Advertisement marketing method and system based on data analysis

Abstract

The present invention discloses an advertising marketing method and system based on data analysis, and the present invention relates to the field of advertising marketing technology. The present invention optimizes the whole process of advertising delivery to realize a complete closed loop from user data collection and analysis to accurate advertising matching and delivery; through the scientific division of software categories and user browsing formats, and the collection of content information and behavior data based on user authorization, the generation accuracy of user interest tags is significantly improved, laying a solid foundation for the personalization and precision of advertising delivery; this method effectively reduces the delivery of invalid advertisements, improves the fit between advertising content and user needs, and optimizes the user's advertising reception experience. At the same time, the present invention combines natural language processing technology and algorithm models, from the matching of interest tags and advertising information to the screening of the best advertisements, to maximize the efficiency of advertising delivery.

Description

Technical Field

[0001] The present invention relates to the technical field of advertising and marketing, and in particular to an advertising and marketing method and system based on data analysis. Background Art

[0002] With the popularization of Internet technology and smart devices, digital advertising marketing has become an important means for enterprises to promote their products and services. However, traditional advertising models usually rely on a single source of user data or extensive target user group division, resulting in a large deviation between advertising content and user needs. This not only reduces the click-through rate and conversion rate of advertisements, but also increases the marketing costs of enterprises, and may also have a negative impact on user experience.

[0003] In recent years, with the rapid development of big data analysis, natural language processing (NLP) and machine learning technologies, data-driven advertising has gradually become an industry trend. Through in-depth mining and analysis of user behavior data, users' interest preferences can be accurately identified to achieve personalized advertising recommendations. However, the current advertising marketing system still has technical barriers in data collection, user interest analysis, advertising matching and delivery, such as insufficient data analysis dimensions and low accuracy in interest tag generation. Therefore, there is an urgent need for a more scientific and systematic advertising marketing method and system to improve the accuracy and effectiveness of advertising delivery, while reducing resource waste, optimizing user experience and improving the company's return on investment. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an advertising marketing method and system based on data analysis, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an advertising marketing method based on data analysis, comprising:

[0006] Step 1: Categorize the software that needs to be advertised, and determine the browsing format of the user in the software based on the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software;

[0007] Step 2: Then obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results;

[0008] Step 3: Match the interest tag with the advertising information to obtain an interest-based advertising set, and determine the best advertisement from the interest-based advertising set;

[0009] Step 4: Place the best advertisement in the corresponding software.

[0010] As a further solution of the present invention: in the step 1, the user browsing content is represented as promotional content published on the software;

[0011] The advertisement placement information indicates that promotional advertisements are allowed to be placed on the software.

[0012] As a further solution of the present invention: in the step 2, the user browsing content and browsing format in the content information are determined, converted into text content, and the main information in the text content is identified to obtain summary tags, and the user's browsing data is collected and analyzed. According to the analysis results, the specific method of obtaining the user's interest tags is as follows:

[0013] AS1: Mark the user's browsing content as a browsing target. When the user browses the browsing target in the software, the browsing format of the browsing target is identified. When the identified browsing format is text, its content is text content, and no processing will be performed at this time;

[0014] When the format of the recognized browsing target is a video, the voice in the video is recognized and converted into text to obtain text content. When the format of the recognized browsing target is pure voice, it is directly converted into text to obtain text content.

[0015] AS2: Then recognize the text content and obtain the summary label of the text content based on the overall text content;

[0016] AS3: Based on user authorization, monitor the user's browsing operations and obtain browsing data. The browsing data includes the total browsing time, number of repetitions, and interval duration. The specific method for obtaining browsing data is as follows:

[0017] When the user's browsing target is determined and the user starts browsing, the timing starts from this node and ends when the user exits the browsing target. The time from the start to the end of the timing is marked as the total browsing time;

[0018] When the user views the content in the browsing target, the content in the browsing target is divided according to the node content to obtain the interval content;

