An intelligent matching system and method for advertising visual elements based on big data

Abstract

The present invention discloses an intelligent matching system and method for advertising visual elements based on big data, which relates to the technical field of big data advertising analysis. The present invention obtains user historical advertising viewing record data, and based on user personalized screening and analysis, determines a periodic user characteristic advertising data set; extracts the corresponding advertising texture element features, constructs a texture element feature path for the corresponding periodic user characteristic advertising data, analyzes the corresponding advertising data score, and determines the optimal periodic user characteristic advertising data; uses the texture element feature path nodes of the optimal periodic user characteristic advertising data as reference nodes, respectively performs clustering processing on the texture element feature path nodes corresponding to the periodic user characteristic advertising data, analyzes the adapted texture element features of the texture element feature path nodes of each optimal user characteristic advertising data, and comprehensively determines the recommended score range of the adapted advertising data for the corresponding user. The present invention can achieve precise user placement of advertisements and reasonable matching of elements.

Description

Technical Field

[0001] The present invention relates to the field of big data advertising analysis, and specifically to an intelligent matching system and method for advertising visual elements based on big data. Background Art

[0002] With the rapid development of the Internet, the amount of data generated every day has increased explosively. These massive data cover rich information such as consumers' browsing behaviors, purchase records, and interest preferences. For the advertising industry, traditional advertising placement and visual design methods are difficult to fully explore the value behind these data and cannot accurately meet the personalized needs of different consumers. At the same time, in an environment of information overload, consumers' attention to advertisements is becoming increasingly scattered, which requires the visual element matching of advertisements to be able to attract users' attention more accurately for precise placement. And currently, the phenomenon of user advertisement pushing mostly involves a large number of placements and multiple repeated typifications, which is likely to cause visual fatigue for users due to the random placement of advertisements and poor experience in data browsing; there is a lack of precise user placement of advertisements and reasonable element matching. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent matching system and method for advertising visual elements based on big data to solve the problems raised in the prior art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions:

[0005] An intelligent matching method for advertising visual elements based on big data, the method comprising the following steps:

[0006] Obtain user historical advertisement viewing record data, determine the analysis period, construct an advertisement viewing data set for the corresponding period, and based on personalized screening and analysis of users, determine the advertisement data set of user characteristics for the period;

[0007] Extract the corresponding advertisement texture element features based on the advertisement data set of user characteristics for the period, construct a texture element feature path for the advertisement data of user characteristics for the corresponding period, and analyze the score of the advertisement data of user characteristics for the corresponding period to determine the best advertisement data of user characteristics for the period; taking the texture element feature path nodes of the best advertisement data of user characteristics for the period as reference nodes, respectively perform clustering processing on the texture element feature path nodes corresponding to the advertisement data of user characteristics for the period, analyze the adaptable texture element features of the texture element feature path nodes of each best advertisement data of user characteristics, and comprehensively determine the recommended score range of the adaptable advertisement data for the corresponding user.

[0008] Further, by calling the user's historical advertisement viewing record database, an analysis period window is determined, the historical advertisement viewing record data of the user within the corresponding period is retrieved, and a corresponding period advertisement viewing data set is constructed; wherein, the user's historical advertisement viewing record database includes a website advertisement viewing record database and an application advertisement viewing record database; the period advertisement viewing data set includes the viewing duration data, advertisement type data, and corresponding advertisement number data of the corresponding advertisement user.

[0009] According to the personalized screening analysis of the average viewing duration of each advertisement data in the period advertisement viewing data set of the corresponding user, the advertisement data with a viewing duration less than the average viewing duration is screened out, and vice versa, and the period user characteristic advertisement data set after the personalized screening analysis is determined.

[0010] Further, according to each period user characteristic advertisement data in the period user characteristic advertisement data set, the texture element features of the corresponding period user characteristic advertisement data are extracted respectively, corresponding texture element feature nodes are constructed, and node connection is performed to construct the texture element feature path of the corresponding period user characteristic advertisement data; wherein the texture element features of the advertisement data include text type texture element features, graphic type texture element features, color type texture element features, etc., and for text type, graphic type, and color type, specific texts are further subdivided, such as Chinese, English, etc.; for graphic type, it is further subdivided into rectangle, circle, etc.; for color type, it is further subdivided into red, blue, etc.; according to the texture element feature paths of the corresponding period user characteristic advertisement data, scoring analysis is performed on each period user characteristic advertisement data respectively; the texture element feature node feature matrix of each period user characteristic advertisement data is constructed respectively; the texture element feature node feature matrix is composed of the proportion weight of the advertisement data of the corresponding texture element feature and the type experience score; wherein the type experience score of the texture element feature is the reference score of the corresponding advertisement texture element feature given by big data; based on the texture element feature node feature matrix of the corresponding period user characteristic advertisement data, texture element feature node scoring analysis is performed, and its calculation formula is

[0011]

[0012] Among them, P[W,Q]n is the scoring data of the texture element feature node n in the corresponding period user characteristic advertisement data; W is the proportion weight of the advertisement data of the texture element feature node n; Q is the advertisement data type experience scoring data of the texture element feature node n.

[0013] According to the texture element feature node scoring data of the corresponding period user characteristic advertisement data, the sum of the node scores on the texture element feature path of the corresponding period user characteristic advertisement data is calculated to determine the scoring data of the corresponding period user characteristic advertisement data.

