Intelligent recommendation method and system based on family circle portrait

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent recommendation method and system based on a family circle portrait. The method comprises the following steps of: analyzing household user data of a platform, performing preliminary fitting analysis on commodities of an advertisement and household user consumption data of the platform to obtain a propagation income basic expected value of the advertisement put on the platform, and taking the propagation income basic expected value as a basic value; performing overall fitting calculation by combining the manufacturing style of the advertisement with the target group style in the home users of the platform and the overall home user style of the platform to obtain a revenue correction value of the advertisement; comprehensive fitting is carried out in combination with the propagation income basic expected value of the advertisement put on the platform and the income correction value of the advertisement to calculate the comprehensive propagation income, and the comprehensive propagation income can show the adaptation degree between the made advertisement style and the putting platform home user in a more detailed manner; therefore, more detailed and specific guidance indexes are provided for advertisement putting, and accurate advertisement putting is facilitated.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an intelligent recommendation method and system based on family circle portraits. Background Art

[0002] Carrier services refer to the various communication services provided by carriers to users, including voice, data, video, etc. Home users are an important customer group for telecom operators, with high market value and loyalty. Therefore, accurately identifying home users and providing personalized services is the key to improving the competitiveness and revenue of telecom operators.

[0003] Chinese patent 202110721422.8 discloses a product recommendation method, device, equipment and storage medium based on dual portraits, including: using product data and user feedback data on products to generate a product portrait of each product, using the user data and product browsing data of the target user to generate a user portrait of the target user, calculating the matching value between the user portrait and each product portrait, collecting the products corresponding to the product portraits whose matching values ​​are greater than the preset threshold as the products to be recommended, screening the recommended products for mutually exclusive key keywords, and recommending the screened products to the target users. In addition, the present invention also relates to blockchain technology, and product data can be stored in the nodes of the blockchain. The present invention also proposes a product recommendation device, electronic device and storage medium based on dual portraits. The present invention can improve the accuracy of product recommendations.

[0004] However, the above patent lacks analysis and adjustment of the compatibility between operators and platforms, which often leads to serious inconsistency between the operator's advertising style and the video platform, resulting in low benefits when advertising is placed, and even causing negative publicity effects when advertising is placed. Summary of the invention

[0005] The present invention discloses an intelligent recommendation method, system and storage medium based on family circle portrait, including: obtaining family user traffic data of an advertising delivery platform, and analyzing characteristic attributes of platform family users; calculating platform family user consumption indicators; obtaining family user style feature set; obtaining target group characteristic information of the advertisement to be delivered; calculating the basic expected value of the communication income that the advertisement to be delivered can obtain on the advertising delivery platform; obtaining the target group family user style feature set; determining the income correction value of the advertisement to be delivered on the advertising delivery platform according to the calculation result; obtaining the comprehensive communication income of the advertisement to be delivered on the advertising delivery platform; judging whether the comprehensive communication income reaches the expected value of the advertisement delivery. The advantages of the present invention are: the adaptability between the advertisement and the family users of the delivery platform can be analyzed in more detail, thereby providing more detailed and specific guidance indicators for the delivery of advertisements, which is convenient for realizing the precise delivery of advertisements.

[0006] In order to solve the above technical problems, an intelligent recommendation method, system and storage medium based on family circle portraits are provided. This technical solution solves the problem that the above patents lack analysis and adjustment of the adaptability between operators and platforms, which often leads to serious inconsistency between the operator's advertising style and the video platform, and then leads to low benefits when advertising is placed, and even causes negative publicity effects when advertising is placed.

[0007] In order to achieve the above purpose, the technical solution adopted by the present invention is: An intelligent recommendation method based on family circle portrait, comprising: Obtaining user traffic data of the advertising delivery platform, the user traffic data including user characteristics and traffic corresponding to the user characteristics, creating a user family profile based on the user characteristics and the traffic corresponding to the user characteristics, and analyzing the characteristic attributes of the platform family users based on the family profile to obtain family user characteristic attribute data; Extracting several features related to household user consumption from household user feature attribute data to obtain a household user consumption feature set, and calculating the platform household user consumption index based on the household user consumption feature set; Extracting several features related to the family user style from the family user feature attribute data to obtain a family user style feature set; Analyze the commodities to be advertised to obtain the target group characteristics information; According to the target group characteristic information of the advertisement to be placed and the consumption index of the platform's household users, the basic expected value of the communication income that the advertisement to be placed can obtain from placing it on the advertising platform is calculated by fitting; Filter out the target group users to be advertised on the platform, extract several features related to user style from the user feature attribute data of the target group users, and obtain a target group user style feature set; Obtaining the style of the advertisement to be placed, and performing comprehensive matching calculation based on the style of the advertisement to be placed, the style of the platform target group users, and the style of the platform users, and determining the revenue correction value of the advertisement to be placed on the advertisement placement platform based on the calculation result; Combined with the revenue correction and the basic expected value of communication revenue, a comprehensive calculation is performed to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; Determine whether the comprehensive communication income reaches the expected value of the advertising. If so, the advertising is considered qualified; if not, the advertising is considered unqualified.