[0019] Then, it is identified whether there is advertising information in the interval content, and the specific advertising information includes product name and brand name. When there is advertising information in the interval content, the interval content is marked as first-level content;

[0020] Then obtain the user's browsing speed for each interval content and mark the browsing speed as ,in, Indicates different interval contents;

[0021] At the same time, the browsing time and the number of repetitions of the interval content by the user are obtained, and the browsing time of the interval content by the user is marked as the interval time. The interval time is the total browsing time of the interval content, specifically including the sum of the time when the user repeatedly views the interval content. The number of repetitions is the number of times the user repeatedly views the interval content after completing browsing.

[0022] AS4: Analyze the browsing data to obtain the user's product interest tags. The specific analysis method of the user's interest tags is as follows:

[0023] The interest value of the interval content is calculated by the following formula calculate:

[0024]

[0025] In the formula, Indicates the duration of the interval. Indicates the total browsing time. is a fixed coefficient, Indicates the number of repetitions of the interval content. and are weight coefficients respectively;

[0026] The interest value of the interval content and the threshold For comparison, The interval content is marked as secondary content, otherwise, The interval content is marked as inertial content, and no processing is performed on the inertial content.

[0027] As a further solution of the present invention: the step AS4 also includes:

[0028] After obtaining the secondary content, first identify whether there is primary content in the secondary content, that is, identify whether there is advertising information in the secondary content. If so, first directly mark its product name and brand name as interest tags;

[0029] If it does not exist, the secondary content is identified at this time. According to the text described in the secondary content, the secondary content and the summary tag are used as association factors respectively, and optical character recognition is used to obtain associated advertising information, and the product name and brand name in the associated advertising information are used as interest tags.

[0030] As a further solution of the present invention: in the step AS3, the specific method of obtaining the summary tag of the text content according to the overall text content is:

[0031] AS21: Identify the main information in the text content and reassemble the main information in the text content to obtain an abbreviated text;

[0032] AS22: Then, the abbreviated text is subjected to label extraction, that is, the number of occurrences of each word in the subject of the abbreviated text is divided by the total number of words in the subject to obtain the frequency of occurrence of the corresponding word, the word with the highest frequency of occurrence is obtained, and it is marked as the keyword of the subject, and the predicate and the object are processed in the same way as the subject to obtain the keywords of the predicate and the object respectively;

[0033] The keywords are then recombined and the resulting sentences are used as summary tags for the text content.

[0034] As a further solution of the present invention: in the step 3, the specific method of matching the interest tag with the advertisement delivery information to obtain the interest advertisement set, and determining the best advertisement from the interest advertisement set is as follows;

[0035] BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ,in, Expressed as the total number of advertising keywords;

[0036] Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; ;

[0037] Use natural language processing technology to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers;

[0038] BS2: For each interest tag and ad keywords Calculate the similarity and form a The similarity matrix M :

[0039]

[0040] In the formula, Expressed as cosine similarity, measuring interest tags With advertising keywords The matching degree of is the dot product of the vectors, and is the norm of the vector;

[0041] BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ;

[0042] BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad :

[0043]

[0044] In the formula, Expressed as the click-through rate of the ad, is the weight coefficient;

[0045] BS5: Select the advertisement with the highest score as the best advertisement.

[0046] An advertising marketing system based on data analysis, comprising:

[0047] The information collection module is used to classify the software that needs to be advertised, and determine the browsing format of the user in the software according to the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software;

[0048] The data analysis module is used to obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results;

[0049] An advertisement matching module is used to match the interest tag with the advertisement delivery information to obtain an interest advertisement set and determine the best advertisement from the interest advertisement set;

[0050] The advertisement delivery module is used to deliver the determined best advertisement in the corresponding software.

