Internet advertisement intelligent publishing system and method

Abstract

The invention belongs to the technical field of Internet advertisements, and particularly relates to an intelligent publishing system and method for Internet advertisements, and the system comprises an advertisement resource unit, an embedded calculation unit and a publishing unit. The advertisement resource unit comprises an advertisement resource packet and a hyperlink packet, the advertisement resource packet comprises a plurality of advertisement resources, and the hyperlink packet comprises a plurality of hyperlinks; the embedded calculation unit comprises a connection module, an embedded module and a conversion module; the connection module classifies the advertisement resources according to the advertisement resource contents and generates advertisement content keywords at the same time; meanwhile, the connection module establishes a jump contact between the advertisement resources and the hyperlink; the publishing unit comprises a target resource and a display terminal, and the target resource displays content through the display terminal; the embedding module is used for reading and retrieving the content of the target resource and matching the advertisement resource with the target resource according to the keyword of the advertisement content; through unit cooperation, accurate advertisement pushing can be achieved, the participation degree and the click rate of users can be improved, and then the conversion rate is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of Internet advertisements, and in particular to an Internet advertisement intelligent publishing system and method. Background Art

[0002] Internet advertising refers to commercial advertising that directly or indirectly promotes goods or provides services in the form of text, pictures, audio, video or other forms through Internet media such as websites, web pages, and Internet applications.

[0003] A Chinese patent application CN102799997B discloses a method and system for publishing Internet advertisements. The purpose of the present invention is to provide a method and system for publishing Internet advertisements, which allows any publisher to publish advertisements on the media through the system on the Internet; at the same time, any media can sign a contract through the system on the Internet and become a contracted media. It is characterized in that: the system includes media and an advertisement publishing system, the media is implanted with an embedded module, the embedded module at least includes a display unit, the advertisement publishing system includes an advertisement publisher registration module, provides an advertisement publisher with an advertisement fee payment account, and the advertisement publishing system and the embedded module exchange and transmit data. The beneficial effect of the present invention is that for a large number of free software, the advertisement publishing system is accessed through the embedded module to form an advertising business and obtain advertising revenue. The free software here includes PC software and mobile terminal software, that is, client software.

[0004] Although the above technology forms an advertising business and obtains advertising revenue, the free software here includes PC software and mobile terminal software, namely client software. However, in the actual publishing process, the conversion rate of advertisements is low because the advertisements are directly embedded in the media, resulting in abrupt advertising content, reducing the conversion rate and even causing user disgust.

[0005] To this end, the present invention provides an Internet advertisement intelligent publishing system and method. Summary of the invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve its technical problem is: an Internet advertisement intelligent publishing system and method of the present invention comprises an advertisement resource unit, an embedded calculation unit and a publishing unit;

[0008] The advertising resource unit includes an advertising resource package and a hyperlink package, wherein the advertising resource package includes a plurality of advertising resources and the hyperlink package includes a plurality of hyperlinks;

[0009] The embedding computing unit includes a connection module, an embedding module and a conversion module;

[0010] The connection module classifies the advertising resources according to the content of the advertising resources and generates keywords for the advertising content at the same time;

[0011] At the same time, the connection module establishes the connection between the advertising resources and the hyperlink jump;

[0012] The publishing unit includes a target resource and a display terminal, and the target resource displays content through the display terminal;

[0013] The embedding module reads and retrieves the content of the target resource, matches the advertising resource to the target resource according to the keywords of the advertising content, and embeds the advertising resource and the corresponding hyperlink into the target resource;

[0014] The conversion module establishes a hyperlink jump command, and the display terminal jumps to the target interface after executing the jump command.

[0015] A method for intelligently publishing Internet advertisements, which is applied to the above-mentioned intelligent publishing system for Internet advertisements, wherein the step of classifying advertisement resources by the connection module comprises:

[0016] A1. Classify the advertising resources into multiple sub-items according to the formats of the advertising resources;

[0017] A2. Classify the advertisements in the sub-projects according to their purpose;

[0018] A3. Classify the purpose-classified advertisements according to the time periods they are suitable for.