[0014] Further, based on the scoring data of the user characteristic advertisement data for each period, the maximum value of the scoring data is output as the best user characteristic advertisement data for the period; then, taking each texture element feature node on the texture element feature path corresponding to the best user characteristic advertisement data for the period as a reference node, node clustering processing is performed on the texture element feature nodes corresponding to the remaining user characteristic advertisement data for each period in the user characteristic advertisement data set for the period; the remaining user characteristic advertisement data for each period refers to the remaining user characteristic advertisement data for the period in the user characteristic advertisement data set for the period after removing the best user characteristic advertisement data for the period; according to the result of the node clustering processing, the clustering adaptation state of the texture element feature nodes corresponding to the best user characteristic advertisement data for the period corresponding to each reference node and the texture element feature nodes corresponding to the remaining user characteristic advertisement data for each period is determined;

[0015] Construct concentric circles with each reference node as the center, and perform density analysis on the texture element feature nodes corresponding to the remaining user characteristic advertisement data in each concentric circle; by gradually increasing the radius of the concentric circle and calculating the ratio of the number of nodes to the area in the concentric circle corresponding to each radius respectively, determine the node density value of the concentric circle corresponding to each radius, and take the concentric circle corresponding to the maximum node density value as the best - fitting node concentric circle of the corresponding reference node; based on the best - fitting node concentric circles corresponding to the texture element feature nodes of the periodic best user characteristic advertisement data respectively, overall plan the types and scores of the texture element feature nodes in the concentric circles divided by the best - fitting nodes. If there are two or more texture element feature nodes of the same type, classify and screen the texture element feature nodes of the same type, and respectively retain the maximum and minimum scores of the texture element feature nodes of the same type; in the best - fitting node concentric circles where the texture element feature nodes of the periodic best user characteristic advertisement data are located respectively, perform a full traversal of the different - type texture element feature nodes starting from the center of the circle. For the existence of texture element feature nodes of the same type, perform a branch traversal respectively, and output the traversal paths of the maximum - value nodes and the minimum - value nodes in the current best - fitting node concentric circle; since after classifying and screening the texture element feature nodes of the same type, only the maximum - score node and the minimum - score node of the same type are retained, through branch traversal, it can be ensured that the types of the texture element feature nodes on a traversal path are different; calculate the sum of the node scores of the maximum - value node traversal path and the minimum - value node traversal path respectively, determine the upper and lower limits of the adapted advertisement recommendation scores for the texture element feature nodes corresponding to the current periodic best user characteristic advertisement data respectively, and construct an adapted advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adapted advertisement recommendation score intervals of the nodes on the paths of the texture element feature nodes corresponding to the current periodic best user characteristic advertisement data, sum and average the upper and lower limits of each interval respectively, output the sum - average value of the upper limits after calculation as the upper limit of the periodic comprehensive adapted advertisement recommendation score interval, and output the sum - average value of the lower limits after calculation as the lower limit of the periodic comprehensive adapted advertisement recommendation score interval, then construct a comprehensive adapted advertisement recommendation score interval corresponding to the current user cycle.

[0016] Furthermore, based on the comprehensive adapted advertisement recommendation score interval within the user cycle, calculate the scores of the pushed advertisement data and perform screening; if the score of the pushed advertisement data is within the comprehensive adapted advertisement recommendation score range, push it to the user; otherwise, screen the advertisement data.

[0017] Implement a cycle self - update loop. By allowing the user to set the update time themselves, update and re - analyze the comprehensive adapted advertisement recommendation score interval of the cycle.

[0018] An intelligent matching system for advertising visual elements based on big data, the system includes an advertising record acquisition module, a texture element feature processing module, an advertising adaptation recommendation analysis module, and a feedback update module;

[0019] The advertising record acquisition module acquires user historical advertising viewing record data, determines the analysis period, constructs an advertising viewing data set for the corresponding period, and based on user personalized screening analysis, determines a periodic user characteristic advertising data set; the texture element feature processing module extracts corresponding advertising texture element features based on the periodic user characteristic advertising data set, constructs a texture element feature path for the corresponding periodic user characteristic advertising data, and analyzes the score of the corresponding periodic user characteristic advertising data to determine the best periodic user characteristic advertising data; the advertising adaptation recommendation analysis module uses the texture element feature path nodes of the best periodic user characteristic advertising data as reference nodes, respectively performs clustering processing on the texture element feature path nodes of the corresponding periodic user characteristic advertising data, analyzes the adapted texture element features of the texture element feature path nodes of each best user characteristic advertising data, and comprehensively determines the recommended score range of the adapted advertising data for the corresponding user; the feedback update module makes a push judgment on the pushed advertisements based on the recommended score range of the adapted advertising data, and performs cyclic update processing on the recommended score range of the periodic adapted advertising data.

[0020] Further, the advertising record acquisition module includes an advertising data acquisition unit and an advertising data personalized processing unit;

[0021] The advertising data acquisition unit determines the analysis period window by calling the user's historical advertising viewing record database, retrieves the user's historical advertising viewing record data within the corresponding period, and constructs an advertising viewing data set for the corresponding period; the advertising viewing data set for the corresponding period includes the viewing duration data, advertising type data, and corresponding advertising number data of the corresponding advertising users;

[0022] The advertising data personalized processing unit performs personalized screening analysis based on the average viewing duration of each advertising data in the periodic advertising viewing data set of the corresponding user; determines the periodic user characteristic advertising data set after personalized screening analysis.