[0008] Preferably, the characteristic attributes of the platform family users analyzed based on the family user traffic data specifically include: Extracting a number of household user features from household user traffic data to obtain a feature data set; Extracting and integrating family user features with the same attributes in the family user feature data set, and assigning weights to the family user features according to the traffic proportions corresponding to the family user features, to obtain several family user feature classification sets; Establishing a relationship between several family user feature classification sets according to the relationship between family user features in several family user feature classification sets; A knowledge graph of household user characteristic attributes is constructed based on the relationship between several household user characteristic classification sets, characteristic data sets and several household user characteristic classification sets.

[0009] Preferably, the step of calculating the platform household user consumption index according to the household user consumption feature set specifically comprises the following steps: Extracting a plurality of household user consumption feature classification sets from a plurality of household user feature classification sets, wherein the household user consumption feature classification sets include at least one household user consumption amount class and a plurality of household user consumption correction feature classification sets; According to the household user consumption correction feature classification set, several household user consumption amount correction weights are generated, and the weighted average is calculated based on the household user consumption amount correction weights and the household user consumption amount categories to obtain the platform household user consumption index.

[0010] Preferably, the step of fitting and calculating the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index specifically comprises the following steps: Extracting and integrating target group features with the same attributes from the target group feature information to be placed advertisements, and obtaining a plurality of target group feature classification sets; According to the matching degree calculation formula, the matching degree between the target group feature classification set and the family user feature classification set is calculated to obtain the advertising target matching degree; Comprehensively calculate the basic expected value of advertising dissemination revenue based on the advertising target matching degree and the platform's household user consumption indicators; The matching degree calculation formula is: In the formula, P is the matching degree, ω i is the influence weight of the feature classification set of the i-th target group, m i is the number of target group features in the i-th target group feature classification set, n i is the number of target group features in the household user feature classification set corresponding to the i-th target group feature classification set, k i is the number of common features in the feature classification set of the i-th target group and the corresponding feature classification set of family users, αi is the similarity between the feature classification set of the ith target group and its corresponding feature classification set of family users, is the weight of the jth identical feature in the i-th target group feature classification set, is the weight of the j-th identical feature in the household user feature classification set corresponding to the i-th target group feature classification set.

[0011] Preferably, the comprehensive matching degree calculation is performed according to the style of the advertisement to be placed, the style of the family users of the platform target group and the style of the family users of the platform, and the revenue correction value of the advertisement to be placed on the advertisement placement platform is determined according to the calculation result, specifically including: Propose a number of family user feature classification sets related to style to form a platform family user style feature set; Extract the features in the family user feature classification set that are the same as the target group feature classification set, and combine them into a platform target group attribute class set; Extracting style features related to the platform target group attribute classification set based on the relationship between several family user feature classification sets, and combining them into several platform target group family user style feature class sets; Extracting style features of the advertisements to be placed and combining them into a set of style feature classes of the advertisements to be placed; According to the matching degree calculation formula, the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform target group and the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform are calculated respectively; According to the correction value calculation formula, calculate the revenue correction value of the advertisement to be placed on the advertisement placement platform; The correction value calculation formula is: Where γ is the correction value, P1 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform target group, P2 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform, and ρ is the conversion rate of family users of the platform.

[0012] Preferably, the advertising delivery analysis method based on preference data further includes advertising style adjustment, and the advertising style adjustment includes: According to the correction value calculation formula, the style characteristics of the advertisement to be placed are adjusted, wherein the adjustment of the style characteristics of the advertisement to be placed includes adding and / or deleting the style characteristics of the advertisement, and adjusting the weight of each style characteristic in the style characteristic class set of the advertisement to be placed; Calculating the style feature data of the advertisement to be placed when the revenue correction value of the advertisement to be placed on the advertisement placement platform is the largest, and obtaining the optimal style feature data of the advertisement to be placed; The advertisement to be delivered is re-created according to the best style feature data of the advertisement to be delivered.

