056] The intervention method is applicable to the operating platform and includes:

[0057] Step S101: receiving general scenario information and manual recommendation strategy information entered by an administrator in a new scenario interface of an operation platform.

[0058] In the specific implementation of step S101, the general scene information (for example, Figure 2 The company information, display end information, brand information, scene name, etc. and manual recommendation strategy information are shown.

[0059] It is understood that the operating platform may be e.g. Figure 3 The intelligent recommendation platform shown. The general information of the scene also includes but is not limited to the scene ID. Among them, the manual recommendation strategy information is used to adjust and intervene in the analysis results of the preset machine learning algorithm. The specific adjustment and intervention process is detailed in the following content.

[0060] In some embodiments, the scene name is specifically a Chinese character that can highlight the characteristics of the scene, with a maximum of 10 characters. The scene ID is specifically an English abbreviation of the scene name, generally in pinyin English, with a maximum of 15 characters.

[0061] Step S102: construct a recommended scene according to the general scene information, and determine the content library corresponding to the recommended scene.

[0062] It should be noted that the content library stores content collected in advance using a preset content collection method. Figure 3 As shown, the material management module in the intelligent recommendation platform obtains content (such as user behavior data, content metadata, external environmental factors (such as holidays, hot events), and other information) from the content center / business party, and builds a content library based on the acquired content. The content library includes but is not limited to the types of content.

[0063] It is understandable that the content center / business party is used to obtain multiple material content packages and integrate content from different sources. The integrated content includes but is not limited to news headlines and news links.

[0064] In the specific implementation of step S102, a recommended scene is constructed based on general scene information (such as enterprise information, display end information, brand information, scene ID and scene name, etc.), and a content library corresponding to the recommended scene is determined based on the general scene information.

[0065] For example Figure 2 As shown in the figure: Based on the general information of the scene, a music recommendation scene is constructed and displayed in the car computer. Further, Figure 3 The material management module shown determines the content library corresponding to the current music recommendation scene, and the content library includes but is not limited to multiple audio data.

[0066] For example, a radio station recommendation scene displayed in the car is constructed based on the display terminal information, scene ID, scene name and other information in the scene general information. Figure 3 The material management module shown determines a content library corresponding to the current radio station recommendation scenario, and the content library includes but is not limited to multiple radio station related data.

[0067] Step S103: Analyze the content in the content library based on a preset machine learning algorithm to obtain analysis results.

[0068] In the specific implementation of step S103, a preset machine learning algorithm (eg Figure 3 The personalized algorithm, relevance algorithm, etc. shown in the figure analyzes the content in the content library, filters out the content that meets the current recommendation scenario, and obtains the analysis result.

[0069] In actual application, Figure 3 After constructing the recommendation scenario, the intelligent recommendation platform shown sends an analysis request to the algorithm platform based on the content in the content library, uses the preset machine learning algorithm in the algorithm platform to analyze the content in the content library, and the algorithm platform feeds back the analysis results to the recommendation scenario.

[0070] It is understandable that if Figure 3 As shown, the algorithm platform includes but is not limited to personalized algorithms, popular algorithms and relevance algorithms.

[0071] The personalization algorithm is used to analyze the content in the content library based on the pre-collected behavior data of users in some recommendation scenarios related to the current recommendation scenario, and obtain a personalized recommendation list (ie, the analysis result of the present invention).

[0072] For example, if the current recommendation scenario is a music recommendation scenario, the user's behavior data in other music recommendation scenarios is collected based on the presets to analyze the content in the content library and obtain a personalized recommendation list.

[0073] In the hot algorithm, based on the principles of "news that have been returned will not be returned (no period limit)" and "if a hot word has matched the content, the hot word will no longer match", the hot words are deduplicated, and then the hot list (i.e. the analysis result of the present invention) is determined by combining the heat value and timeliness.

[0074] In the correlation algorithm, the similarity vector value between item 1 and item 2 is determined, and based on the similarity vector value, the similarity is calculated to determine the analysis result.

[0075] It should be noted that if Figure 3 As shown in the figure, the intelligent recommendation platform includes a strategy library module. In the strategy library module, administrators can configure manual recommendation strategies and preset machine learning algorithms on the algorithm list page. In the strategy library module, basic information of the algorithm can also be displayed and general rules and algorithms (without code and parameters) can be uploaded.

[0076] In some specific embodiments, when the manual recommendation strategy information includes intervention parameters, the parameters of the preset machine learning algorithm are adjusted according to the intervention parameters, and the process returns to step S103.