[0023] Further, the texture element feature processing module includes an advertising texture element path construction unit and an advertising data score analysis unit;

[0024] The advertising texture element path construction unit extracts the texture element features of the corresponding periodic user characteristic advertising data respectively according to the periodic user characteristic advertising data in the periodic user characteristic advertising data set, constructs corresponding texture element feature nodes, and performs node connection to construct a texture element feature path for the corresponding periodic user characteristic advertising data;

[0025] The advertisement data scoring and analysis unit performs scoring and analysis on the advertisement data of user characteristics in each period respectively according to the texture element feature paths corresponding to the advertisement data of user characteristics in each period; it constructs the feature matrices of the corresponding texture element feature nodes for the texture element feature nodes of the advertisement data of user characteristics in each period respectively; the feature matrix of the texture element feature node is composed of the proportion weight of the advertisement data corresponding to the texture element feature and the type experience score; based on the feature matrix of the texture element feature node corresponding to the advertisement data of user characteristics in each period, the scoring and analysis of the texture element feature node is carried out; according to the scoring data of the texture element feature node corresponding to the advertisement data of user characteristics in each period, the sum of the scores of the nodes on the texture element feature path corresponding to the advertisement data of user characteristics in each period is calculated to determine the scoring data corresponding to the advertisement data of user characteristics in each period.

[0026] Further, the advertisement adaptation and recommendation analysis module includes an advertisement feature node clustering processing unit and an advertisement adaptation and recommendation analysis unit;

[0027] The advertisement feature node clustering processing unit outputs the advertisement data with the maximum scoring data as the best advertisement data of user characteristics in the period based on the scoring data of the advertisement data of user characteristics in each period; then, taking each texture element feature node on the texture element feature path corresponding to the best advertisement data of user characteristics in the period as a reference node, node clustering processing is carried out on the texture element feature nodes corresponding to the remaining advertisement data of user characteristics in each period in the advertisement data set of user characteristics in the period; according to the result of the node clustering processing, the clustering adaptation state between the texture element feature node corresponding to the best advertisement data of user characteristics in the period corresponding to each reference node and the texture element feature nodes corresponding to the remaining advertisement data of user characteristics in each period is determined;

[0028] The advertisement adaptation recommendation analysis unit constructs concentric circles with each reference node as the center, and respectively performs density analysis on the texture element feature nodes corresponding to the remaining user characteristic advertisement data in each concentric circle; by gradually increasing the radius of the concentric circle and respectively calculating the ratio of the number of nodes to the area in the concentric circle corresponding to each radius, the node density value of the concentric circle corresponding to each radius is determined, and the concentric circle corresponding to the maximum node density value is taken as the best adaptation node concentric circle of the corresponding reference node; respectively based on the best adaptation node concentric circles corresponding to the texture element feature nodes of the periodic best user characteristic advertisement data, the types and scores of the texture element feature nodes in the best adaptation node division concentric circle are coordinated. If there are two or more texture element feature nodes of the same type, the texture element feature nodes of the same type are classified and screened, and the maximum and minimum scores of the texture element feature nodes of the same type are respectively retained; respectively in the best adaptation node concentric circles where the texture element feature nodes of the periodic best user characteristic advertisement data are located, a full traversal of the different type texture element feature nodes is performed starting from the center of the circle. For the existence of texture element feature nodes of the same type, branch traversals are respectively performed, and the maximum value node traversal path and the minimum value node traversal path in the current best adaptation node concentric circle are output; respectively calculate the sum of the node scores of the maximum value node traversal path and the minimum value node traversal path, respectively determine the upper and lower limits of the adaptation advertisement recommendation score for each texture element feature node corresponding to the current periodic best user characteristic advertisement data, and construct an adaptation advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adaptation advertisement recommendation score intervals of the nodes on the path of the texture element feature nodes corresponding to the current periodic best user characteristic advertisement data, and perform a sum and average calculation on both the upper and lower limits of each interval. The calculated upper limit sum average value is output as the upper limit of the periodic comprehensive adaptation advertisement recommendation score interval, and the calculated lower limit sum average value is output as the lower limit of the periodic comprehensive adaptation advertisement recommendation score interval, thus constructing a comprehensive adaptation advertisement recommendation score interval corresponding to the current user cycle.

[0029] Further, the feedback update module includes an advertisement push judgment unit and a data self-update unit;

[0030] The advertisement push judgment unit calculates and screens the score of the pushed advertisement data based on the comprehensive adaptation advertisement recommendation score interval within the user cycle; if the score of the pushed advertisement data is within the comprehensive adaptation advertisement recommendation score range, it is pushed to the user; otherwise, the advertisement data is screened out;

