An Internet-based AI-powered advertising and marketing push method

Abstract

This invention discloses an AI-based advertising and marketing push method for the internet, specifically relating to the field of advertising and marketing technology. It includes analyzing search hot words and trending topics using natural language processing and text analysis techniques; determining characteristic trend words on the internet platform by statistically analyzing word frequency; identifying characteristic trend words on the internet platform and in push advertisements in real time; using knowledge graph technology and sentiment analysis to determine the trend changes of push advertisements; and determining the potential popularity of push advertisements by analyzing the relationship between the volume of push advertisement traffic and the number of ad clicks on the internet platform, thus determining whether to continue supporting push advertisement traffic. This invention helps avoid excessive investment in inefficient advertisements and optimizes advertising push strategies in real time based on market and user feedback, thereby maximizing the effectiveness of advertising.

Description

Technical Field

[0001] This invention relates to the field of advertising and marketing technology, and more specifically, to an internet-based artificial intelligence advertising and marketing push method. Background Technology

[0002] With the rapid development of the Internet, the advertising and marketing industry has ushered in unprecedented opportunities and challenges. Especially on Internet platforms, ad push has become an important business model. Existing ad push technology mainly relies on traditional keyword matching, user behavior analysis and traffic optimization algorithms. However, when faced with rapid market changes and diversified user interests, traditional methods are difficult to effectively adjust ad push strategies and are not easy to accurately identify target users during push, resulting in wasted advertising traffic and budget.

[0003] To address the aforementioned shortcomings, a technical solution is provided. Summary of the Invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, embodiments of the present invention provide an Internet-based artificial intelligence advertising and marketing push method to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An internet-based AI-powered advertising and marketing push method specifically includes the following steps:

[0007] S1: By using natural language processing and text analysis techniques to analyze search hot words and popular topics, and by statistically analyzing the frequency of word occurrences, we can determine the characteristic trend words of the Internet platform.

[0008] S2: By identifying the characteristic trend words in the Internet platform and the characteristic trend words of push advertisements in real time, knowledge graph technology and sentiment analysis technology are used to determine the trend change information of push advertisements.

[0009] S3: By analyzing the relationship between the volume of traffic pushed by internet platforms and the number of ad clicks, we can determine the potential popularity of the pushed ads using correlation analysis methods.

[0010] S4: Conduct a comprehensive analysis of the popularity potential information of the push ads and determine whether to continue to support the push ads with traffic.

[0011] In a preferred embodiment, determining the characteristic trend words of the Internet platform includes:

[0012] We capture trending search terms and hot topics from internet platforms in real time, and extract characteristic trend words from these platforms using text analysis technology. These characteristic trend words are identified by statistically analyzing the frequency of their occurrence in the captured trending search terms and hot topics, setting a frequency threshold, and determining words that exceed the frequency threshold as characteristic trend words.

[0013] In a preferred embodiment, information on the potential popularity of the pushed advertisement is determined.

[0014] The trend changes in pushed advertisements are represented by the hot topic correlation coefficient and the hot topic index trend coefficient;

[0015] The logic for obtaining the hotspot correlation coefficient is as follows: Construct a knowledge graph based on the feature trend words in the Internet platform, where nodes in the knowledge graph represent entities, entities are various trend feature words, edges in the knowledge graph represent the correlation between entities, and the correlation is the weight between entities.

[0016] Based on historical records on internet platforms, the co-occurrence frequency between entities is determined, and the weights between entities are calculated using the following formula: Where QZ(A,B) represents the weight between entity A and entity B, CS(A,B) represents the number of times entity A and entity B appear simultaneously as feature trend terms in the historical records of the internet platform, and ∑ C∈all terms CS(A,C) is the sum of the co-occurrence counts of entity A with all other trend feature terms;

[0017] Determine the set of trending terms characteristic of internet platforms, and represent the set of trending terms characteristic of internet platforms as: LW i =[LW1,LW2,LW3,……,LW I The system uses word segmentation technology to determine the vocabulary of push ads, obtains the characteristic trend vocabulary of push ads, and determines the set of characteristic trend vocabulary of push ads. The set of characteristic trend vocabulary of push ads is represented as: QS n =[QS1,QS2,QS3,……,QS N ];

[0018] The hotspot correlation coefficient is calculated using the following formula: Among them, GL rd This represents the hotspot correlation coefficient.