[0013] An intelligent recommendation system based on family circle portrait, used to implement the above-mentioned intelligent recommendation method based on family circle portrait, comprising: A platform analysis module, the platform analysis module is used to analyze the characteristic attributes of the platform family users according to the family user traffic data, obtain the family user characteristic attribute data, extract several characteristics related to the family user consumption from the family user characteristic attribute data, obtain the family user consumption feature set, calculate the platform family user consumption index according to the family user consumption feature set, extract several characteristics related to the family user style from the family user characteristic attribute data, and obtain the family user style feature set; An advertisement analysis module, which is used to analyze the commodity subject to be advertised to obtain target group characteristic information and the style of the advertisement to be placed; A fitting analysis module, which is used to fit and calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index, determine the revised value of the revenue of the advertisement to be placed on the advertisement placement platform, and perform a comprehensive calculation based on the revised value and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertisement placement platform and determine whether the comprehensive communication revenue reaches the expected value of the advertisement placement; An advertisement adjustment module, wherein the advertisement adjustment module is used to adjust the advertisement style.

[0014] Optionally, the platform analysis module includes: A family user characteristic attribute analysis unit, the family user characteristic attribute analysis unit is used to analyze the characteristic attributes of the platform family users according to the family user traffic data; A household user consumption analysis unit, the household user consumption analysis unit is used to calculate the platform household user consumption index according to the household user consumption feature set; A family user style analysis unit is used to extract several features related to the family user style from the family user feature attribute data to obtain a family user style feature set.

[0015] Optionally, the fitting analysis module includes: A basic revenue analysis unit, which is used to calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index; A revenue correction unit, the revenue correction unit is used to determine a revenue correction value of the advertisement to be placed on the advertisement placement platform; A comprehensive revenue analysis unit, which is used to perform a comprehensive calculation based on the revenue correction and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; A judgment unit is used to judge whether the comprehensive communication income reaches the expected value of the advertisement delivery.

[0016] Furthermore, a computer-readable storage medium is proposed, on which a computer-readable program is stored. When the computer-readable program is called, it executes the above-mentioned intelligent recommendation method based on family circle portrait.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes an intelligent recommendation method based on family circle portrait, which analyzes the family user data of the platform and then performs preliminary fitting analysis on the advertised products and the family user consumption data of the platform to obtain the basic expected value of the communication revenue of the advertisement placed on the platform, and takes the basic expected value of the communication revenue as the basic value, combines the production style of the advertisement with the target group style of the family users of the platform and the overall family user style of the platform to perform overall fitting calculation, and then obtains the revised value of the advertisement revenue, and combines the basic expected value of the communication revenue of the advertisement placed on the platform and the revised value of the advertisement revenue to perform comprehensive fitting calculation to obtain the comprehensive communication revenue, which can show in more detail the compatibility between the produced advertisement style and the family users of the delivery platform, thereby providing more detailed and specific guiding indicators for the delivery of advertisements, and facilitating the precise delivery of advertisements. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of the method of the present invention; Figure 2 A flow chart of a method for analyzing characteristic attributes of platform family users according to user traffic data in the present invention; Figure 3 It is a flow chart of the method for calculating the consumption index of the family users of the platform in the present invention; Figure 4 A flow chart of the method for obtaining the basic expected value of the propagation income in the present invention; Figure 5 The present invention is a flow chart of a method for determining a revenue correction value of an advertisement to be placed on an advertisement placement platform. DETAILED DESCRIPTION