[0077] It is understandable that after the administrator previews the analysis results in the operation platform, he can enter the intervention parameters for the preset machine learning algorithm in the operation platform, and use the intervention parameters to adjust the parameters of the preset machine learning algorithm to achieve the purpose of quickly adjusting the analysis results, thereby improving the recommendation accuracy of the adjusted preset machine learning algorithm.

[0078] Step S104: adjusting the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

[0079] It should be noted that the manual recommendation strategy information includes but is not limited to user portraits, time period preference setting information and weight information.

[0080] In the process of implementing step S104, the analysis results are adjusted according to the manual recommendation strategy information input by the administrator in the operation platform, specifically, the analysis results are adjusted according to the user portrait, time period preference setting information and weight information, so as to obtain the data to be recommended for the recommendation scenario.

[0081] The specific implementation process of the adjustment is as follows (Process A1-Process A3):

[0082] Process A1: Based on the user profile contained in the manual recommendation strategy information, select the content matching the user profile from the analysis results.

[0083] It should be noted that user portraits are idealized user models created based on real user data, which represent the characteristics, needs, behaviors and goals of a specific user group. User portraits include but are not limited to basic user information (such as age, gender, occupation), psychological characteristics (such as values, lifestyle), behavioral habits (such as shopping habits, online behavior), etc.

[0084] In the specific implementation process A1, the user portrait is extracted from the manual recommendation strategy information, and the content that meets the current recommendation scenario (i.e., the analysis result) is analyzed to extract the key information and features of the content; the key information and features of the content are matched with the user portrait to obtain the content that matches the user portrait.

[0085] Among them, content analysis includes but is not limited to text analysis (such as keyword extraction, topic modeling), image analysis (such as image recognition, object detection) and video analysis (such as scene recognition, action recognition), etc.

[0086] Process A2: Based on the time period preference setting information included in the manual recommendation strategy information, multiple time period preference contents matching the time period preference setting information are screened out from the contents matching the user portrait.

[0087] It should be noted that the time period preference setting information generally refers to the user's preference or behavior pattern in a specific time period.

[0088] In the specific implementation process A2, based on the time period preference setting information included in the manual recommendation strategy information, the content matching the user portrait is classified into time periods to obtain a plurality of time period preference contents matching the time period preference setting information.

[0089] It is understandable that filtering out multiple time period preference contents matching the time period preference setting information according to the time period preference setting information helps to more accurately meet the needs and interests of the user in a specific time period.

[0090] Process A3: The weights of the preferred contents in multiple time periods are adjusted according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended in the recommendation scenario.

[0091] In the specific implementation process A3, the weight of the preferred content of each time period is adjusted according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

[0092] It is understandable that adjusting the preferred content of multiple time periods according to the weight information and obtaining the content to be recommended in the recommendation scenario can help improve the accuracy and relevance of the recommendation; through weight control, it can ensure that high-quality and high-value content is recommended at the right time, thereby maximizing the influence of the content, increasing user stickiness and promoting content diversity.

[0093] For example Figure 2 The content shown, obtains the corresponding user portrait according to the user group name "1" in the manual recommendation strategy information, and obtains the corresponding content according to the content group selection "1", so as to filter out multiple time period preference contents that match the time period preference setting information from the content.

[0094] It is important to note that when Figure 2 When the “Is it effective” button in the manual intervention column is triggered, it indicates to perform intervention operation on the recommended data according to the intervention information in the manual intervention column. The specific implementation process is as shown in step S105.

[0095] Step S105: When the manual recommendation strategy information includes intervention information, an intervention operation is performed on the data to be recommended based on the intervention information to obtain final data to be recommended.

[0096] It is understandable that intervention information is usually used to make some form of modification or adjustment to the recommended data in order to affect the final result.

[0097] In the specific implementation of step S105, when the manual recommendation strategy information includes intervention information, the intervention target in the data to be recommended is determined according to the intervention information, and the intervention target is intervened to obtain the final data to be recommended.

[0098] In some embodiments, the specific intervention process is as follows (process B1 and process B2):

[0099] Process B1: When the manual recommendation strategy information includes intervention information, target data to be recommended whose type is consistent with the intervention type included in the intervention information is obtained from the data to be recommended.

[0100] For example: When the intervention type included in the intervention information in the manual recommendation strategy information is of the commodity type, all the recommended data of the commodity type in the data to be recommended are determined as the target recommended data.