[0051] In the advertisement matching module, the specific method of matching the interest tag with the advertisement delivery information to obtain an interest advertisement set, and determining the best advertisement from the interest advertisement set is:

[0052] BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ;

[0053] Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; ;

[0054] Use natural language processing technology to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers;

[0055] BS2: For each interest tag and interest-based advertising Calculate the similarity and form a The similarity matrix M :

[0056]

[0057] In the formula, Expressed as cosine similarity, measuring interest tags With advertising The matching degree of is the dot product of the vectors, and is the norm of the vector;

[0058] BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ;

[0059] BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad :

[0060]

[0061] In the formula, Expressed as the click-through rate of the ad, is the weight coefficient;

[0062] BS5: Select the advertisement with the highest score as the best advertisement.

[0063] The present invention provides an advertising marketing method and system based on data analysis. Compared with the prior art, it has the following beneficial effects:

[0064] The present invention optimizes the entire process of advertisement delivery, thereby realizing a complete closed loop from user data collection and analysis to accurate advertisement matching and delivery; through the scientific classification of software categories and user browsing formats, as well as the collection of content information and behavior data based on user authorization, the generation accuracy of user interest tags is significantly improved, laying a solid foundation for the personalization and precision of advertisement delivery; this method effectively reduces the delivery of invalid advertisements, improves the fit between advertisement content and user needs, and optimizes the user's advertisement reception experience.

[0065] At the same time, the present invention combines natural language processing technology and algorithm models to maximize the efficiency of advertising delivery, from matching interest tags with advertising information to screening of optimal advertisements. While improving the conversion rate of advertising delivery, it also provides advertisers and platforms with efficient data support and optimization strategies, thereby improving the commercial value and return on investment of advertising delivery while saving resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The present invention will be further described below in conjunction with the accompanying drawings.

[0067] Figure 1 A flowchart of the steps of an advertising and marketing method based on data analysis of the present invention;

[0068] Figure 2 This is a structural framework diagram of an advertising and marketing system based on data analysis in the present invention. DETAILED DESCRIPTION

[0069] 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.

[0070] Embodiment 1

[0071] See also Figure 1-Figure 2 ,The present invention provides an advertising marketing method based on data analysis, comprising:

[0072] Step 1: Categorize the software that needs to be advertised, and determine the browsing format of the user in the software based on the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software;

[0073] It should be noted that the software that needs to be advertised is divided into categories, including shopping type, social media type, social e-commerce type and short video platform type. The classification method is that professional staff classify the mainstream software on the market in advance. When obtaining the software that needs to be advertised, it can be directly determined which category the software belongs to;

[0074] According to the classification of the software, determine the browsing format of the user in the software. The specific browsing format includes pictures, videos, texts, and audios, which are determined by professional and experienced personnel. For example, for shopping-type software, it can be determined that the browsing format of the user in the software is picture and text format. For short video platform-type software, it can be determined that the browsing format of the user in the software is audio format.

[0075] The user browsing content refers to the promotional content released on the software, for example, users can browse short video content released by others on TikTok;

[0076] The advertising placement information refers to the advertisements that are allowed to be placed on the software for promotion. The specific advertising placement information includes personal advertising promotion and advertiser advertising promotion;

[0077] By dividing the software that needs to be advertised into categories such as shopping, social media, social e-commerce, short videos, etc., and determining the user's browsing format (such as pictures, videos, text, audio, etc.) according to different categories of software, the accuracy of advertising can be significantly improved;

[0078] Step 2: Then obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results;

[0079] The specific method of determining the user's browsing content and browsing format in the content information, converting it into text content, identifying the main information in the text content, obtaining summary tags, and collecting the user's browsing data, analyzing the browsing data, and obtaining the user's interest tags according to the analysis results is as follows:

[0080] AS1: Mark the user's browsing content as a browsing target. When the user browses the browsing target in the software, the browsing format of the browsing target is identified. When the identified browsing format is text, its content is text content, and no processing will be performed at this time;

[0081] When the format of the recognized browsing target is a video, the voice in the video is recognized and converted into text to obtain text content. When the format of the recognized browsing target is pure voice, it is directly converted into text to obtain text content.