[0019] Preferably, the step of generating advertisement content keywords by the connection module includes:

[0020] B1. Remove noise and stop words from the advertisement text in the advertisement resources, and then perform word segmentation on the text;

[0021] B2. Generate candidate keywords based on the frequency of each segmentation;

[0022] B3. Use the N-gram model to continuously combine n candidate keywords into advertising content keywords;

[0023] Among them, n=1, 2, …, n.

[0024] Preferably, the method of matching the keywords of the advertisement content with the target resources is:

[0025] Obtaining the overlap number X: Processing the target resource text, based on the word overlap, comparing the keywords of the advertisement content with the words in the target resource one by one, and calculating the overlap number X;

[0026] Compare the overlap quantity values: establish a benchmark value Y for the overlap quantity, and compare the overlap quantity with the benchmark value;

[0027] The keywords of the advertisement content in the part X≥Y are matched with the corresponding target resources;

[0028] X<Y part of the advertising content keywords:

[0029] Convert to word vectors: Convert the keywords of the advertisement content and the target resource text into word vectors, and calculate the matching degree M between the keywords of the advertisement content and the target resource text;

[0030] Match the remaining part: establish a comparison benchmark P for the matching degree, compare the matching degree and the comparison benchmark, and match the remaining advertising content keywords with the target resource text based on the comparison results.

[0031] Preferably, the step of calculating the matching degree between the advertisement content keywords and the target resource text includes:

[0032] C1. Encode the keywords of the advertisement content and the target resource text through the pre-trained language model through BERT.

[0033] C2. Define the semantic vectors corresponding to the keywords in the advertisement content and the semantic vectors corresponding to the target resource text as;

[0034] C3. Calculate the matching degree based on the semantic vector.

[0035] Preferably, the matching degree is calculated in step C3 as follows:

[0036]

[0037] Among them, kt is the vector dot product, and |k||t| is the modulus of the vector.

[0038] Preferably, the title of the time period classification in step A3 is T, where i=1, 2, ..., i;

[0039] According to the titles classified by time period, the matching degree between the keywords of the advertisement content and the target resource text is calculated respectively, and the data relationship of the time matching degree is obtained;

[0040] The steps of calculating the time matching data relationship include:

[0041] D1. Collect time period data and matching data;

[0042] D2. Set the function expression;

[0043] D3. Derive function expressions.

[0044] Preferably, what is collected in step D1 is:

[0045] (t 1 ,y 1 )、(t 2 ,y 2)…(t i ,y i );

[0046] Among them, t is the time corresponding to the title, y is the matching degree, and i=1, 2, ..., i.

[0047] Preferably, in step D2, the function expression set is:

[0048] y=mt+b, where b is the matching degree at the initial moment.

[0049] Preferably, the specific derivation method of the function expression in step D3 is:

[0050]

[0051] in,

[0052] After substitution, derive the function expression.

[0053] The beneficial effects of the present invention are as follows:

[0054] 1. The present invention reads and retrieves the content of the target resource through an embedding module, matches the advertising resource with the target resource according to the keywords of the advertising content, and embeds the advertising resource and the corresponding hyperlink into the target resource. Keyword extraction can help users locate relevant information more quickly. By embedding advertisements related to user interests and behaviors in the target resource, the advertising content can be highly matched with the needs, interests and context of the target user, thereby enhancing the relevance of the advertisement. This precise advertising push can increase user participation and click-through rate, thereby increasing conversion rate.

[0055] 2. The present invention calculates the matching degree between the keywords of the advertisement content and the text of the target resource according to the titles classified by time periods, and obtains the data relationship of the time matching degree. In different time periods, the corresponding advertisement content is published to the target resource according to the matching degree, and relevant products are provided at the most appropriate time. Through accurate timing selection, the advertisement can more effectively attract the user's attention, enhance the purchase intention, reduce interference, and improve the response rate, thereby ultimately enhancing the sales conversion rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The present invention will be further described below in conjunction with the accompanying drawings.