[0101] Process B2: Based on the intervention content and display order in the intervention information, perform an intervention operation on the target recommended data to obtain the final recommended data.

[0102] That is to say, perform an intervention operation on the target recommended data according to the intervention content and display order in the intervention information to ensure that these target recommended data have a higher priority in the final recommended list, thereby improving the diversity of the data to be recommended and enhancing the user experience.

[0103] It can be understood that, for example Figure 2 in the intervention information includes an intervention type and intervention content. If the intervention type is "post", then obtain the target recommended data of the post type from the data to be recommended, and perform an intervention operation on the target recommended data based on the intervention content and display order to obtain the final recommended data.

[0104] It should be noted that the presentation form of the data to be recommended can be in list form. In step S105, directly adjust the list according to the manual recommendation strategy information. For example: if the manual recommendation strategy information contains specific recall content (such as "You are the one every year") and a specific order of the specific recall content (such as 1), then adjust "You are the one every year" to the first position in the list. Another example Figure 3 In the scene management module shown, the administrator sets the number of recommended items per single recommendation. Based on this, adjust the number of recommended items per single recommendation in the list to obtain the final recommended data for the current recommended scene.

[0105] Step S106: When receiving an access request, determine the recommended scene accessed by the access request.

[0106] In the specific process of implementing step S106, when receiving a user's access request, determine the recommended scene accessed by the access request according to the recommended scene information carried in the access request.

[0107] Step S107: Extract the recommended data that matches the recommended scene accessed by the access request and display it.

[0108] In the specific process of implementing step S107, obtain the recommended data that matches the recommended scene accessed by the access request, and display the recommended data to the user through the interaction interface.

[0109] In practical applications, the intervention method for the current media asset recommendation scene can be verified through experiments. Specifically, for example Figure 3The AB experiment module shown. Through the AB experiment module, create the experiment name, experiment recommendation scenario, experiment time interval, experiment type and experiment group information, and display the experiment recommendation data to the user. Through the data report module, collect user feedback data on the experiment recommendation data, generate data reports corresponding to different scenarios, indicators and time, and display them to the administrator through the interactive interface.

[0110] It should be noted that the AB experiment module and the data report module can obtain data or store data from the data center respectively. The data center includes but is not limited to the main wide table data and user behavior data. The data center also supports report data calculation and AB experiment data calculation.

[0111] In some specific embodiments, user feedback data (or business indicators) for the data to be recommended are collected; the intervention effect is evaluated based on the feedback data (or business indicators) to obtain an evaluation result; and a preset machine learning algorithm is trained based on the evaluation result to obtain an optimized preset machine learning algorithm.

[0112] Understandably, feedback data can provide direct signals indicating whether the recommendations are in line with the user's interests and needs. By evaluating the effectiveness of interventions, recommendations can be better personalized and user satisfaction can be improved.

[0113] It is important to note that, for example Figure 2 Schematic diagram of the new scenario interface shown. The general scenario information entered by the administrator in the new scenario interface can be used to create a new recommendation scenario. When it is detected that the administrator triggers the "algorithm strategy" button in the new scenario interface and clicks the "finish" button, the recommendation scenario is constructed based on the general scenario information, and the data to be recommended for the current recommendation scenario is determined through the pre-set algorithm strategy (including but not limited to personalized recommendation, related recommendation, and popular recommendation).

[0114] When it is detected that the administrator triggers the "Manual Strategy" button in the new scenario interface and clicks the "Finish" button, the analysis results output by the preset machine learning algorithm are adjusted according to the "User Group Name" determined by the administrator in the "User Group Selection" of the "Manual Recommendation Strategy" column, and the "Content Group Name" determined in the "Content Group Selection". Further, when it is detected that the "Is it effective" button in the manual intervention column is triggered, the recommended data is directly intervened according to the intervention type, intervention content and display order to obtain the final recommended data.

[0115] In an embodiment of the present invention, general scenario information and manual recommendation strategy information entered by an administrator are received; a recommendation scenario is constructed and a content library is determined based on the general scenario information; the data to be recommended is adjusted through the manual recommendation strategy, and the data to be recommended is intervened based on the intervention information in the manual recommendation strategy information, so that the final data to be recommended is more in line with user interests and the accuracy of media asset recommendations is improved. The present invention enables administrators to adjust recommendation strategies according to actual needs through an intelligent manual adjustment mechanism, which helps to meet recommendation needs in specific scenarios, reduces the complexity and time cost of manual intervention, and achieves the purpose of efficient operation. Furthermore, intervention information enables the recommendation system to quickly respond to market changes and business strategy adjustments, thereby improving content promotion efficiency.