[0019] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0020] Embodiment 1: Reference Figure 1 As shown, an intelligent recommendation method based on family circle portrait includes: obtaining user traffic data of an advertising delivery platform, the user traffic data includes user characteristics and traffic corresponding to the user characteristics, performing user family portrait according to the user characteristics and the traffic corresponding to the user characteristics, and analyzing characteristic attributes of family users of the platform according to the family portrait to obtain characteristic attribute data of the family users; Extracting several features related to household user consumption from household user feature attribute data to obtain a household user consumption feature set, and calculating the platform household user consumption index based on the household user consumption feature set; Extracting several features related to the family user style from the family user feature attribute data to obtain a family user style feature set; Analyze the commodities to be advertised to obtain the target group characteristics information; According to the target group characteristic information of the advertisement to be placed and the consumption index of the platform's household users, the basic expected value of the communication income that the advertisement to be placed can obtain from placing it on the advertising platform is calculated by fitting; Filter out the target group users to be advertised on the platform, extract several features related to user style from the user feature attribute data of the target group users, and obtain a target group user style feature set; Obtaining the style of the advertisement to be placed, and performing comprehensive matching calculation based on the style of the advertisement to be placed, the style of the platform target group users, and the style of the platform users, and determining the revenue correction value of the advertisement to be placed on the advertisement placement platform based on the calculation result; Combined with the revenue correction and the basic expected value of communication revenue, a comprehensive calculation is performed to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; Determine whether the comprehensive communication income reaches the expected value of the advertising. If so, the advertising is considered qualified; if not, the advertising is considered unqualified.

[0021] After analyzing the household user data of the platform, a preliminary fitting analysis is performed on the advertised products and the household user consumption data of the platform to obtain the basic expected value of the communication revenue of the advertisements on the platform. Taking the basic expected value of the communication revenue as the base value, the overall fitting calculation is performed based on the production style of the advertisement, the style of the target group among the household users of the platform and the overall household user style of the platform to obtain the revised value of the advertisement revenue. The comprehensive communication revenue is calculated by comprehensive fitting based on the basic expected value of the communication revenue of the advertisements on the platform and the revised value of the advertisement revenue. The comprehensive communication revenue can show in more detail the degree of compatibility between the produced advertisement style and the household users of the delivery platform.

[0022] See also Figure 2 As shown, based on the analysis of the traffic data of home users, the characteristic attributes of home users on the platform include: Extracting a number of household user features from household user traffic data to obtain a feature data set; Extracting and integrating family user features with the same attributes in the family user feature data set, and assigning weights to the family user features according to the traffic proportions corresponding to the family user features, to obtain several family user feature classification sets; Establishing a relationship between several family user feature classification sets according to the relationship between family user features in several family user feature classification sets; A knowledge graph of household user characteristic attributes is constructed based on the relationship between several household user characteristic classification sets, characteristic data sets and several household user characteristic classification sets.

[0023] It is understandable that there are many different characteristics among the family users of the platform, and the related characteristics usually have the same attributes, such as age attribute, gender attribute, etc. By extracting and integrating the family user characteristics with the same attributes in the family user characteristic data set, and assigning weights according to the proportion of each characteristic, a family user characteristic classification set is obtained; At the same time, based on the relationship between family user characteristics in the family user characteristic classification set, for example, among the male family users of a certain platform, those aged between 18 and 25 account for 38%. We establish multiple family user characteristic classification sets, and use them to construct a family user characteristic attribute knowledge graph to facilitate subsequent family user attribute analysis.

[0024] See also Figure 3 As shown, calculating the platform household user consumption index according to the household user consumption feature set specifically includes the following steps: Extracting a number of household user consumption feature classification sets from a number of household user feature classification sets, wherein the household user consumption feature classification sets include at least one household user consumption amount class and a number of household user consumption correction feature classification sets; According to the household user consumption correction feature classification set, several household user consumption amount correction weights are generated, and the weighted average is calculated based on the household user consumption amount correction weights and the household user consumption amount categories to obtain the platform household user consumption index.

[0025] Due to the differences in spending power among different groups, this solution calculates the platform's household user consumption index based on the modified characteristics of household user consumption. For example, among the male group on a certain platform, 2.5% of them spend more than 10,000 yuan on the platform, 12.5% ​​of them spend between 5,000 and 10,000 yuan, 23% of them spend between 1,000 and 5,000 yuan, 56% of them spend between 100 and 1,000 yuan, and 6% of them spend less than 100 yuan. Through weighted calculation, the consumption index of the male group on the platform is 2,188.5 yuan. In this way, several consumption indicators related to the characteristics related to household user consumption are calculated. This indicator can show the consumption power level of the platform's household users in detail and clearly, thereby providing a detailed and specific indicator for the consumption benefits that can be brought about by advertising.