[0057] Figure 1 It is a framework diagram of the advertising intelligent publishing system of the present invention;

[0058] Figure 2 is a flow chart of advertising resource classification in the present invention;

[0059] Figure 3 is a flow chart of generating keywords for advertisement content in the present invention;

[0060] Figure 4 It is a flow chart of calculating the matching degree between the keywords of the advertisement content and the target resource text in the present invention;

[0061] Figure 5 It is a flow chart of calculating the time matching degree data relationship in the present invention. DETAILED DESCRIPTION

[0062] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0063] like Figures 1 to 5 As shown, the present invention includes an advertising resource unit, an embedding calculation unit, and a publishing unit;

[0064] The advertising resource unit includes an advertising resource package and a hyperlink package, wherein the advertising resource package includes a plurality of advertising resources and the hyperlink package includes a plurality of hyperlinks;

[0065] The embedding computing unit includes a connection module, an embedding module and a conversion module;

[0066] The connection module classifies the advertising resources according to the content of the advertising resources and generates keywords for the advertising content at the same time;

[0067] A method for intelligently publishing Internet advertisements, which is applied to the above-mentioned intelligent publishing system for Internet advertisements, wherein the step of classifying advertisement resources by the connection module comprises:

[0068] A1. Classify advertising resources into multiple sub-items according to the format of advertising resources; among them, the common sub-item classification method is image resources, video resources, text resources, and audio resources, which are classified according to their dissemination methods.

[0069] A2. Classify the advertisements within the sub-projects according to their advertising purposes. Common advertising purposes include brand promotion, product sales, user conversion, traffic guidance, etc.

[0070] A3. Classify purpose-classified advertisements according to the time periods they are suitable for. Among them, various types of advertisements are adaptable to specific time periods, such as winter down jacket sales advertisements, late-night dark circle removal product sales advertisements, etc. When such advertisements are pushed, they are adapted to the corresponding release time, and can provide relevant products at the most appropriate time based on the specific needs and behavior patterns of users. Through precise timing selection, advertisements can more effectively attract user attention, enhance purchasing desire, reduce interference, and increase response rate, thereby ultimately enhancing sales conversion.

[0071] Meanwhile, the connection module establishes the connection between the advertising resources and the hyperlink jump. For common advertisements, after the advertisement content is released, it is necessary to jump to the corresponding interface through the hyperlink of the advertisement content. For example, in product sales advertisements, the hyperlink jumps to the corresponding product sales interface on the online shopping platform, and the traffic guidance duration connection jumps to the corresponding APP.

[0072] The publishing unit includes the target resource and the display terminal. The target resource displays the content through the display terminal. Common display terminals include smartphones, tablets, and computers, etc.

[0073] The embedding module reads and retrieves the content of the target resource, matches the advertising resources with the target resource according to the keywords in the advertisement content, and embeds the advertising resources and the corresponding hyperlinks into the target resource. Keyword extraction can help users locate relevant information more quickly. By embedding advertisements related to users' interests and behaviors in the target resource, the advertisement content can be highly matched with the needs, interests, and context of the target users, enhancing the relevance of the advertisement. This precise advertisement push can improve user engagement and click-through rate, and thus increase the conversion rate.

[0074] The form of native advertisements integrates the advertisement content with the natural content of the page, making it visually unobtrusive and easier for users to accept. Compared with traditional display advertisements (such as pop-up advertisements, banner advertisements, etc.), native advertisements do not cause too much interference to the user experience. Users are more willing to accept the advertisement content, enhancing brand affinity, thereby increasing the adhesion of the advertisement to users and reducing users' aversion to advertisements.

[0075] The conversion module establishes a hyperlink jump command. After the display terminal executes the jump command, it jumps to the target interface. Taking a smartphone as an example, common jump commands include click, shake, etc.

[0076] The steps for the connection module to generate keywords for the advertisement content include:

[0077] B1. Remove noise from the advertisement text in the advertising resources, remove stop words, and then segment the text;

[0078] Remove noise and clean up irrelevant information in the advertisement text, such as HTML tags, special characters, symbols in the advertisement, etc.;

[0079] Remove stop words. Stop words refer to those words that appear frequently but contribute little to text analysis. For example, "de", "shi", "zai" in Chinese, and "the", "and", "is" in English, etc.

[0080] B2. Generate candidate keywords according to the frequency of each word segment;

[0081] B3. Continuously combine n candidate keywords into keywords for the advertisement content through the N-gram model;

[0082] Among them, n=1, 2, …, n.