[0082] AS2: Then recognize the text content and obtain the summary label of the text content based on the overall text content;

[0083] The specific method of obtaining the summary tag of the text content according to the overall text content is:

[0084] AS21: Identify the main information in the text content and reassemble the main information in the text content to obtain an abbreviated text;

[0085] It should be noted that the main information is represented by deleting the modifiers in the text, and the remaining words are the main components. The specific main components in the text include subject, predicate, and object, and the modifiers include attributive, adverbial, complement, etc. In this embodiment, NLP technology is used to identify the text content. The specific identification process is an existing technology and will not be repeated here.

[0086] AS22: Then, the abbreviated text is subjected to label extraction, that is, the number of occurrences of each word in the subject of the abbreviated text is divided by the total number of words in the subject to obtain the frequency of occurrence of the corresponding word, the word with the highest frequency of occurrence is obtained, and it is marked as the keyword of the subject, and the predicate and the object are processed in the same way as the subject to obtain the keywords of the predicate and the object respectively;

[0087] Then, the keywords are recombined and the resulting sentences are used as summary tags of the text content;

[0088] AS3: Based on user authorization, monitor the user's browsing operations and obtain browsing data. The browsing data includes the total browsing time, number of repetitions, and interval duration. The specific method for obtaining browsing data is as follows:

[0089] When the user's browsing target is determined and the user starts browsing, the timing starts from this node and ends when the user exits the browsing target. The time from the start to the end of the timing is marked as the total browsing time;

[0090] When the user views the content in the browsing target, the content in the browsing target is divided according to the node content to obtain the interval content;

[0091] The specific node content is set by professional publishers, and then it is identified whether there is advertising information in the interval content. The specific advertising information includes product names and brand names. When there is advertising information in the interval content, the interval content is marked as first-level content;

[0092] Then obtain the user's browsing speed for each interval content and mark the browsing speed as ,in, Indicates different interval contents;

[0093] It should be noted that the browsing speed here refers to the first browsing speed when the user browses the interval content. When the user repeatedly views the interval content, the browsing speed is only the speed of the first viewing;

[0094] At the same time, the browsing time and the number of repetitions of the interval content by the user are obtained, and the browsing time of the interval content by the user is marked as the interval time. The interval time is the total browsing time of the interval content, specifically including the sum of the time when the user repeatedly views the interval content. The number of repetitions is the number of times the user repeatedly views the interval content after completing browsing.

[0095] AS4: Analyze the browsing data to obtain the user's product interest tags. The specific analysis method of the user's interest tags is as follows:

[0096] The interest value of the interval content is calculated by the following formula calculate:

[0097]

[0098] In the formula, Indicates the duration of the interval. Indicates the total browsing time. is a fixed coefficient, Indicates the number of repetitions of the interval content. and are weight coefficients respectively;

[0099] The interest value of the interval content and the threshold For comparison, The interval content is marked as secondary content, otherwise, The interval content is marked as inertial content, and no processing is performed on the inertial content. The threshold It is set by a person skilled in the art;

[0100] After obtaining the secondary content, first identify whether there is primary content in the secondary content, that is, identify whether there is advertising information in the secondary content. If so, first directly mark its product name and brand name as interest tags;

[0101] If it does not exist, the secondary content is identified at this time, and according to the text described in the secondary content, the secondary content and the summary tag are used as association factors respectively, optical character recognition is used to obtain the associated advertising information, and the product name and brand name in the associated advertising information are used as interest tags. The processing process of optical character recognition is a prior art and will not be described in detail here;

[0102] By collecting user operation data on different browsing content and formats, converting it into a unified text form, and using NLP technology to extract the main information to generate summary tags and user interest tags, it can comprehensively and accurately reflect the user's interest preferences; at the same time, through data analysis of user browsing behavior (including total browsing time, interval duration, number of repetitions, etc.), it can further explore the user's deep interest in specific content; this analysis method not only improves the accuracy of user interest identification, but also provides data support for personalized matching of advertising, effectively avoiding the problem of inefficient delivery caused by generalized interest prediction, and ensuring that the advertising content is highly relevant to user needs;