[0019] In a preferred embodiment, the hotspot index trend coefficient includes:

[0020] The logic for obtaining the trend coefficient of the hot topic index is as follows: The vocabulary of the push ad comments is determined using word segmentation technology; the sentiment score of each word in the push ad comments is determined using a sentiment dictionary; the total sentiment score of the push ad comments is determined; and the total sentiment score of each push ad comment is marked as: DFm Where m = 1, 2, 3, ..., M, M is a positive integer, and m is the number of the push advertisement comment;

[0021] The trend coefficient of the trending index for push ads is calculated using the following formula: Among them, ZS qs DZ is used to measure the trend of trending topics for push ads. m The number of likes for each push ad comment.

[0022] In a preferred embodiment, determining the potential popularity of the pushed advertisement includes:

[0023] The potential popularity of pushed ads is represented by a traffic anonymity coefficient;

[0024] The logic for obtaining the traffic anonymity coefficient is as follows: A monitoring period is set, the volume of traffic for each push of the advertisement within the monitoring period is determined, and the volume of traffic for each round of pushes within the monitoring period is marked as: LL k Where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of each data transmission during the monitoring period;

[0025] Determine the increase in clicks for each push ad during the monitoring period, and label this increase as: ZJ k ;

[0026] The flow concealment coefficient is calculated using the following formula: Where XG is the traffic concealment coefficient. To monitor the average amount of traffic pushed each time within a specified time period, This is to monitor the average increase in clicks for each pushed ad within a given time period.

[0027] In a preferred embodiment, a comprehensive analysis is performed on the popularity potential information and popularity potential information of the pushed advertisement, including:

[0028] By comprehensively analyzing the trend changes and potential popularity of pushed advertisements, a push traffic evaluation model is constructed by weighting the hot topic correlation coefficient, hot topic index trend coefficient, and traffic concealment coefficient. This model generates a push traffic evaluation coefficient, the formula for which is as follows: Among them, PG tl The coefficients for evaluating the flow of traffic are α1, α2, and α3, which are the proportional coefficients of the hotspot correlation coefficient, the hotspot index trend coefficient, and the traffic concealment coefficient, respectively. α1, α2, and α3 are all greater than 0.

[0029] In a preferred embodiment, determining whether to continue providing push traffic support for the pushed advertisement includes:

[0030] Set a push evaluation coefficient threshold and compare the push evaluation coefficient of the pushed advertisement with the push evaluation coefficient threshold;

[0031] If the push evaluation coefficient of the push advertisement is greater than the push evaluation coefficient threshold, then the push advertisement will continue to receive traffic support.

[0032] If the push evaluation coefficient of the pushed advertisement is less than the push evaluation coefficient threshold, an early warning signal is generated, and new pushes to the pushed advertisement are stopped. The user is further analyzed through artificial intelligence technology to identify different user groups and determine the push evaluation coefficient of the pushed advertisement in different user groups. The push evaluation coefficient of different user groups is compared with the push evaluation coefficient threshold. For user groups whose push evaluation coefficient is greater than the push evaluation coefficient threshold, push traffic support is continuously provided until the push evaluation coefficient of the pushed advertisement is less than the push evaluation coefficient threshold.

[0033] The technical effects and advantages of this invention are as follows:

[0034] This invention analyzes search terms and trending topics using natural language processing and text analysis techniques to identify characteristic trend words on internet platforms. It then uses knowledge graph and sentiment analysis techniques to determine the potential popularity of pushed advertisements. Furthermore, it employs relevance analysis to further determine the potential popularity of pushed advertisements. By comprehensively analyzing these potential popularity and trending trends, the invention determines whether to continue supporting the pushed advertisements with increased traffic. This invention helps avoid over-investment in inefficient advertising and optimizes advertising strategies in real time based on market and user feedback, thereby maximizing the effectiveness of advertising campaigns. Attached Figure Description

[0035] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0036] Figure 1 This is a flowchart illustrating an internet-based artificial intelligence-powered advertising and marketing push method according to the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1

[0039] Figure 1 This is a flowchart illustrating an internet-based artificial intelligence-powered advertising and marketing push method according to the present invention, which specifically includes the following steps:

[0040] S1: By using natural language processing and text analysis techniques to analyze search hot words and popular topics, and by statistically analyzing the frequency of word occurrences, we can determine the characteristic trend words of the Internet platform.

[0041] S2: By identifying the characteristic trend words in the Internet platform and the characteristic trend words of push advertisements in real time, knowledge graph technology and sentiment analysis technology are used to determine the trend change information of push advertisements.