[0031] The data self-update unit realizes a periodic self-update loop, and updates and re-analyzes the comprehensive adaptation advertisement recommendation score interval by the user's self-set update time.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] The present invention performs personalized advertisement data acquisition and screening processing for users. Based on the processed advertisement data, it extracts advertisement texture element features, determines the advertisement data texture element feature path, and calculates the advertisement data score. It determines the optimal characteristic advertisement data for the user cycle, and accordingly performs node clustering analysis on each node of the optimal characteristic advertisement data to determine the adapted texture element nodes, analyzes the adapted advertisement recommendation score range, thereby making accurate judgments on user advertisement placement and achieving self-loop update. The present invention can more accurately attract users' attention with the visual element combination of advertisements for accurate placement, improve the effective browsing status of users' advertisement data, and improve the full utilization of advertisement placement resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic structural diagram of an intelligent advertisement visual element matching system based on big data according to the present invention;

[0035] Figure 2 It is a schematic flowchart of an intelligent advertisement visual element matching method based on big data according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] Embodiment: As Figure 1 shown, the present invention provides a technical solution:

[0038] An intelligent advertisement visual element matching system based on big data, the system includes an advertisement record acquisition module, a texture element feature processing module, an advertisement adaptation recommendation analysis module, and a feedback update module;

[0039] The advertisement record acquisition module acquires the user's historical advertisement viewing record data, determines the analysis period, constructs an advertisement viewing data set for the corresponding period, and based on personalized screening analysis of the user, determines the advertisement data set of the user characteristics for the period; the texture element feature processing module extracts the corresponding advertisement texture element features based on the advertisement data set of the user characteristics for the period, constructs a texture element feature path for the advertisement data of the user characteristics for the corresponding period, and analyzes the score of the advertisement data of the user characteristics for the corresponding period to determine the best advertisement data of the user characteristics for the period; the advertisement adaptation recommendation analysis module uses the texture element feature path nodes of the best advertisement data of the user characteristics for the period as reference nodes, respectively performs clustering processing on the texture element feature path nodes corresponding to the advertisement data of the user characteristics for the period, analyzes the adapted texture element features of the texture element feature path nodes of each best advertisement data of the user characteristics, and comprehensively determines the recommended score range of the adapted advertisement data for the corresponding user; the feedback update module makes a push judgment on the pushed advertisement based on the recommended score range of the adapted advertisement data, and performs cyclic update processing on the recommended score range of the adapted advertisement data for the period.

[0040] Further, the advertisement record acquisition module includes an advertisement data acquisition unit and an advertisement data personalized processing unit;

[0041] The advertisement data acquisition unit determines the analysis period window by calling the user's historical advertisement viewing record database, retrieves the user's historical advertisement viewing record data within the corresponding period, and constructs an advertisement viewing data set for the corresponding period; the advertisement viewing data set for the period includes the viewing duration data, advertisement type data, and corresponding advertisement number data of the corresponding advertisement user.

[0042] The advertisement data personalized processing unit performs personalized screening analysis based on the average viewing duration of each advertisement data in the advertisement viewing data set for the corresponding period; determines the advertisement data set of the user characteristics for the period after the personalized screening analysis.

[0043] Further, the texture element feature processing module includes an advertisement texture element path construction unit and an advertisement data score analysis unit;

[0044] The advertisement texture element path construction unit extracts the texture element features of the advertisement data of the user characteristics for the corresponding period respectively according to the advertisement data of the user characteristics for the period in the advertisement data set of the user characteristics for the period, constructs the corresponding texture element feature nodes, and performs node connection to construct the texture element feature path of the advertisement data of the user characteristics for the corresponding period;

[0045] The advertisement data scoring and analysis unit performs scoring and analysis on the advertisement data of user characteristics in each period respectively according to the texture element feature paths corresponding to the advertisement data of user characteristics in each period; it constructs a feature matrix of corresponding texture element feature nodes for the texture element feature nodes of the advertisement data of user characteristics in each period respectively; the feature matrix of the texture element feature nodes is composed of the proportion weight of the advertisement data corresponding to the texture element feature and the type experience score; based on the feature matrix of the texture element feature nodes corresponding to the advertisement data of user characteristics in each period, scoring and analysis of the texture element feature nodes is carried out; according to the scoring data of the texture element feature nodes corresponding to the advertisement data of user characteristics in each period, summation calculation of the scores of the nodes on the texture element feature paths corresponding to the advertisement data of user characteristics in each period is carried out to determine the scoring data of the advertisement data of user characteristics in each period.

[0046] Further, the advertisement adaptation and recommendation analysis module includes an advertisement feature node clustering processing unit and an advertisement adaptation and recommendation analysis unit;

[0047] The advertisement feature node clustering processing unit outputs the advertisement data with the maximum scoring data as the best advertisement data of user characteristics in the period based on the scoring data of the advertisement data of user characteristics in each period; then, taking each texture element feature node on the texture element feature path corresponding to the best advertisement data of user characteristics in the period as a reference node, node clustering processing is carried out on the texture element feature nodes corresponding to the remaining advertisement data of user characteristics in each period in the advertisement data set of user characteristics in the period; according to the result of the node clustering processing, the clustering adaptation status of the texture element feature nodes corresponding to the best advertisement data of user characteristics in the period corresponding to each reference node and the texture element feature nodes corresponding to the remaining advertisement data of user characteristics in each period is determined;