[0103] Step 3: Match the interest tag with the advertising information to obtain an interest-based advertising set, and determine the best advertisement from the interest-based advertising set;

[0104] The specific method of matching the interest tag with the advertisement delivery information to obtain an interest advertisement set, and determining the best advertisement from the interest advertisement set is:

[0105] BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ;

[0106] Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; ,in, Expressed as the total number of advertising keywords;

[0107] Use natural language processing technology (such as Word2Vec, TF-IDF or Transformer model) to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers;

[0108] BS2: For each interest tag and ad keywords Calculate the similarity and form a The similarity matrix M :

[0109]

[0110] In the formula, Expressed as cosine similarity, measuring interest tags With advertising keywords The matching degree of is the dot product of the vectors, and is the norm of the vector;

[0111] BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ; Among them, the threshold Set up by professional staff;

[0112] BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad :

[0113]

[0114] In the formula, Expressed as the click-through rate of the ad, is the weight coefficient;

[0115] BS5: Select the advertisement with the highest score as the best advertisement;

[0116] By matching user interest tags with keywords in advertising information and using natural language processing technology to convert interest tags and advertising keywords into vectors for cosine similarity calculation, this step can effectively screen out the ads that users are most likely to be interested in and form an interest-based advertising set. At the same time, by further calculating the comprehensive score of the ads and selecting the ads with the highest score as the best ads, this data-driven matching method significantly improves the accuracy and delivery effect of ads. In addition, this process significantly reduces the frequency of low-relevance ads, bringing users a more valuable advertising experience, while optimizing the utilization of advertisers' delivery resources and improving the conversion rate of ads.

[0117] Step 4: Place the best advertisement in the corresponding software.

[0118] Embodiment 2

[0119] In the specific implementation process, this embodiment is based on the first embodiment and is different from the first embodiment in that this embodiment further provides an advertising marketing system based on data analysis, including:

[0120] The information collection module is used to classify the software that needs to be advertised, and determine the browsing format of the user in the software according to the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software;

[0121] The data analysis module is used to obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results;

[0122] An advertisement matching module is used to match the interest tag with the advertisement delivery information to obtain an interest advertisement set and determine the best advertisement from the interest advertisement set;

[0123] An advertisement delivery module, used for delivering the determined best advertisement in the corresponding software;

[0124] In the advertisement matching module, the specific method of matching the interest tag with the advertisement delivery information to obtain an interest advertisement set, and determining the best advertisement from the interest advertisement set is:

[0125] BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ;

[0126] Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; m;

[0127] Use natural language processing technology to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers;

[0128] BS2: For each interest tag and interest-based advertising Calculate the similarity and form a The similarity matrix M :

[0129]

[0130] In the formula, Expressed as cosine similarity, measuring interest tags With advertising The matching degree of is the dot product of the vectors, and is the norm of the vector;

[0131] BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ;

[0132] BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad :

[0133]

[0134] In the formula, Expressed as the click-through rate of the ad, is the weight coefficient;

[0135] BS5: Select the advertisement with the highest score as the best advertisement.

[0136] Embodiment 3

[0137] The specific implementation process of this embodiment includes the entire implementation process of the above two groups of embodiments.