[0042] S3: By analyzing the relationship between the volume of traffic pushed by internet platforms and the number of ad clicks, we can determine the potential popularity of the pushed ads using correlation analysis methods.

[0043] S4: Conduct a comprehensive analysis of the popularity potential information of the push ads and determine whether to continue to support the push ads with traffic.

[0044] Ad streaming is a dynamic optimization process based on user data, ad content, and advertiser needs. By leveraging big data, machine learning, and real-time bidding technologies, ad streaming can achieve precise targeting, personalized recommendations, and performance optimization, thereby improving ad conversion rates and advertisers' ROI. By combining ad streaming with trend-based keyword analysis, it can capture real-time fluctuations and trends in user interests, ensuring high relevance and timeliness of ad content, thus improving ad effectiveness. This includes:

[0045] Featured trend words refer to the changing trends in search frequency, social discussion volume, webpage clicks, etc. of certain specific words or keywords within a certain period of time. These words can reflect users' interests, needs, or the hot topics of a certain event.

[0046] Using natural language processing (NLP) and text analysis techniques, we can track the popularity changes of specific words or phrases in real time, determine which words are becoming hot topics, and perform cluster analysis on frequently occurring keywords to identify potential hot words.

[0047] Based on the analyzed hot words or trending topics, we customize related advertising content, create ads that match the trend of characteristic words through automated ad generation technology, and push ads containing characteristic words to the audience.

[0048] Real-time data analysis is performed using stream processing frameworks (such as Apache Kafka, Apache File, etc.) to quickly identify and respond to trend changes in feature words. Valuable trend information is extracted from massive amounts of user behavior data, search data, and social media data through big data technology analysis.

[0049] By identifying trending keywords in internet platforms and push ads in real time, knowledge graph and sentiment analysis technologies are used to determine the trend changes in push ads. These trend changes are then represented by hotspot correlation coefficients and hotspot index trend coefficients.

[0050] The advantages of the hotspot correlation coefficient are as follows:

[0051] The hot topic correlation coefficient can capture the characteristic trend words of Internet platforms (such as hot topics and popular keywords), and process the advertising content through natural language processing (NLP), and further use knowledge graphs to determine the correlation between the characteristic trend words of Internet platforms and advertisements.

[0052] The hot topic relevance coefficient analyzes trending words on internet platforms to capture popular topics or keywords that users are interested in in a timely manner, enabling the advertising push system to push advertising content that is highly consistent with users' interests at the right time, thereby improving the relevance of the advertisements.

[0053] By using knowledge graphs to analyze the semantic relationship between advertisements and trending topics, the advertising system can accurately target advertisements to user groups that are relevant to the trend and have high attention to it, avoiding blind advertising and reducing resource waste.

[0054] By leveraging the hot topic relevance coefficient, the advertising system can avoid pushing irrelevant ads to users, thereby preventing ineffective traffic consumption.

[0055] The logic for obtaining the hot topic correlation coefficient is as follows: real-time capture of hot search terms and trending topics in the Internet platform, extraction of characteristic trend words in the Internet platform through text analysis technology, and determination of characteristic trend words by statistically analyzing the frequency of occurrence of words in the captured hot search terms and trending topics, setting a frequency threshold, and identifying words with a frequency threshold greater than the frequency threshold as characteristic trend words.

[0056] Construct a knowledge graph based on feature trend words in the Internet platform. In the knowledge graph, nodes represent entities, entities are various trend feature words, edges represent the degree of association between entities, and the degree of association is the weight between entities.

[0057] It should be noted that search terms and trending topics are dynamic and change, so the construction of a knowledge graph is a continuous optimization process. By constantly capturing different trend feature words in real time, a knowledge graph of feature trend words based on the historical records of Internet platforms is formed.

[0058] Based on historical records on internet platforms, the co-occurrence frequency between entities is determined, and the weights between entities are calculated using the following formula: Where QZ(A,B) represents the weight between entity A and entity B, CS(A,B) represents the number of times entity A and entity B appear simultaneously as feature trend terms in the historical records of the internet platform, and ∑ C∈all terms CS(A,C) is the sum of the co-occurrence counts of entity A with all other trend feature terms;

[0059] Determine the set of trending terms characteristic of internet platforms, and represent the set of trending terms characteristic of internet platforms as: LW i =[LW1,LW2,LW3,……,LW I The system uses word segmentation technology to determine the vocabulary of push ads, obtains the characteristic trend vocabulary of push ads, and determines the set of characteristic trend vocabulary of push ads. The set of characteristic trend vocabulary of push ads is represented as: QS n =[QS1,QS2,QS3,……,QS N ];

[0060] The hotspot correlation coefficient is calculated using the following formula: Among them, GL rd This represents the hotspot correlation coefficient.