[0048] The advertisement adaptation recommendation analysis unit constructs concentric circles with each reference node as the center, and respectively performs density analysis on the texture element feature nodes corresponding to the remaining user characteristic advertisement data in each concentric circle; by gradually increasing the radius of the concentric circle and respectively calculating the ratio of the number of nodes to the area in the concentric circle corresponding to each radius, the node density value of the concentric circle corresponding to each radius is determined, and the concentric circle corresponding to the maximum node density value is taken as the best adaptation node concentric circle of the corresponding reference node; respectively based on the best adaptation node concentric circles corresponding to the texture element feature nodes of the periodic best user characteristic advertisement data, overall plan the types and scores of the texture element feature nodes in the best adaptation node divided concentric circles. If there are two or more texture element feature nodes of the same type, the texture element feature nodes of the same type are classified and screened, and the maximum and minimum scores of the texture element feature nodes of the same type are respectively retained; respectively in the best adaptation node concentric circles where the texture element feature nodes of the periodic best user characteristic advertisement data are located, perform a full traversal of the different type texture element feature nodes starting from the center of the circle. For the existence of texture element feature nodes of the same type, perform a branch traversal respectively, and output the maximum value node traversal path and the minimum value node traversal path in the current best adaptation node concentric circle; respectively calculate the sum of the node scores of the maximum value node traversal path and the minimum value node traversal path, respectively determine the upper and lower limits of the adaptation advertisement recommendation score corresponding to each texture element feature node of the current periodic best user characteristic advertisement data, and construct an adaptation advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adaptation advertisement recommendation score intervals of the nodes on the path of the texture element feature nodes corresponding to the current periodic best user characteristic advertisement data, and perform a sum average calculation on both the upper and lower limits of each interval. The calculated upper limit sum average value is output as the upper limit of the periodic comprehensive adaptation advertisement recommendation score interval, and the calculated lower limit sum average value is output as the lower limit of the periodic comprehensive adaptation advertisement recommendation score interval, then construct a comprehensive adaptation advertisement recommendation score interval corresponding to the current user cycle.

[0049] Further, the feedback update module includes an advertisement push judgment unit and a data self-update unit;

[0050] The advertisement push judgment unit calculates and screens the score of the pushed advertisement data based on the comprehensive adaptation advertisement recommendation score interval within the user cycle; if the score of the pushed advertisement data is within the comprehensive adaptation advertisement recommendation score range, it is pushed to the user; otherwise, the advertisement data is screened out;

[0051] The data self-update unit realizes a periodic self-update loop, and updates and re-analyzes the comprehensive adaptation advertisement recommendation score interval by the user's self-set update time;

[0052] As Figure 2 shown, the present invention provides another technical solution:

[0053] An intelligent matching method for advertising visual elements based on big data, the method comprising the following steps:

[0054] Obtain user historical advertising viewing record data, determine the analysis period, construct an advertising viewing data set for the corresponding period, and based on user personalized screening analysis, determine the periodic user characteristic advertising data set;

[0055] Extract the corresponding advertising texture element features based on the periodic user characteristic advertising data set, construct the texture element feature path of the periodic user characteristic advertising data, and analyze the scores of the periodic user characteristic advertising data to determine the best periodic user characteristic advertising data; taking the texture element feature path nodes of the best periodic user characteristic advertising data as reference nodes, perform clustering processing on the corresponding texture element feature path nodes of the periodic user characteristic advertising data respectively, analyze the adapted texture element features of the texture element feature path nodes of each best user characteristic advertising data, and comprehensively determine the recommended score range of the adapted advertising data for the corresponding user.

[0056] Further, by calling the user's historical advertising viewing record database, determine the analysis period window, retrieve the user's historical advertising viewing record data within the corresponding period, and construct an advertising viewing data set for the corresponding period; wherein, the user's historical advertising viewing record database includes a website advertising viewing record database and an application advertising viewing record database; the periodic advertising viewing data set includes the viewing duration data, advertising type data, and corresponding advertising number data of the corresponding advertising users;

[0057] Perform personalized screening analysis according to the average viewing duration of each advertising data in the periodic advertising viewing data set of the corresponding user, screen out the advertising data with a viewing duration less than the average viewing duration, and retain the rest, to determine the periodic user characteristic advertising data set after personalized screening analysis.

[0058] Further, according to the periodic user characteristic advertisement data in the periodic user characteristic advertisement dataset, extract the texture element features of the corresponding periodic user characteristic advertisement data respectively, construct the corresponding texture element feature nodes, and perform node connection to construct the texture element feature path of the corresponding periodic user characteristic advertisement data; wherein the texture element features of the advertisement data include text type texture element features, graphic type texture element features, color type texture element features, etc., and specific texts are further subdivided for text type, graphic type, and color type, such as Chinese, English, etc.; for the graphic type, it is further subdivided into rectangle, circle, etc.; for the color type, it is further subdivided into red, blue, etc.; according to the texture element feature paths of the corresponding periodic user characteristic advertisement data, perform scoring analysis on the corresponding periodic user characteristic advertisement data respectively; construct the corresponding texture element feature node feature matrix for the texture element feature nodes of the corresponding periodic user characteristic advertisement data respectively; the texture element feature node feature matrix is composed of the proportion weight of the advertisement data corresponding to the texture element feature and the type experience score; wherein the type experience score of the texture element feature is the reference score of the corresponding advertisement texture element feature given by big data; based on the texture element feature node feature matrix of the corresponding periodic user characteristic advertisement data, perform texture element feature node scoring analysis, and its calculation formula is

[0059]

[0060] wherein, P[W,Q]n is the scoring data of the texture element feature node n in the corresponding periodic user characteristic advertisement data; W is the proportion weight of the advertisement data of the texture element feature node n; Q is the advertisement data type experience scoring data of the texture element feature node n;

[0061] According to the texture element feature node scoring data of the corresponding periodic user characteristic advertisement data, perform the sum calculation of the node scores on the texture element feature path of the corresponding periodic user characteristic advertisement data to determine the scoring data of the corresponding periodic user characteristic advertisement data.