[0116] Corresponding to an intervention method applied to a recommendation scenario provided in the above embodiment of the present invention, see Figure 4 , showing a schematic diagram of the structure of an intervention device applied to a recommendation scenario provided by an embodiment of the present invention.

[0117] The intervention device includes: a receiving unit 401 , a construction unit 402 , an analysis unit 403 , an adjustment unit 404 , an intervention unit 405 , a determination unit 406 and a display unit 407 .

[0118] The receiving unit 401 is used to receive the scenario general information and manual recommendation strategy information entered by the administrator in the new scenario interface of the operation platform.

[0119] The construction unit 402 is used to construct a recommended scene according to the general scene information, and determine a content library corresponding to the recommended scene, wherein the content library stores content collected in advance using a preset content collection method.

[0120] The analysis unit 403 is used to analyze the content in the content library based on a preset machine learning algorithm to obtain an analysis result.

[0121] The adjustment unit 404 is used to adjust the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

[0122] The intervention unit 405 is used to perform an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended when the manual recommendation strategy information includes intervention information.

[0123] The determination unit 406 is configured to determine the recommended scene accessed by the access request when an access request is received.

[0124] The display unit 407 is used to extract and display the to-be-recommended data that matches the recommendation scenario accessed by the access request.

[0125] In an embodiment of the present invention, general scenario information and manual recommendation strategy information entered by an administrator are received; a recommendation scenario is constructed and a content library is determined based on the general scenario information; the data to be recommended is adjusted through the manual recommendation strategy, and the data to be recommended is intervened based on the intervention information in the manual recommendation strategy information, so that the final data to be recommended is more in line with user interests and the accuracy of media asset recommendations is improved. The present invention enables administrators to adjust recommendation strategies according to actual needs through an intelligent manual adjustment mechanism, which helps to meet recommendation needs in specific scenarios, reduces the complexity and time cost of manual intervention, and achieves the purpose of efficient operation. Furthermore, intervention information enables the recommendation system to quickly respond to market changes and business strategy adjustments, thereby improving content promotion efficiency.

[0126] Combination Figure 4 According to the content shown, the intervention device also includes: a collection unit, an evaluation unit and an optimization unit.

[0127] The collecting unit is used to collect user feedback data on the data to be recommended.

[0128] The evaluation unit is used to evaluate the intervention effect based on the feedback data and obtain the evaluation result.

[0129] The optimization unit is used to train the preset machine learning algorithm based on the evaluation result to obtain an optimized preset machine learning algorithm.

[0130] Combination Figure 4 As shown in the content, the adjustment unit 404 includes: a first screening module, a second screening module and an adjustment module.

[0131] The first screening module is used to screen out content matching the user portrait from the analysis results according to the user portrait included in the manual recommendation strategy information.

[0132] The second screening module is used to screen out a plurality of time period preference contents matching the time period preference setting information from the contents matching the user portrait based on the time period preference setting information included in the manual recommendation strategy information.

[0133] The adjustment module is used to adjust the weights of the preferred contents of multiple time periods according to the weight information contained in the manual recommendation strategy information to obtain the data to be recommended in the recommendation scenario.

[0134] Combination Figure 4 The content shown, the intervention unit 405, includes: an acquisition module and an intervention module.

[0135] The acquisition module is used to acquire target data to be recommended whose type is consistent with the intervention type included in the intervention information from the data to be recommended when the manual recommendation strategy information includes intervention information.

[0136] The intervention module is used to perform intervention operations on the target data to be recommended based on the intervention content and display order in the intervention information to obtain the final data to be recommended.

[0137] Combination Figure 4 As shown in the content, the intervention device also includes: a parameter adjustment unit, which is used to adjust the parameters of the preset machine learning algorithm according to the intervention parameters when the manual recommendation strategy information includes intervention parameters, and return to the execution analysis unit 403.

[0138] An embodiment of the present invention also provides an electronic device, comprising: a processor and a memory, the processor and the memory being connected via a communication bus; wherein the processor is used to call and execute a program stored in the memory; the memory is used to store the program, and the program is used to implement an intervention method applied to a recommendation scenario as provided in an embodiment of the present invention.

[0139] It should be noted that the electronic equipment in this article can be a server, PC, PAD, mobile phone, ECU (Electronic Control Unit), VCU (Vehicle Control Unit), MCU (MicroController Unit), HCU (Hybrid Control Unit), etc.