[0061] As the formula shows, the larger the hot topic correlation coefficient, the higher the correlation coefficient usually means that the characteristic trend words (such as popular topics and hot keywords) on the Internet platform and the characteristic words (keywords, descriptions, etc. in the push advertisement) have a high degree of semantic matching, which means that the traffic of the push advertisement can be increased.

[0062] The advantage of the hotspot index trend coefficient is that:

[0063] By using sentiment lexicon and sentiment analysis technology, it is possible to accurately identify the sentiment tendencies of users in ad reviews (such as positive, negative or neutral), which enables advertising platforms to gain a deeper understanding of users' reactions to ads and adjust ad content or push strategies based on sentiment feedback.

[0064] By analyzing user sentiment, advertising platforms can enhance ad exposure based on comments with positive sentiment, avoid the impact of negative sentiment, and ultimately improve ad performance and user satisfaction.

[0065] Advertising platforms can use sentiment analysis to avoid pushing unpopular or negatively impacting ads to users, reducing ad waste and ensuring that ad budgets are spent precisely on effective ad delivery.

[0066] The logic for obtaining the trend coefficient of the hot topic index is as follows: The vocabulary of the push ad comments is determined using word segmentation technology; the sentiment score of each word in the push ad comments is determined using a sentiment dictionary; the total sentiment score of the push ad comments is determined; and the total sentiment score of each push ad comment is marked as: DF m Where m = 1, 2, 3, ..., M, M is a positive integer, and m is the number of the push advertisement comment;

[0067] It should be noted that for Chinese text in ad reviews, commonly used tools include Jieba (for Chinese word segmentation) or deep learning-based word segmentation methods such as BERT. By assigning a sentiment score to each word, the total sentiment score of the ad review is the sum of the sentiment scores of each word in the ad review.

[0068] The trend coefficient of the trending index for push ads is calculated using the following formula: Among them, ZS qs DZ is used to measure the trend of trending topics for push ads. m The number of likes for each push ad comment.

[0069] As the formula shows, the larger the trend coefficient of the hot topic index, the more popular the content of the pushed advertisement is with the public. This indicates that the advertisement has a greater chance of exposure, user participation, interaction possibility and market potential, which means that the traffic of the pushed advertisement can be increased.

[0070] By analyzing the relationship between the volume of traffic and the number of clicks on push advertisements on internet platforms, correlation analysis is used to determine the potential popularity of push advertisements, and this potential popularity is represented by a traffic anonymity coefficient.

[0071] The advantages of the traffic concealment coefficient are:

[0072] The traffic anonymity coefficient can effectively detect mismatches between traffic and clicks. For example, if the number of clicks does not increase accordingly even if the traffic pushed increases, the traffic anonymity deviation coefficient will increase. This deviation may reflect a poor match between the advertising content and the target audience, or an inaccurate traffic source, and can promptly identify and correct the phenomenon of inflated traffic.

[0073] The traffic anonymity coefficient can clearly reveal whether the effect of the advertising push has achieved the expected results. If the traffic anonymity coefficient is low, specific optimization measures can be taken, such as adjusting the advertising content, accurately targeting the target audience, and improving the advertising display method.

[0074] The logic for obtaining the traffic anonymity coefficient is as follows: A monitoring period is set, the volume of traffic for each push of the advertisement within the monitoring period is determined, and the volume of traffic for each round of pushes within the monitoring period is marked as: LL k Where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of each data transmission during the monitoring period;

[0075] Determine the increase in clicks for each push ad during the monitoring period, and label this increase as: ZJ k ;

[0076] The flow concealment coefficient is calculated using the following formula: Where XG is the traffic concealment coefficient. To monitor the average amount of traffic pushed each time within a specified time period, To monitor the average increase in clicks for each pushed ad within a given time period;

[0077] It should be noted that the monitoring period is a specific time period used to analyze the relationship between the volume of traffic pushed by the push advertisement and the number of ad clicks. The monitoring period is set by professionals in the field. The larger the volume of traffic pushed by the push advertisement, the larger the corresponding number of ad clicks. When the number of ad clicks does not change significantly due to the volume of traffic pushed by the push advertisement, it means that the push advertisement has lost its potential and the push advertisement needs to be stopped.