[0062] Further, based on the scoring data of the periodic user characteristic advertisement data, output the maximum value of the scoring data as the periodic best user characteristic advertisement data; then, taking each texture element feature node on the texture element feature path corresponding to the periodic best user characteristic advertisement data as a reference node, perform node clustering processing on the texture element feature nodes corresponding to the remaining periodic user characteristic advertisement data in the periodic user characteristic advertisement dataset; the remaining periodic user characteristic advertisement data refers to the remaining periodic user characteristic advertisement data in the periodic user characteristic advertisement dataset after removing the periodic best user characteristic advertisement data; according to the node clustering processing result, determine the clustering adaptation status of each reference node corresponding to the texture element feature node of the periodic best user characteristic advertisement data and the texture element feature nodes corresponding to the remaining periodic user characteristic advertisement data;

[0063] Construct concentric circles with each reference node as the center, and perform density analysis on the texture element feature nodes corresponding to the remaining user characteristic advertisement data in each concentric circle; by gradually increasing the radius of the concentric circle and calculating the ratio of the number of nodes to the area in the concentric circle corresponding to each radius respectively, determine the node density value of the concentric circle corresponding to each radius, and take the concentric circle corresponding to the maximum node density value as the best matching node concentric circle of the corresponding reference node; respectively based on the best matching node concentric circles corresponding to the texture element feature nodes of the best user characteristic advertisement data in the period, overall plan the types and scores of the texture element feature nodes in the concentric circles divided by the best matching nodes. If there are two or more texture element feature nodes of the same type, classify and screen the texture element feature nodes of the same type, and respectively retain the maximum and minimum scores of the texture element feature nodes of the same type; respectively perform a full traversal of the texture element feature nodes of different types starting from the center of the concentric circles where the texture element feature nodes of the best user characteristic advertisement data in the period are located. For the texture element feature nodes of the same type, perform a branch traversal respectively, and output the traversal paths of the maximum value nodes and the minimum value nodes in the current best matching node concentric circle; since after classifying and screening the texture element feature nodes of the same type, only the maximum score node and the minimum score node of the same type are retained, through branch traversal, it can be ensured that the types of the texture element feature nodes on a traversal path are different; respectively calculate the sum of the node scores of the traversal paths of the maximum value nodes and the minimum value nodes, respectively determine the upper and lower limits of the adapted advertisement recommendation scores corresponding to the texture element feature nodes of the current best user characteristic advertisement data in the period, and construct the adapted advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adapted advertisement recommendation score intervals of the nodes on the paths of the texture element feature nodes corresponding to the best user characteristic advertisement data in the current period, and perform a sum and average calculation on the upper and lower limits of each interval. Output the sum average value of the calculated upper limits as the upper limit of the comprehensive adapted advertisement recommendation score interval in the period, and output the sum average value of the calculated lower limits as the lower limit of the comprehensive adapted advertisement recommendation score interval in the period, then construct the comprehensive adapted advertisement recommendation score interval corresponding to the current user period.

[0064] Furthermore, based on the comprehensive adapted advertisement recommendation score interval in the user period, calculate the scores of the pushed advertisement data and perform screening; if the score of the pushed advertisement data is within the comprehensive adapted advertisement recommendation score range, push it to the user; otherwise, screen out the advertisement data.

[0065] Implement a cycle of self-update. By allowing the user to set the update time, re-analyze and update the comprehensive adapted advertisement recommendation score interval in the period.