[0138] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0139] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An advertising marketing method based on data analysis, characterized in that: include: Step 1: Categorize the software that needs to be advertised, and determine the browsing format of the user in the software based on the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software; Step 2: Then obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results; Step 3: Match the interest tag with the advertising information to obtain an interest-based advertising set, and determine the best advertisement from the interest-based advertising set; Step 4: Place the best advertisement in the corresponding software; In the step 1, the user browsing content is represented as promotional content published on the software; The advertisement placement information indicates that promotional advertisements are allowed to be placed on the software; In the step 2, the user's browsing content and browsing format in the content information are determined, converted into text content, and the main information in the text content is identified to obtain summary tags. At the same time, the user's browsing data is collected and analyzed. According to the analysis results, the specific method of obtaining the user's interest tags is as follows: AS1: Mark the user's browsing content as a browsing target. When the user browses the browsing target in the software, the browsing format of the browsing target is identified. When the identified browsing format is text, its content is text content, and no processing will be performed at this time; When the format of the recognized browsing target is a video, the voice in the video is recognized and converted into text to obtain text content. When the format of the recognized browsing target is pure voice, it is directly converted into text to obtain text content. AS2: Then recognize the text content and obtain the summary label of the text content based on the overall text content; AS3: Based on user authorization, monitor the user's browsing operations and obtain browsing data. The browsing data includes the total browsing time, number of repetitions, and interval duration. The specific method for obtaining browsing data is as follows: When the user's browsing target is determined and the user starts browsing, the timing starts from this node and ends when the user exits the browsing target. The time from the start to the end of the timing is marked as the total browsing time; When the user views the content in the browsing target, the content in the browsing target is divided according to the node content to obtain the interval content; Then, it is identified whether there is advertising information in the interval content, and the specific advertising information includes product name and brand name. When there is advertising information in the interval content, the interval content is marked as first-level content; Then obtain the user's browsing speed for each interval content and mark the browsing speed as ,in, Indicates different interval contents; At the same time, the browsing time and the number of repetitions of the interval content by the user are obtained, and the browsing time of the interval content by the user is marked as the interval time. The interval time is the total browsing time of the interval content, specifically including the sum of the time when the user repeatedly views the interval content. The number of repetitions is the number of times the user repeatedly views the interval content after completing browsing. AS4: Analyze the browsing data to obtain the user's product interest tags. The specific analysis method of the user's interest tags is as follows: The interest value of the interval content is calculated by the following formula calculate: In the formula, Indicates the duration of the interval. Indicates the total browsing time. is a fixed coefficient, Indicates the number of repetitions of the interval content. and are weight coefficients respectively; The interest value of the interval content and the threshold For comparison, The interval content is marked as secondary content, otherwise, The interval content is marked as inertial content, and no processing is performed on the inertial content.

2. The advertising and marketing method based on data analysis according to claim 1, characterized in that: In step AS3, the specific method of obtaining the summary tag of the text content according to the overall text content is: AS21: Identify the main information in the text content and reassemble the main information in the text content to obtain an abbreviated text; AS22: Then, the abbreviated text is subjected to label extraction, that is, the number of occurrences of each word in the subject of the abbreviated text is divided by the total number of words in the subject to obtain the frequency of occurrence of the corresponding word, the word with the highest frequency of occurrence is obtained, and it is marked as the keyword of the subject, and the predicate and the object are processed in the same way as the subject to obtain the keywords of the predicate and the object respectively; The keywords are then recombined and the resulting sentences are used as summary tags for the text content.

3. The advertising and marketing method based on data analysis according to claim 1, characterized in that: In the step 3, the specific method of matching the interest tag with the advertisement delivery information to obtain an interest advertisement set, and determining the best advertisement from the interest advertisement set is: BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ; Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; ,in, Expressed as the total number of advertising keywords; Use natural language processing technology to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers; BS2: For each interest tag and ad keywords Calculate the similarity and form a The similarity matrix M : In the formula, Expressed as cosine similarity, measuring interest tags With advertising keywords The matching degree of is the dot product of the vectors, and is the norm of the vector; BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ; BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad : In the formula, Expressed as the click-through rate of the ad, is the weight coefficient; BS5: Select the advertisement with the highest score as the best advertisement.