[0083] N-gram models are often used to predict or generate the next word in a text. Given the first N words, the model can predict the next most likely word (the programming and specific operation steps of the N-gram model are well known to those skilled in the art, so I will not go into detail here).

[0084] The way to match the keywords of the advertising content with the target resources is:

[0085] Obtaining the overlap number X: Processing the target resource text, based on the word overlap, comparing the keywords of the advertisement content with the words in the target resource one by one, and calculating the overlap number X;

[0086] Compare the overlap quantity values: establish a benchmark value Y for the overlap quantity, and compare the overlap quantity with the benchmark value;

[0087] The keywords of the advertisement content in the part X≥Y are matched with the corresponding target resources;

[0088] Matching is performed by comparing overlapping parts, and the processing method is intuitive and quick. As a preliminary operation during matching, some advertising content keywords fail to match the target resources after the operation.

[0089] X<Y part of the advertising content keywords:

[0090] Convert to word vectors: Convert the keywords of the advertisement content and the target resource text into word vectors, and calculate the matching degree M between the keywords of the advertisement content and the target resource text;

[0091] The steps for calculating the matching degree between the keywords of the advertisement content and the target resource text include:

[0092] C1. Encode the advertising content keywords and target resource text through BERT;

[0093] C2, define the semantic vectors corresponding to the keywords in the advertising content and the semantic vectors corresponding to the target resource text; the semantic vectors corresponding to the keywords in the advertising content The semantic vector corresponding to the target resource text is

[0094] C3. Based on semantic vector and Calculate the matching degree.

[0095] The method for calculating the matching degree in step C3 is:

[0096]

[0097] Among them, kt is the vector dot product, and |k||t| is the modulus of the vector.

[0098] Match the remaining part: establish a comparison benchmark P for the matching degree, compare the matching degree and the comparison benchmark, and match the remaining advertising content keywords with the target resource text based on the comparison results.

[0099] Matching is completed by calculating the matching degree, which improves the relevance of the advertising content and the target resource content. When the advertisement is embedded in the content that the user already trusts and likes, the user is more likely to accept the advertisement and think that the advertisement itself is valuable. This increase in trust can enhance the effectiveness of the advertisement, thereby increasing the click-through rate, interaction rate and ultimate purchase conversion.

[0100] At the same time, the matched ads are embedded in the target resources. Since the ad content is closely related to the content that users are interested in, users are more likely to click, interact, and even share the ad content, further expanding the ad dissemination effect.

[0101] The title of the time period classification in step A3 is T, where i=1, 2, ..., i, such as the title is divided by hours.

[0102] According to the titles classified by time period, the matching degree between the keywords of the advertising content and the text of the target resource is calculated respectively, and the time matching data relationship is obtained. In different time periods, the corresponding advertising content is published to the target resource according to the matching degree, and relevant products are provided at the most appropriate time. Through precise timing selection, advertising can more effectively attract users' attention, enhance purchasing intention, reduce interference, and increase response rate, thereby ultimately enhancing sales conversion rate.

[0103] The steps of calculating the time matching data relationship include:

[0104] D1. Collect time period data and matching data;

[0105] What is collected in step D1 is:

[0106] (t 1 ,y 1 )、(t 2 ,y 2 )…(t i ,y i );

[0107] Among them, t is the time corresponding to the title, y is the matching degree, and i=1, 2, ..., i.

[0108] D2. Set the function expression;

[0109] In step D2, the function expression is set as:

[0110] y=mt+b, where b is the matching degree at the initial moment

[0111] D3. Derive function expressions;

[0112] The specific derivation method of the function expression in step D3 is:

[0113]

[0114] in,

[0115] After substitution, the function expression is derived (the function expression is derived through the least squares method, which will not be elaborated here). By deriving the function expression and substituting it into the calculation, the advertising content with the corresponding matching degree with the target resource can be pushed within the corresponding time period, and then the push time is used as the basis for the advertising content to improve the user conversion rate.

[0116] The above-mentioned front, back, left, right, top and bottom are all based on the figures in the specification. Figure 1 As a benchmark, according to the person's observation perspective, the side of the device facing the observer is defined as the front, the left side of the observer is defined as the left, and so on.

[0117] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the scope of protection of the present invention.

