Internet-based multimedia advertisement intelligent pushing method and system
By analyzing and sorting ad push methods, the problem of overloading advertising information caused by insufficient data interconnection between APP platforms is solved, the advertising push is optimized, and the user experience and content processing efficiency is improved.
Patent Information
- Application Number
- CN202411119689.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-22
AI Technical Summary
Due to the lack of data interconnection between major APP platforms, the user's mobile client receives a large number of repeated advertisement pushes, causing advertising information to be overloaded, and the advertisement is no longer related to the user's current needs or interests, which affects the user's content processing efficiency and Experience.
By obtaining the historical behavior and access information of the target object client, calculating recommended values, performing data analysis and preprocessing, extracting target keywords, sorting and attenuating advertising messages, avoiding the push of duplicate and low-value messages, and optimizing the time node and order of advertising push .
It reduces the pressure of information overload, improves the efficiency of user content browsing, improves the actual experience of users, and avoids pushing irrelevant advertisements when users are busy.
Smart Images

Figure CN119006066A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of advertisement recommendation, and in particular to an Internet-based multimedia advertisement intelligent push method and system. Background Art
[0002] With the development of computer network technology, message overload has emerged. How to efficiently obtain messages of interest from massive amounts of content has become an important demand. It is also based on this that more and more applications and platforms have introduced personalized content recommendation systems to solve the above information overload problem, and achieve accurate content recommendation through personalized matching of content and users. The various network service providers that provide applications will push various application content to the mobile terminals used by users through the network for users to view; the device information detected by the mobile terminal will also be pushed to the front end of the mobile terminal for users to view.
[0003] Existing advertising push methods mainly push advertisements by obtaining the user's corresponding browsing history, consumption history and other information. In real life, when users search for information or purchase goods online, due to the lack of data interconnection between major APP platforms, users' mobile clients often receive a large number of repeated advertising pushes, causing advertising information overload, especially when the advertisements are repeated and no longer relevant to the user's current needs or interests. If users view and process these contents, it will not only waste a lot of time and affect the user's content processing efficiency, but also cause users to feel bored and have a poor actual user experience. Summary of the invention
[0004] The purpose of the present invention is to provide a method and system for intelligently pushing multimedia advertisements based on the Internet to solve the following technical problems:
[0005] Due to the lack of data interconnection between major APP platforms, users' mobile clients often receive a large number of repeated advertising pushes, causing advertising information overload. Especially when the advertisements are repeated and no longer relevant to the user's current needs or interests, if the user views and processes these contents, it will not only waste a lot of time and affect the user's content processing efficiency, but also cause the user to feel bored and the actual user experience is very poor.
[0006] The purpose of the present invention can be achieved through the following technical solutions:
[0007] A multimedia advertisement intelligent push method based on the Internet comprises the following steps:
[0008] S1, obtaining the historical behavior of the target object client, the historical behavior includes the generation time of all application recommendation messages to the target object and the target object's access information; the target object's access information includes the target object's access time, access duration and purchase data for any application;
[0009] S2, select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; perform data analysis on the time period corresponding to the recommendation value to obtain the time node, mark the time node and the corresponding recommendation value as the recommendation parameter, and obtain the recommendation parameters of all applications;
[0010] S3, extracting text data from the purchase data and preprocessing it to obtain target keywords, extracting all the messages to be recommended in the target object client storage module and searching for the messages to be recommended according to the target keywords, marking the messages to be recommended containing the target keywords as target messages; sorting the target messages according to the recommendation parameters, generating a target message queue, and attenuating the target message recommendation values in turn according to the target message queue, and the attenuation factor is (1 / 2)^(n-1), where n is the number of target messages;
[0011] S4, determining all the messages to be recommended at any time node according to the recommendation parameters, and sorting the messages to be recommended from large to small according to the recommendation values and generating a recommended message sequence, and selecting the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
[0012] As a further solution of the present invention: in S2, the specific process of recommending parameters is:
[0013] A unit time is preset and divided into several equal time periods, any application is selected and the recommended messages are classified according to the time period to which the generation time of the recommended messages of the application belongs, so as to obtain several sub-datasets; the first access time T after the generation time t of any recommended message is obtained, and the time difference between the generation time t and the access time T is calculated to obtain the response time, and the recommended value W of the sub-dataset is calculated according to the response time, and the time periods in which all the recommended values are less than the preset threshold value within the first m unit times of the current time are obtained, where m is a positive integer, and when the recommended value of any time period is less than the preset threshold value for more than n+β / 2×n times, where β is a preset coefficient, the time period is marked as a forbidden push node, and the remaining time period after the unit time of the target object is excluding the forbidden push node is marked as a time node, and the time node and the recommended value corresponding to the time node are marked as a user parameter.