[0066] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A method for intelligently matching advertising visual elements based on big data, characterized by: The method comprises the following steps: Obtain the user's historical ad viewing record data, determine the analysis period, build the corresponding period ad viewing data set, and determine the period user characteristic ad data set based on user personalized screening analysis; Based on the periodic user characteristic advertisement data set, the corresponding advertisement texture element feature extraction is performed, the texture element feature path of the corresponding periodic user characteristic advertisement data is constructed, and the corresponding periodic user characteristic advertisement data score is analyzed to determine the periodic best user characteristic advertisement data; with the texture element feature path node of the periodic best user characteristic advertisement data as the reference node, the texture element feature path node corresponding to the periodic user characteristic advertisement data is clustered, the adapted texture element features of the texture element feature path node of each best user characteristic advertisement data are analyzed, and the recommended score interval of the adapted advertisement data of the corresponding user is comprehensively determined; Specifically include: Based on the scoring data of the user characteristic advertising data of each period, the maximum value of the scoring data is output as the best user characteristic advertising data of the period; each texture element feature node on the texture element feature path corresponding to the best user characteristic advertising data of the period is used as a reference node, and node clustering processing is performed on the texture element feature nodes corresponding to the remaining user characteristic advertising data of each period in the period user characteristic advertising data set; Concentric circles are constructed with each reference node as the center, and density analysis is performed on the texture element feature nodes corresponding to the remaining periodic user characteristic advertising data contained in the concentric circles; the node density value of the concentric circles corresponding to each radius is determined by gradually increasing the radius of the concentric circles, and the ratio of the number of nodes to the area in the concentric circles corresponding to each radius is calculated respectively, and the concentric circle corresponding to the maximum node density value is taken as the best fitting node concentric circle corresponding to the reference node; based on the best fitting node concentric circles corresponding to each texture element feature node of the periodic best user characteristic advertising data, the type and score of each texture element feature node in the concentric circle are comprehensively divided by the best fitting node; if there are two or more texture element feature nodes of the same type, the texture element feature nodes of the same type are classified and screened out, and the maximum and minimum scores of the texture element feature nodes of the same type are taken and retained respectively; in the best fitting node concentric circles where each texture element feature node of the periodic best user characteristic advertising data is located, the center of the circle is taken as the starting point. Perform a full traversal of the texture element feature nodes of different types, and for the texture element feature nodes of the same type, perform branch traversal respectively, and output the maximum value node traversal path and the minimum value node traversal path in the concentric circle of the current best adaptation node; calculate the sum of the node scores of the maximum value node traversal path and the minimum value node traversal path respectively, and determine the upper and lower limits of the adaptation advertisement recommendation score of each texture element feature node corresponding to the best user characteristic advertising data of the current period, and construct the adaptation advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adaptation advertisement recommendation score interval of each node on the texture element feature node path corresponding to the best user characteristic advertising data of the current period, sum and average the upper and lower limits of each interval, output the calculated upper limit sum average value as the upper limit of the period comprehensive adaptation advertisement recommendation score interval, and output the calculated lower limit sum average value as the lower limit of the period comprehensive adaptation advertisement recommendation score interval, and construct the comprehensive adaptation advertisement recommendation score interval corresponding to the current user period; Based on the comprehensive adaptive advertising recommendation score range within the user cycle, the push advertising data is scored and filtered; if the score of the push advertising data is within the comprehensive adaptive advertising recommendation score area, the user will be pushed; otherwise, the advertising data will be filtered out.

2. The method for intelligently matching advertising visual elements based on big data according to claim 1, characterized in that: By calling the user's historical advertisement viewing record database, determining the analysis period window, retrieving the user's historical advertisement viewing record data in the corresponding period, and constructing a corresponding period advertisement viewing data set; the period advertisement viewing data set includes the viewing time data of the corresponding advertisement user, the advertisement type data, and the corresponding advertisement number data; A personalized screening analysis is performed based on the average viewing time of each advertisement data in the periodic advertisement viewing data set of the corresponding user to determine the periodic user characteristic advertisement data set after the personalized screening analysis.

3. The method for intelligently matching advertising visual elements based on big data according to claim 2 is characterized by: According to the periodic user characteristic advertisement data in the periodic user characteristic advertisement data set, texture element features of the corresponding periodic user characteristic advertisement data are extracted respectively, corresponding texture element feature nodes are constructed, and the nodes are connected in series to construct texture element feature paths of the corresponding periodic user characteristic advertisement data; according to the texture element feature paths of the corresponding periodic user characteristic advertisement data, scoring analysis is performed on the respective periodic user characteristic advertisement data; the texture element feature node feature matrix of the texture element feature nodes of the respective periodic user characteristic advertisement data is constructed; the texture element feature node feature matrix is ​​composed of the corresponding texture element feature advertisement data proportion weight and type experience score combination; based on the texture element feature node feature matrix of the corresponding periodic user characteristic advertisement data, texture element feature node scoring analysis is performed; According to the texture element feature node scoring data corresponding to each period user characteristic advertisement data, the node scoring summation calculation on the texture element feature path corresponding to each period user characteristic advertisement data is performed to determine the scoring data corresponding to each period user characteristic advertisement data.

4. The method for intelligently matching advertising visual elements based on big data according to claim 3 is characterized by: Achieve periodic self-update cycle, and update and re-analyze the periodic comprehensive adaptation advertising recommendation score interval by allowing users to set the update time.

5. An intelligent matching system for advertising visual elements based on big data, using an intelligent matching method for advertising visual elements based on big data as described in any one of claims 1 to 4, characterized in that: The system includes an advertisement record acquisition module, a texture element feature processing module, an advertisement adaptation recommendation analysis module and a feedback update module; The advertising record acquisition module acquires the user's historical advertising viewing record data, determines the analysis period, constructs the corresponding period advertising viewing data set, and determines the period user characteristic advertising data set based on the user personalized screening analysis; the texture element feature processing module extracts the corresponding advertising texture element features based on the period user characteristic advertising data set, constructs the texture element feature path of the corresponding period user characteristic advertising data, analyzes the corresponding period user characteristic advertising data score, and determines the period best user characteristic advertising data; the advertising adaptation recommendation analysis module uses the texture element feature path nodes of the period best user characteristic advertising data as reference nodes, performs clustering processing on the texture element feature path nodes corresponding to the period user characteristic advertising data, analyzes the adaptation texture element features of the texture element feature path nodes of each best user characteristic advertising data, and comprehensively determines the corresponding user's adaptation advertising data recommendation score interval; the feedback update module makes a push judgment on the push advertisement based on the adaptation advertising data recommendation score interval, and performs a cyclic update processing on the period adaptation advertising data recommendation score interval.