[0140] An embodiment of the present invention further provides a vehicle, which includes the electronic device mentioned above.

[0141] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0142] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0143] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intervention method applied to a recommendation scenario, characterized in that: The method comprises: Receive the scenario general information and manual recommendation strategy information entered by the administrator in the new scenario interface of the operation platform; Constructing a recommended scenario according to the scenario general information, and determining a content library corresponding to the recommended scenario, wherein the content library stores content collected in advance using a preset content collection method; Analyze the content in the content library based on a preset machine learning algorithm to obtain an analysis result; Adjusting the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario; When the manual recommendation strategy information includes intervention information, performing an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended; If an access request is received, determining a recommended scene accessed by the access request; The data to be recommended that matches the recommendation scenario accessed by the access request is extracted and displayed.

2. The method according to claim 1, characterized in that After extracting and displaying the data to be recommended that matches the recommendation scenario accessed by the access request, the method further includes: Collecting user feedback data regarding the data to be recommended; Evaluate the intervention effect according to the feedback data to obtain an evaluation result; The preset machine learning algorithm is trained based on the evaluation result to obtain an optimized preset machine learning algorithm.

3. The method according to claim 1, characterized in that: The step of adjusting the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario includes: According to the user portrait included in the manual recommendation strategy information, filtering out content matching the user portrait from the analysis results; Based on the time period preference setting information included in the manual recommendation strategy information, a plurality of time period preference contents matching the time period preference setting information are screened out from the contents matching the user portrait; The weights of the plurality of time period preference contents are adjusted according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

4. The method according to claim 1, characterized in that: When the manual recommendation strategy information includes intervention information, performing an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended includes: When the manual recommendation strategy information includes intervention information, acquiring target data to be recommended whose type is consistent with the intervention type included in the intervention information from the data to be recommended; Based on the intervention content and display order in the intervention information, an intervention operation is performed on the target data to be recommended to obtain final data to be recommended.

5. The method according to claim 1, characterized in that: After analyzing the content in the content library based on a preset machine learning algorithm and obtaining the analysis result, the method further includes: When the manual recommendation strategy information includes intervention parameters, the parameters of the preset machine learning algorithm are adjusted according to the intervention parameters, and the step of analyzing the content in the content library based on the preset machine learning algorithm to obtain the analysis results is returned.

6. An intervention device applied to a recommendation scenario, characterized in that: The device comprises: A receiving unit, used to receive the scenario general information and manual recommendation strategy information entered by the administrator in the new scenario interface of the operation platform; A construction unit, configured to construct a recommendation scene according to the general scene information, and determine a content library corresponding to the recommendation scene, wherein the content library stores content collected in advance using a preset content collection method; An analysis unit, configured to analyze the content in the content library based on a preset machine learning algorithm to obtain an analysis result; An adjustment unit, configured to adjust the analysis result according to the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario; an intervention unit, configured to, when the manual recommendation strategy information includes intervention information, perform an intervention operation on the data to be recommended based on the intervention information to obtain final data to be recommended; a determination unit, configured to, upon receiving an access request, determine a recommended scene accessed by the access request; The display unit is used to extract and display the data to be recommended that matches the recommendation scenario accessed by the access request.

7. The device according to claim 6, characterized in that The device also includes: A collecting unit, used to collect user feedback data on the data to be recommended; An evaluation unit, used to evaluate the intervention effect according to the feedback data to obtain an evaluation result; An optimization unit is used to train the preset machine learning algorithm based on the evaluation result to obtain an optimized preset machine learning algorithm.

8. The device according to claim 6, characterized in that The adjustment unit comprises: A first screening module, configured to screen out content matching the user portrait from the analysis results according to the user portrait included in the manual recommendation strategy information; A second screening module is used to screen out a plurality of time period preference contents matching the time period preference setting information from the contents matching the user portrait based on the time period preference setting information included in the manual recommendation strategy information; The adjustment module is used to adjust the weights of the plurality of time period preference contents according to the weight information included in the manual recommendation strategy information to obtain the data to be recommended for the recommendation scenario.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the processor and the memory are connected via a communication bus; wherein the processor is used to call and execute a program stored in the memory; The memory is used to store a program, and the program is used to implement the intervention method applied to the recommendation scenario as described in any one of claims 1 to 5.

10. A vehicle, characterized in that: The vehicle includes the electronic device according to claim 9.

Images