[0014] As a further solution of the present invention, the sub-datasets are also preprocessed, and the specific process of the preprocessing is as follows:
[0015] Get any access time T and access duration H, construct the access time interval [T, T+H], and delete the recommended messages within the access time interval.
[0016] As a further solution of the present invention: the calculation process of the recommended value is:
[0017] wi=1 / e -(t-T) ;
[0018]
[0019] Among them, wi is the recommendation score of any recommended message, t is the response time, T is the first request time, n is the number of preprocessed recommended messages in any sub-dataset, and W is the recommendation value.
[0020] As a further solution of the present invention: in S3, the specific process of sorting is:
[0021] Obtain the recommended parameters of the application corresponding to all target messages in turn, determine the time nodes of all target messages according to the recommended parameters, select any time node and obtain the target message corresponding to the time node, sort them in descending order of the recommended value to obtain the target message group Xi, where Xi∈[q1, q2, ..., qs], s represents the number of target messages in the time node, i is the number of the time node, and then obtain the target message group corresponding to all time nodes and mark them as X1, X2, ..., Xz, z is the number of time nodes, and then sort the target messages according to the time node numbers.
[0022] As a further solution of the present invention: S3 further includes:
[0023] If there is a target message group corresponding to any time node that contains multiple target messages from the same application, then the arrangement order of the multiple targets in the target message group is obtained and the target message recommendation values are attenuated in turn according to the arrangement order. The attenuation factor is (1 / 2)^(n'-1), where n' is the number of target messages of the same application.
[0024] As a further solution of the present invention: S4 also includes successively attenuating the recommendation values of the recommendation messages of the same application in the recommendation queue, and the attenuation factor is (1 / 2)^(n"-1), where n" is the n"th content recommended by the same application.
[0025] A multimedia advertisement intelligent push system based on the Internet, comprising:
[0026] Data acquisition module: used to acquire the historical behavior of the target object client, the historical behavior includes the generation time of all applications recommending messages to the target object and the target object's access information; the target object's access information includes the target object's access time to any application, access duration and purchase data;
[0027] Data analysis module: used to select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; and perform data analysis on the time period corresponding to the recommendation value to obtain the time node, mark the time node and the corresponding recommendation value as the recommendation parameter, and obtain the recommendation parameters of all applications;
[0028] Data optimization module: used to extract text data from the purchase data and perform preprocessing to obtain target keywords, extract all to-be-recommended messages in the target object client storage module and search for the to-be-recommended messages according to the target keywords, and mark the to-be-recommended messages containing the target keywords as target messages; sort the target messages according to the recommendation parameters, generate a target message queue, and sequentially decay the target message recommendation values according to the target message queue, and the decay factor is (1 / 2)^(n-1), where n is the number of target messages;
[0029] Result generation module: used to determine all the messages to be recommended at any time node according to the recommendation parameters, sort the messages to be recommended from large to small according to the recommendation value and generate a recommended message sequence, and select the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
[0030] Beneficial effects of the present invention:
[0031] The present invention obtains the historical behavior of the target client and classifies the recommended message data to obtain several sub-processing data sets, calculates the time difference between the click time of any recommended message and the recommended time to obtain the response time, calculates the recommendation score of any sub-processing data set according to the response time and the click behavior of the recommended message, and analyzes the time period corresponding to the recommended value to obtain the time node, and marks the time node and the recommended value corresponding to the time node as a user parameter to realize the processing of the time node of the user advertising information push, so as to avoid advertising push when the user is working or busy, resulting in that the recommended application content cannot be viewed in time by the user even if it meets the content that the current user is interested in, resulting in a very poor user experience;
[0032] The purchase data is extracted and preprocessed to obtain target keywords, which are used to retrieve and calibrate the recommended messages, and the recommended messages are further divided into target messages and remaining recommended messages. The target messages are sorted and the recommended value is decayed according to the recommendation parameters. It can be understood that the existing advertising push method mainly pushes advertisements by obtaining the user's corresponding browsing records, consumption records and other information. When the user places an order for a certain product, due to the lack of data interconnection between major APP platforms and the corresponding advertisement recommendation based on the user's search record, a large number of repeated recommendation messages that are less in line with the user's current needs appear, which makes the user feel bored and the actual user experience is very poor. Therefore, the present invention sorts the target messages and decays the recommended value according to the recommendation parameters, and sorts all the recommended messages from large to small according to the recommended value, and selects the top N recommended messages for push, thereby avoiding pushing too many repeated or low-value messages to users, reducing the pressure of information overload, improving the efficiency of user content browsing, and enhancing the actual user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] The present invention will be further described below in conjunction with the accompanying drawings.