[0118] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An Internet advertising intelligent publishing system, characterized by: It includes an advertising resource unit, an embedding calculation unit and a publishing unit; The advertising resource unit includes an advertising resource package and a hyperlink package, wherein the advertising resource package includes a plurality of advertising resources, and the hyperlink package includes a plurality of hyperlinks; The embedding computing unit includes a connection module, an embedding module and a conversion module; The connection module classifies the advertising resources according to the content of the advertising resources and generates keywords for the advertising content at the same time; At the same time, the connection module establishes the connection between the advertising resources and the hyperlink jump; The publishing unit includes a target resource and a display terminal, and the target resource displays content through the display terminal; The embedding module reads and retrieves the content of the target resource, matches the advertising resource to the target resource according to the keywords of the advertising content, and embeds the advertising resource and the corresponding hyperlink into the target resource; The conversion module establishes a hyperlink jump command, and the display terminal jumps to the target interface after executing the jump command.

2. A method for intelligently publishing Internet advertisements, which is applied to the intelligent publishing system for Internet advertisements according to claim 1, characterized in that: The step of classifying the advertising resources by the connection module includes: A1. Classify the advertising resources into multiple sub-items according to the formats of the advertising resources; A2. Classify the advertisements in the sub-projects according to their purpose; A3. Classify the purpose-classified advertisements according to the time periods they are suitable for.

3. The method for intelligently publishing Internet advertisements according to claim 2, characterized in that: The step of generating advertisement content keywords by the connection module includes: B1. Remove noise and stop words from the advertisement text in the advertisement resources, and then perform word segmentation on the text; B2. Generate candidate keywords based on the frequency of each segmentation; B3. Use the N-gram model to continuously combine n candidate keywords into advertising content keywords; Among them, n=1, 2, …, n.

4. The method for intelligently publishing Internet advertisements according to claim 3, characterized in that: The way to match the keywords of the advertising content with the target resources is: Obtaining the overlap number X: Processing the target resource text, based on the word overlap, comparing the keywords of the advertisement content with the words in the target resource one by one, and calculating the overlap number X; Compare the overlap quantity values: establish a benchmark value Y for the overlap quantity, and compare the overlap quantity with the benchmark value; The keywords of the advertisement content in the part X≥Y are matched with the corresponding target resources; X<Y part of the advertising content keywords: Convert to word vectors: Convert the keywords of the advertisement content and the target resource text into word vectors, and calculate the matching degree M between the keywords of the advertisement content and the target resource text; Match the remaining part: establish a comparison benchmark P for the matching degree, compare the matching degree and the comparison benchmark, and match the remaining advertising content keywords with the target resource text based on the comparison results.

5. The method for intelligently publishing Internet advertisements according to claim 4, characterized in that: The steps for calculating the matching degree between the keywords of the advertisement content and the target resource text include: C1. Encode the advertising content keywords and target resource text through BERT; C2, define the semantic vectors corresponding to the keywords in the advertisement content and the semantic vectors corresponding to the target resource text; C3. Calculate the matching degree based on the semantic vector.

6. The method for intelligently publishing Internet advertisements according to claim 5, characterized in that: The method for calculating the matching degree in step C3 is: Among them, kt is the vector dot product, and |k||t| is the modulus of the vector.

7. The method for intelligently publishing Internet advertisements according to claim 6, characterized in that: The title of the time period classification in step A3 is T, where i=1, 2, ..., i; According to the titles classified by time period, the matching degree between the keywords of the advertisement content and the target resource text is calculated respectively, and the data relationship of the time matching degree is obtained; The steps of calculating the time matching data relationship include: D1. Collect time period data and matching data; D2. Set the function expression; D3. Derive function expressions.

8. The method for intelligently publishing Internet advertisements according to claim 7, characterized in that: What is collected in step D1 is: (t1,y1)、(t2,y2)…(t i ,y i ); Among them, t is the time corresponding to the title, y is the matching degree, and i=1, 2, ..., i.

9. The method for intelligently publishing Internet advertisements according to claim 8, characterized in that: In step D2, the function expression is set as: y=mt+b, where b is the matching degree at the initial moment.

10. The method for intelligently publishing Internet advertisements according to claim 9, characterized in that: The specific derivation method of the function expression in step D3 is: in, After substitution, derive the function expression.

Images