[0026] See also Figure 4 As shown, the following steps are specifically used to calculate the basic expected value of the communication benefits that the advertisement to be placed can obtain on the advertisement placement platform based on the target group characteristic information of the advertisement to be placed and the platform household user consumption index: Extracting and integrating target group features with the same attributes from the target group feature information to be placed advertisements, and obtaining a plurality of target group feature classification sets; According to the matching degree calculation formula, the matching degree between the target group feature classification set and the family user feature classification set is calculated to obtain the advertising target matching degree; Comprehensively calculate the basic expected value of advertising dissemination revenue based on the advertising target matching degree and the platform's household user consumption indicators; The matching degree calculation formula is: In the formula, P is the matching degree, ω i is the influence weight of the feature classification set of the i-th target group, m i is the number of target group features in the i-th target group feature classification set, n i is the number of target group features in the household user feature classification set corresponding to the i-th target group feature classification set, k i is the number of common features in the feature classification set of the i-th target group and the corresponding feature classification set of family users, α iis the similarity between the feature classification set of the ith target group and its corresponding feature classification set of family users, is the weight of the jth identical feature in the i-th target group feature classification set, is the weight of the j-th identical feature in the household user feature classification set corresponding to the i-th target group feature classification set.

[0027] Each commodity subject to be advertised has a different consumer target group. Therefore, calculating the degree of fit between the consumer target group of the commodity subject and the household users of the advertising platform is an important indicator for determining whether the advertising is reasonable. For example, it is obviously inappropriate to place an advertisement about women's clothing on a platform where the proportion of male household users is too large. Therefore, this scheme proposes to calculate the matching degree between the target group feature classification set and the household user feature classification set, and calculate the basic expected value of the communication revenue based on this matching degree combined with the platform household user consumption index; The matching degree in this scheme is calculated by combining two indicators. On the one hand, it is the influence weight of the attribute class. For example, for products specifically for women, the important weight of the gender attribute should be relatively high. On the other hand, it is the fit within the attribute class. The fit within the attribute class is determined by the proportion of the number of the same features and the weight of the same features. For example, a certain product is specifically for female family users. In the target group gender classification set, there is only one feature, female, and the weight of female is 100%. In the platform family user gender classification set, there are both male and female features, and the weight of male is 56% and the weight of female is 44%. In this case, the fit of the gender classification set is .

[0028] See also Figure 5 As shown, a comprehensive matching degree is calculated based on the style of the advertisement to be placed, the style of the family users of the platform target group and the style of the family users of the platform, and the revenue correction value of the advertisement to be placed on the advertisement placement platform is determined based on the calculation result, specifically including: Propose a number of family user feature classification sets related to style to form a platform family user style feature set; Extract the features in the family user feature classification set that are the same as the target group feature classification set, and combine them into a platform target group attribute class set; Extracting style features related to the platform target group attribute classification set based on the relationship between several family user feature classification sets, and combining them into several platform target group family user style feature class sets; Extracting style features of the advertisements to be placed and combining them into a set of style feature classes of the advertisements to be placed; According to the matching degree calculation formula, the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform target group and the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform are calculated respectively; According to the correction value calculation formula, calculate the revenue correction value of the advertisement to be placed on the advertisement placement platform; The correction value calculation formula is: Where γ is the correction value, P1 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform target group, P2 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform, and ρ is the conversion rate of family users of the platform.

[0029] In this scheme, the basic expected value of communication revenue is calculated based on the assumption that the delivered advertisements can fully attract the target group and have no appeal to non-target groups. However, this calculation method obviously has its limitations and does not conform to the actual law of advertising dissemination. Therefore, a correction value calculation is used to correct the communication revenue. According to this scheme, the correction calculation is performed in two aspects. On the one hand, it is the matching degree between the style feature class set of the advertisements to be delivered and the style feature class set of the platform's target group family users. This value reflects the degree of attraction of the delivered advertisements to the target group. On the other hand, it is the product of the matching degree between the style feature class set of the advertisements to be delivered and the style feature class set of the platform's family users and the platform's family user conversion rate. This value reflects the attraction value of the delivered advertisements to ordinary family users of the platform. Among them, the platform's family user conversion rate represents the proportion of potential consumer groups other than the target group. This value is usually set between 5% and 10%.

[0030] Embodiment 2: An intelligent recommendation method based on family circle portrait, based on the first embodiment, this embodiment also includes advertisement style adjustment, and the advertisement style adjustment includes: According to the correction value calculation formula, the style features of the advertisement to be placed are adjusted, and the style features of the advertisement to be placed include adding and / or deleting the style features of the advertisement, and adjusting the weight of each style feature in the style feature class set of the advertisement to be placed; Calculating the style feature data of the advertisement to be placed when the revenue correction value of the advertisement to be placed on the advertisement placement platform is the largest, and obtaining the optimal style feature data of the advertisement to be placed; The advertisement to be delivered is re-created according to the best style feature data of the advertisement to be delivered.