[0034] Figure 1 The present invention is a flowchart of a multimedia advertisement intelligent push method based on the Internet. DETAILED DESCRIPTION
[0035] 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.
[0036] See also Figure 1 As shown, the present invention is a multimedia advertisement intelligent push method based on the Internet, comprising the following steps:
[0037] S1, obtaining the historical behavior of the target object client, the historical behavior includes the generation time of all application recommendation messages to the target object and the target object's access information; the target object's access information includes the target object's access time, access duration and purchase data for any application;
[0038] S2, select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; perform data analysis on the time period corresponding to the recommendation value to obtain the time node, mark the time node and the corresponding recommendation value as the recommendation parameter, and obtain the recommendation parameters of all applications;
[0039] S3, extracting text data from the purchase data and preprocessing it to obtain target keywords, extracting all the messages to be recommended in the target object client storage module and searching for the messages to be recommended according to the target keywords, marking the messages to be recommended containing the target keywords as target messages; sorting the target messages according to the recommendation parameters, generating a target message queue, and attenuating the target message recommendation values in turn according to the target message queue, and the attenuation factor is (1 / 2)^(n-1), where n is the number of target messages;
[0040] S4, determining all the messages to be recommended at any time node according to the recommendation parameters, and sorting the messages to be recommended from large to small according to the recommendation values and generating a recommended message sequence, and selecting the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
[0041] The present invention obtains the historical behavior of the target object client and classifies the recommended message data to obtain several sub-processing data sets, calculates the time difference between the click time of any recommended message and the recommended time to obtain the response time, calculates the recommendation score of any sub-processing data set according to the response time and the click behavior of the recommended message, and analyzes the time period corresponding to the recommended value to obtain the time node, and marks the time node and the recommended value corresponding to the time node as a user parameter, so as to realize the processing of the time node of the user advertising information push, avoid advertising push when the user is working or busy, resulting in the user not being able to view the recommended application content in time even if it meets the content that the current user is interested in, resulting in a very poor user experience; the purchase data is extracted and preprocessed to obtain the target keyword, and the target keyword is used to retrieve and calibrate the recommended message, and the recommended message is further divided into The target messages and the remaining messages to be recommended are sorted and the recommendation values are decayed according to the recommendation parameters. It can be understood that the existing advertising push method mainly pushes advertisements by obtaining the user's corresponding browsing records, consumption records and other information. When the user places an order for a certain product, due to the lack of data interconnection between the major APP platforms and the corresponding advertisement recommendations based on the user's search records, a large number of repeated recommendation messages that are less relevant to the user's current needs appear, which makes the user feel bored and the actual user experience is very poor. Therefore, the present invention sorts the target messages and decays the recommendation values according to the recommendation parameters, and sorts all the messages to be recommended from large to small according to the recommendation values, and selects the top N recommended messages for push, thereby avoiding pushing too many repeated or low-value messages to users, reducing the pressure of information overload, improving the efficiency of user content browsing, and enhancing the actual user experience.
[0042] In a preferred embodiment of the present invention, in S2, the specific process of recommending parameters is:
[0043] A unit time is preset and divided into several equal time periods, any application is selected and the recommended messages are classified according to the time period to which the generation time of the recommended messages of the application belongs, so as to obtain several sub-datasets; the first access time T after the generation time t of any recommended message is obtained, and the time difference between the generation time t and the access time T is calculated to obtain the response time, and the recommended value W of the sub-dataset is calculated according to the response time, and the time periods in which all the recommended values are less than the preset threshold value within the first m unit times of the current time are obtained, where m is a positive integer, and when the recommended value of any time period is less than the preset threshold value for more than n+β / 2×n times, where β is a preset coefficient, the time period is marked as a forbidden push node, and the remaining time period after the unit time of the target object is excluding the forbidden push node is marked as a time node, and the time node and the recommended value corresponding to the time node are marked as a user parameter.