4. An advertising marketing system based on data analysis, characterized in that: include: The information collection module is used to classify the software that needs to be advertised, and determine the browsing format of the user in the software according to the classification of the software; at the same time, based on the user's authorization, collect the content information of the user in the software, determine the user's browsing content, and determine the advertising information in the software; The data analysis module is used to obtain the user's content information, determine the user's browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, and collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags based on the analysis results; An advertisement matching module, used to match interest tags with advertisement delivery information to obtain an interest advertisement set, and determine the best advertisement from the interest advertisement set; An advertisement delivery module, used for delivering the determined best advertisement in the corresponding software; In the work content of the information collection module, the user browsing content is represented as the promotional content published on the software; The advertisement placement information indicates that promotional advertisements are allowed to be placed on the software; The working content of the data analysis module is to determine the user browsing content and browsing format in the content information, convert it into text content, identify the main information in the text content, obtain summary tags, collect the user's browsing data, analyze the browsing data, and obtain the user's interest tags according to the analysis results in the following specific ways: AS1: Mark the user's browsing content as a browsing target. When the user browses the browsing target in the software, the browsing format of the browsing target is identified. When the identified browsing format is text, its content is text content, and no processing will be performed at this time; When the format of the recognized browsing target is a video, the voice in the video is recognized and converted into text to obtain text content. When the format of the recognized browsing target is pure voice, it is directly converted into text to obtain text content. AS2: Then recognize the text content and obtain the summary label of the text content based on the overall text content; AS3: Based on user authorization, monitor the user's browsing operations and obtain browsing data. The browsing data includes the total browsing time, number of repetitions, and interval duration. The specific method for obtaining browsing data is as follows: When the user's browsing target is determined and the user starts browsing, the timing starts from this node and ends when the user exits the browsing target. The time from the start to the end of the timing is marked as the total browsing time; When the user views the content in the browsing target, the content in the browsing target is divided according to the node content to obtain the interval content; Then, it is identified whether there is advertising information in the interval content, and the specific advertising information includes product name and brand name. When there is advertising information in the interval content, the interval content is marked as first-level content; Then obtain the user's browsing speed for each interval content and mark the browsing speed as ,in, Indicates different interval contents; At the same time, the browsing time and the number of repetitions of the interval content by the user are obtained, and the browsing time of the interval content by the user is marked as the interval time. The interval time is the total browsing time of the interval content, specifically including the sum of the time when the user repeatedly views the interval content. The number of repetitions is the number of times the user repeatedly views the interval content after completing browsing. AS4: Analyze the browsing data to obtain the user's product interest tags. The specific analysis method of the user's interest tags is as follows: The interest value of the interval content is calculated by the following formula calculate: In the formula, Indicates the duration of the interval. Indicates the total browsing time. is a fixed coefficient, Indicates the number of repetitions of the interval content. and are weight coefficients respectively; The interest value of the interval content and the threshold For comparison, The interval content is marked as secondary content, otherwise, The interval content is marked as inertial content, and no processing is performed on the inertial content.

5. The advertising and marketing system based on data analysis according to claim 4, characterized in that: In the advertisement matching module, the specific method of matching the interest tag with the advertisement delivery information to obtain an interest advertisement set, and determining the best advertisement from the interest advertisement set is: BS1: Record all interest tags as : ;in, Expressed as the total number of interest tags; ; Extract the keywords of each advertisement in the advertisement delivery information to form an advertisement keyword set, and record the advertisement keyword set as , ; ; Use natural language processing technology to convert interest tags and advertising keywords into d-dimensional vectors and store them in and in; and , ,in, Represented as the field of real numbers; BS2: For each interest tag and interest-based advertising Calculate the similarity and form a The similarity matrix M : In the formula, Expressed as cosine similarity, measuring interest tags With advertising The matching degree of is the dot product of the vectors, and is the norm of the vector; BS3: Select the cosine similarity greater than the threshold Ad keywords As a candidate, the advertising keywords The corresponding advertisements are constructed as interest-based advertising sets ; BS4: Interest-based advertising , the interest-based advertising set is calculated using the following formula The overall score of each ad : In the formula, Expressed as the click-through rate of the ad, is the weight coefficient; BS5: Select the advertisement with the highest score as the best advertisement.

Images