[0078] As the formula shows, the larger the traffic anonymity coefficient, the stronger the correlation between traffic and clicks. When the push traffic is large, the clicks of the pushed ads are large. Conversely, the smaller the traffic anonymity coefficient, the less significant the increase in clicks when traffic increases, or even the negative correlation. This indicates that the pushed traffic may not be effectively converted into clicks. This may be due to factors such as the ad content, target audience matching, or the ad presentation method not attracting users to click. It is necessary to reduce the push traffic for the pushed ads.

[0079] By comprehensively analyzing the trend changes and potential popularity of pushed advertisements, a push traffic evaluation model is constructed by weighting the hot topic correlation coefficient, hot topic index trend coefficient, and traffic concealment coefficient. This model generates a push traffic evaluation coefficient, the formula for which is as follows: Among them, PG tl The coefficients for evaluating the flow of traffic are α1, α2, and α3, which are the proportional coefficients of the hotspot correlation coefficient, the hotspot index trend coefficient, and the traffic concealment coefficient, respectively. α1, α2, and α3 are all greater than 0.

[0080] As can be seen from the formula, the larger the hot topic correlation coefficient, hot topic index trend coefficient, and traffic concealment coefficient, the larger the push evaluation coefficient, indicating that push ads can increase push traffic to users. Conversely, the smaller the hot topic correlation coefficient, hot topic index trend coefficient, and traffic concealment coefficient, the smaller the push evaluation coefficient, indicating that push ads should reduce push traffic to users.

[0081] A threshold for the push traffic evaluation coefficient is set, and the push traffic evaluation coefficient of the pushed advertisement is compared with the threshold. If the push traffic evaluation coefficient of the pushed advertisement is greater than the threshold, the push traffic support for the pushed advertisement will continue. If the push traffic evaluation coefficient of the pushed advertisement is less than the threshold, an early warning signal is generated, and new push traffic to the pushed advertisement is stopped. Through artificial intelligence technology, users are further analyzed to identify different user groups and determine the push traffic evaluation coefficient of the pushed advertisement for different user groups. The push traffic evaluation coefficient of different user groups is compared with the push traffic evaluation coefficient threshold. User groups whose push traffic evaluation coefficient is greater than the threshold will continue to receive push traffic support until the push traffic evaluation coefficient of the pushed advertisement is less than the threshold.

[0082] This invention analyzes search terms and trending topics using natural language processing and text analysis techniques to identify characteristic trend words on internet platforms. It uses knowledge graph and sentiment analysis techniques to determine the trend changes in pushed advertisements and uses relevance analysis to determine the potential popularity of these advertisements. By comprehensively analyzing the potential popularity information and the potential popularity data, this invention determines whether to continue supporting the push advertisements with increased traffic. This invention helps avoid over-investment in inefficient advertising and optimizes advertising strategies in real time based on market and user feedback, thereby maximizing the effectiveness of advertising campaigns.

[0083] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0084] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0085] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0086] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0087] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An Internet-based artificial intelligence advertising marketing push method, characterized in that, Specifically comprising the following steps: S1: by using natural language processing and text analysis techniques to analyze search hot words and hot topics, by counting the frequency of the appearance of the vocabulary, determine the characteristic trend vocabulary of the internet platform; S2: by determining the characteristic trend vocabulary of the internet platform and the characteristic trend vocabulary of the push advertisement in real time, using knowledge graph technology and sentiment analysis technology, determine the trend change information of the push advertisement; S3: by the relationship between the flow size and the advertisement click volume of the push advertisement in the internet platform, using correlation analysis method, determine the heat potential information of the push advertisement; S4: the trend change information and heat potential information of the push advertisement are analyzed comprehensively, judge whether to continue to support the push flow of the push advertisement; Determine the characteristic trend vocabulary of the internet platform, including: Real-time capture of search hot words and hot topics in the internet platform, extract the characteristic trend vocabulary in the internet platform through text analysis technology, the characteristic trend vocabulary is counted by the frequency of the appearance of the vocabulary in the captured search hot words and hot topics, set the frequency threshold, determine the vocabulary greater than the frequency threshold as the characteristic trend vocabulary; Determine the trend change information of the push advertisement, including: The trend change information of the push advertisement is represented by the hot spot correlation coefficient and the hot spot index trend coefficient; The logic of obtaining the hot spot correlation coefficient is: constructing a knowledge graph based on the characteristic trend vocabulary in the internet platform, wherein the nodes in the knowledge graph represent entities, and the entities are various trend characteristic vocabularies, the edges in the knowledge graph represent the correlation degree between entities, and the correlation degree is the weight between entities; Based on the history record of the Internet platform, the co-occurrence frequency between entities is determined, and the weight between entities is calculated, and the calculation formula is: ; wherein, is the weight between entity A and entity B, is the number of times that entity A and entity B appear simultaneously as characteristic trend words in the history record of the Internet platform, is the sum of the co-occurrence times of entity A and all other trend characteristic words; Determine the Internet platform feature trend vocabulary set, and represent the Internet platform feature trend vocabulary set as: Determine the vocabulary of the push advertisement through the word segmentation technology, obtain the feature trend vocabulary of the push advertisement, determine the push advertisement feature trend vocabulary set, and represent the push advertisement feature trend vocabulary set as: ; The heat spot correlation coefficient is calculated, and the calculation formula is: ; wherein, is the heat spot correlation coefficient.