6. The intelligent matching system for advertising visual elements based on big data according to claim 5 is characterized by: The advertisement record acquisition module includes an advertisement data acquisition unit and an advertisement data personalization processing unit; The advertisement data acquisition unit determines the analysis period window by calling the user's historical advertisement viewing record database, retrieves the user's historical advertisement viewing record data in the corresponding period, and constructs a corresponding period advertisement viewing data set; the period advertisement viewing data set includes the viewing time data of the corresponding advertisement user, the advertisement type data, and the corresponding advertisement number data; The advertisement data personalization processing unit performs personalized screening analysis according to the average viewing time of each advertisement data in the periodic advertisement viewing data set of the corresponding user; and determines the periodic user characteristic advertisement data set after the personalized screening analysis.

7. The intelligent matching system for advertising visual elements based on big data according to claim 6 is characterized by: The texture element feature processing module includes an advertisement texture element path construction unit and an advertisement data scoring analysis unit; The advertisement texture element path construction unit extracts texture element features of the corresponding periodic user characteristic advertisement data according to each periodic user characteristic advertisement data in the periodic user characteristic advertisement data set, constructs corresponding texture element feature nodes, and constructs texture element feature paths of the corresponding periodic user characteristic advertisement data by connecting the nodes in series; The advertising data scoring and analysis unit performs scoring and analysis on the advertising data with user characteristics of each period respectively according to the texture element feature path corresponding to the advertising data with user characteristics of each period; it constructs the corresponding texture element feature node feature matrix for the texture element feature nodes of the advertising data with user characteristics of each period respectively; the texture element feature node feature matrix is ​​composed of the advertising data proportion weight and type experience score combination of the corresponding texture element feature; based on the texture element feature node feature matrix corresponding to the advertising data with user characteristics of each period, a texture element feature node scoring analysis is performed; based on the texture element feature node scoring data corresponding to the advertising data with user characteristics of each period, the node scores on the texture element feature path corresponding to the advertising data with user characteristics of each period are summed up and calculated to determine the scoring data corresponding to the advertising data with user characteristics of each period.

8. The intelligent matching system for advertising visual elements based on big data according to claim 7 is characterized by: The advertisement adaptation recommendation analysis module includes an advertisement feature node clustering processing unit and an advertisement adaptation recommendation analysis unit; The advertising feature node clustering processing unit outputs the maximum value of the scoring data as the advertising data with the best user characteristics for the period based on the scoring data of the advertising data with user characteristics for each period; takes each texture element feature node on the texture element feature path corresponding to the advertising data with the best user characteristics for the period as a reference node, and performs node clustering processing on the texture element feature nodes corresponding to the remaining advertising data with user characteristics for each period in the advertising data set with user characteristics for the period; The advertisement adaptation recommendation analysis unit constructs concentric circles with each reference node as the center, and respectively performs density analysis on the texture element feature nodes corresponding to the remaining periodic user characteristic advertisement data contained in the concentric circles; by gradually increasing the radius of the concentric circles, and respectively calculating the ratio of the number of nodes to the area in the concentric circles corresponding to each radius, the node density value of the concentric circles corresponding to each radius is determined, and the concentric circle corresponding to the maximum node density value is taken as the best adaptation node concentric circle corresponding to the reference node; based on the best adaptation node concentric circles corresponding to each texture element feature node of the periodic best user characteristic advertisement data, the type and score of each texture element feature node in the concentric circle of the best adaptation node division are comprehensively analyzed, and if there are two or more texture element feature nodes of the same type, the texture element feature nodes of the same type are classified and screened out, and the maximum and minimum scores of the texture element feature nodes of the same type are respectively retained; in the best adaptation node concentric circles where each texture element feature node of the periodic best user characteristic advertisement data is located, Taking the center of the circle as the starting point, perform a full traversal of the texture element feature nodes of different types. For the texture element feature nodes of the same type, perform branch traversal respectively, and output the maximum value node traversal path and the minimum value node traversal path in the concentric circle of the current best adaptation node; calculate the sum of the node scores of the maximum value node traversal path and the minimum value node traversal path respectively, and determine the upper and lower limits of the adaptation advertisement recommendation score of each texture element feature node corresponding to the best user characteristic advertising data of the current period respectively, and construct the adaptation advertisement recommendation score interval corresponding to the current texture element feature node; integrate the adaptation advertisement recommendation score interval of each node on the texture element feature node path corresponding to the best user characteristic advertising data of the current period, sum and average the upper and lower limits of each interval, output the calculated upper limit sum average value as the upper limit of the period comprehensive adaptation advertisement recommendation score interval, and output the calculated lower limit sum average value as the lower limit of the period comprehensive adaptation advertisement recommendation score interval, and construct the comprehensive adaptation advertisement recommendation score interval corresponding to the current user period.

9. The intelligent matching system for advertising visual elements based on big data according to claim 8 is characterized by: The feedback update module includes an advertisement push judgment unit and a data self-update unit; The advertising push judgment unit calculates and screens the push advertising data based on the comprehensive adaptation advertising recommendation score interval within the user cycle; The score of the pushed advertising data is placed in the comprehensive adaptation advertising recommendation score area for user push; Otherwise, the advertising data will be filtered out; The data self-update unit implements a periodic self-update cycle, and updates and re-analyzes the periodic comprehensive adaptation advertisement recommendation score interval through the user setting the update time.

Images