[0044] In another preferred embodiment of the present invention, the sub-dataset is further preprocessed, and the specific process of the preprocessing is:
[0045] Get any access time T and access duration H, construct the access time interval [T, T+H], and delete the recommended messages within the access time interval.
[0046] In another preferred embodiment of the present invention, the calculation process of the recommended value is:
[0047] wi=1 / e -(t-T) ;
[0048]
[0049] Among them, wi is the recommendation score of any recommended message, t is the response time, T is the first request time, n is the number of preprocessed recommended messages in any sub-dataset, and W is the recommendation value.
[0050] In another preferred embodiment of the present invention, in S3, the specific process of sorting is:
[0051] Obtain the recommended parameters of the application corresponding to all target messages in turn, determine the time nodes of all target messages according to the recommended parameters, select any time node and obtain the target message corresponding to the time node, sort them in descending order of the recommended value to obtain the target message group Xi, where Xi∈[q1, q2, ..., qs], s represents the number of target messages in the time node, i is the number of the time node, and then obtain the target message group corresponding to all time nodes and mark them as X1, X2, ..., Xz, z is the number of time nodes, and then sort the target messages according to the time node numbers.
[0052] In another preferred embodiment of the present invention, said S3 further includes:
[0053] If there is a target message group corresponding to any time node that contains multiple target messages from the same application, then the arrangement order of the multiple targets in the target message group is obtained and the target message recommendation values are attenuated in turn according to the arrangement order. The attenuation factor is (1 / 2)^(n'-1), where n' is the number of target messages of the same application.
[0054] In another preferred embodiment of the present invention, S4 also includes successively attenuating the recommendation values of the recommendation messages of the same application in the recommendation queue, and the attenuation factor is (1 / 2)^(n"-1), where n" is the n"th content recommended by the same application.
[0055] A multimedia advertisement intelligent push system based on the Internet, comprising:
[0056] Data acquisition module: used to acquire the historical behavior of the target object client, the historical behavior includes the generation time of all applications recommending messages to the target object and the target object's access information; the target object's access information includes the target object's access time to any application, access duration and purchase data;
[0057] Data analysis module: used to select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; and perform data analysis on the time period corresponding to the recommendation value to obtain the time node, mark the time node and the corresponding recommendation value as the recommendation parameter, and obtain the recommendation parameters of all applications;
[0058] Data optimization module: used to extract text data from the purchase data and perform preprocessing to obtain target keywords, extract all to-be-recommended messages in the target object client storage module and search for the to-be-recommended messages according to the target keywords, and mark the to-be-recommended messages containing the target keywords as target messages; sort the target messages according to the recommendation parameters, generate a target message queue, and sequentially decay the target message recommendation values according to the target message queue, and the decay factor is (1 / 2)^(n-1), where n is the number of target messages;
[0059] Result generation module: used to determine all the messages to be recommended at any time node according to the recommendation parameters, sort the messages to be recommended from large to small according to the recommendation value and generate a recommended message sequence, and select the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
[0060] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.
Claims
1. A multimedia advertisement intelligent push method based on the Internet, characterized in that: The following steps are involved: S1, obtaining the historical behavior of the target object client, the historical behavior includes the generation time of all application recommendation messages to the target object and the target object's access information; the target object's access information includes the target object's access time, access duration and purchase data for any application; S2, select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; Perform data analysis on the time period corresponding to the recommended value to obtain the time node, mark the time node and the corresponding recommended value as the recommended parameter, and obtain the recommended parameters for all applications; S3, extracting text data from the purchase data and preprocessing it to obtain target keywords, extracting all the messages to be recommended in the target object client storage module and searching for the messages to be recommended according to the target keywords, marking the messages to be recommended containing the target keywords as target messages; sorting the target messages according to the recommendation parameters, generating a target message queue, and attenuating the target message recommendation values in turn according to the target message queue, and the attenuation factor is (1 / 2)^(n-1), where n is the number of target messages; S4, determining all the messages to be recommended at any time node according to the recommendation parameters, and sorting the messages to be recommended from large to small according to the recommendation values and generating a recommended message sequence, and selecting the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
2. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 1, characterized in that: In S2, the specific process of recommending parameters is as follows: A unit time is preset and divided into several equal time periods, any application is selected and the recommended messages are classified according to the time period to which the generation time of the recommended messages of the application belongs, so as to obtain several sub-datasets; the first access time T after the generation time t of any recommended message is obtained, and the time difference between the generation time t and the access time T is calculated to obtain the response time, and the recommended value W of the sub-dataset is calculated according to the response time, and the time periods in which all the recommended values are less than the preset threshold value within the first m unit times of the current time are obtained, where m is a positive integer, and when the recommended value of any time period is less than the preset threshold value for more than n+β / 2×n times, where β is a preset coefficient, the time period is marked as a forbidden push node, and the remaining time period after the unit time of the target object is excluding the forbidden push node is marked as a time node, and the time node and the recommended value corresponding to the time node are marked as a user parameter.
3. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 2, characterized in that: The method also includes preprocessing the sub-datasets, and the specific process of the preprocessing is as follows: Get any access time T and access duration H, construct the access time interval [T, T+H], and delete the recommended messages within the access time interval.
4. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 2, characterized in that: The calculation process of the recommended value is: wi=1 / e -(t-T) 4 Among them, wi is the recommendation score of any recommended message, t is the response time, T is the first request time, n is the number of preprocessed recommended messages in any sub-dataset, and W is the recommendation value.
5. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 1, characterized in that: In S3, the specific process of sorting is as follows: Obtain the recommended parameters of the application corresponding to all target messages in turn, determine the time nodes of all target messages according to the recommended parameters, select any time node and obtain the target message corresponding to the time node, sort them in descending order of the recommended value to obtain the target message group Xi, where Xi∈[q1, q2, ..., qs], s represents the number of target messages in the time node, i is the number of the time node, and then obtain the target message group corresponding to all time nodes and mark them as X1, X2, ..., Xz, z is the number of time nodes, and then sort the target messages according to the time node numbers.
6. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 1, characterized in that: The S3 further includes: If there is a target message group corresponding to any time node that contains multiple target messages from the same application, then the arrangement order of the multiple targets in the target message group is obtained and the target message recommendation values are attenuated in turn according to the arrangement order. The attenuation factor is (1 / 2)^(n'-1), where n' is the number of target messages of the same application.
7. The method for intelligently pushing multimedia advertisements based on the Internet according to claim 1, characterized in that: The S4 also includes successively attenuating the recommendation values of the recommendation messages of the same application in the recommendation queue, and the attenuation factor is (1 / 2)^(n"-1), where n" is the n"th content recommended by the same application.
8. An Internet-based multimedia advertising intelligent push system, characterized in that: include: Data acquisition module: used to acquire the historical behavior of the target object client, the historical behavior includes the generation time of all applications recommending messages to the target object and the target object's access information; the target object's access information includes the target object's access time to any application, access duration and purchase data; Data analysis module: used to select any application and calculate the recommendation value according to the generation time and access time of the corresponding recommendation message; Perform data analysis on the time period corresponding to the recommended value to obtain the time node, mark the time node and the corresponding recommended value as the recommended parameter, and obtain the recommended parameters for all applications; Data optimization module: used to extract text data from the purchase data and perform preprocessing to obtain target keywords, extract all to-be-recommended messages in the target object client storage module and search for the to-be-recommended messages according to the target keywords, and mark the to-be-recommended messages containing the target keywords as target messages; sort the target messages according to the recommendation parameters, generate a target message queue, and sequentially decay the target message recommendation values according to the target message queue, and the decay factor is (1 / 2)^(n-1), where n is the number of target messages; Result generation module: used to determine all the messages to be recommended at any time node according to the recommendation parameters, sort the messages to be recommended from large to small according to the recommendation value and generate a recommended message sequence, and select the top N recommended messages in the recommended message queue to recommend to the target object within the time node.
Citation Information
Patent Citations
Commodity sequence calculation method and device, computer equipment and storage medium
CN110569432A
Electronic bidding and tendering transaction platform bidding and tendering information recommendation method and system
CN115098766A
Advertisement resource recommendation method and system based on Internet game platform
CN116503115A
Intelligent advertisement pushing method based on big data
CN116739675A
Network content recommendation method based on big data
CN118094015A
Cited By
Intelligent advertisement pushing method based on Internet advertisement platform
CN120634648A