2. The Internet-based artificial intelligence advertising marketing pushing method according to claim 1, characterized in that, Hot spot index trend coefficient, including: The obtaining logic of the hotspot index trend coefficient is: determining the words of the push advertisement comment through a word segmentation technique, determining the sentiment scores of the words of the push advertisement comment using a sentiment dictionary, determining the total sentiment scores of the push advertisement comments, and marking the total sentiment scores of the push advertisement comments as: wherein m=1, 2, 3, …, M, M is a positive integer, and m is the number of the push advertisement comment. The hot spot index trend coefficient of the push advertisement is calculated, and the calculation formula is: ; wherein, is the hot spot index trend coefficient of the push advertisement, is the like quantity of each push advertisement comment.

3. The Internet-based artificial intelligence advertising marketing pushing method according to claim 2, characterized in that, Determine the heat potential information of the push advertisement, including: The heat potential information of the push advertisement is represented by the flow concealment coefficient; The acquisition logic of the flow concealment coefficient is: setting a monitoring time period, determining the push flow size of each time of the push advertisement in the monitoring time period, and marking the push flow size of each round of the push advertisement in the monitoring time period as: Wherein, k=1, 2, 3, …, K, K is a positive integer, and k is the number of each delivery flow in the monitoring time period. The increase of the push advertisement click quantity after each push of the push advertisement in the monitoring time period is determined, and the increase of the push advertisement click quantity after each push of the push advertisement in the monitoring time period is marked as: ; The flow concealment coefficient is calculated, and the calculation formula is: ; wherein, XG is the flow concealment coefficient, is the average value of the flow of each push in the monitoring time period, is the average value of the increase of the advertisement click volume of each push in the monitoring time period.

4. The Internet-based artificial intelligence advertising marketing pushing method according to claim 3, characterized in that, The trend change information and heat potential information of the push advertisement are analyzed comprehensively, including: Through comprehensive analysis of the trend change information of the pushed advertisement and the hotness potential information of the pushed advertisement, a push flow evaluation model is constructed by weighted calculation of a hot spot correlation coefficient, a hot spot index trend coefficient and a flow concealment coefficient, and a push flow evaluation coefficient is generated, and a calculation formula of the push flow evaluation coefficient is: ; wherein, is the push flow evaluation coefficient, , , a proportion coefficient of the hot spot correlation coefficient, the hot spot index trend coefficient and the flow concealment coefficient respectively, , , all are greater than 0.

5. The Internet-based artificial intelligence advertising marketing pushing method according to claim 4, characterized in that, Determine whether to continue to support the push flow of the push advertisement, including: Set the push flow evaluation coefficient threshold, compare the push flow evaluation coefficient of the push advertisement with the push flow evaluation coefficient threshold; If the push flow evaluation coefficient of the push advertisement is greater than the push flow evaluation coefficient threshold, continue to support the push flow of the push advertisement; If the push flow evaluation coefficient of the push advertisement is less than the push flow evaluation coefficient threshold, generate an early warning signal, stop increasing new push flow for the push advertisement, further analyze the users through artificial intelligence technology, identify different user groups, determine the push flow evaluation coefficient of the push advertisement in different user groups, compare the push flow evaluation coefficient of different user groups with the push flow evaluation coefficient threshold, continue to support the push flow of the user groups greater than the push flow evaluation coefficient threshold, until the push flow evaluation coefficient of the push advertisement is less than the push flow evaluation coefficient threshold.

Images