[0031] In this embodiment, by adaptively adjusting the advertising style according to the style of the advertising delivery platform, it can effectively ensure that the advertising style is more closely aligned with the delivery platform, thereby achieving the best advertising delivery effect and ensuring the maximum dissemination benefit of the advertising delivery.

[0032] To further illustrate this solution, an intelligent recommendation system based on family circle portrait is proposed, which is used to implement the intelligent recommendation method based on family circle portrait as described above, including: The platform analysis module is used to analyze the characteristic attributes of the platform family users according to the family user traffic data, obtain the family user characteristic attribute data, extract several characteristics related to the family user consumption from the family user characteristic attribute data, obtain the family user consumption feature set, calculate the platform family user consumption index according to the family user consumption feature set, extract several characteristics related to the family user style from the family user characteristic attribute data, and obtain the family user style feature set; Advertisement analysis module, which is used to analyze the commodity subject to be advertised to obtain the target group characteristic information and the style of the advertisement to be placed; The fitting analysis module is used to fit and calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertising platform according to the characteristic information of the target group of the advertisement to be placed and the consumption index of the platform's household users, determine the revised value of the revenue of the advertisement to be placed on the advertising platform, and perform comprehensive calculation based on the revised value and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform and determine whether the comprehensive communication revenue reaches the expected value of the advertisement placement; Advertisement adjustment module: The advertisement adjustment module is used to adjust the advertisement style.

[0033] The platform analysis module includes: A family user characteristic attribute analysis unit, which is used to analyze the characteristic attributes of family users on the platform according to the family user traffic data; A household user consumption analysis unit, which is used to calculate the household user consumption index of the platform according to the household user consumption feature set; The family user style analysis unit is used to extract several features related to the family user style from the family user feature attribute data to obtain a family user style feature set.

[0034] The fitting analysis module includes: A basic revenue analysis unit, which is used to calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index; A revenue correction unit, the revenue correction unit is used to determine the revenue correction value of the advertisement to be placed on the advertisement placement platform; A comprehensive revenue analysis unit is used to perform comprehensive calculations based on revenue correction and basic expected value of communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; The judgment unit is used to judge whether the comprehensive communication income reaches the expected value of the advertisement delivery.

[0035] The working process of the above-mentioned intelligent recommendation system based on family circle portrait is as follows: Step 1: The family user characteristic attribute analysis unit analyzes the characteristic attributes of the platform family users based on the family user traffic data; Step 2: The household user consumption analysis unit calculates the platform household user consumption index according to the household user consumption feature set; Step 3: The advertising analysis module analyzes the product subject to be advertised to obtain the target group characteristic information of the advertisement to be placed; Step 4: The basic revenue analysis unit calculates the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information and the platform household user consumption index; Step 5: The advertisement analysis module obtains the style of the advertisement to be placed and obtains an advertisement style feature set; Step 6: The family user style analysis unit extracts several features related to the family user style from the family user feature attribute data to obtain a family user style feature set; Step 7: The revenue correction unit calculates the comprehensive matching degree of the style of the advertisement to be placed, the style of the family users of the platform target group, and the style of the family users of the platform, and determines the revenue correction value of the advertisement to be placed on the advertisement placement platform; Step 8: The comprehensive revenue analysis unit performs comprehensive calculation based on the revenue correction and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; Step 9: The judgment unit judges whether the comprehensive communication income reaches the expected value of the advertisement delivery; Step 10: The advertising adjustment module adjusts the advertising style according to the calculation results of the revenue correction value.

[0036] Furthermore, the present solution also proposes a computer-readable storage medium, on which a computer-readable program is stored, and when the computer-readable program is called, the intelligent recommendation method based on the family circle portrait as described above is executed; It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0037] In summary, the advantages of the present invention are that it can analyze the compatibility between advertisements and home users of the delivery platform in more detail, thereby providing more detailed and specific guidance indicators for the delivery of advertisements, thereby facilitating the precise delivery of advertisements.

[0038] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An intelligent recommendation method based on family circle portrait, characterized in that The steps include: Obtaining user traffic data of the advertising delivery platform, the user traffic data including user characteristics and traffic corresponding to the user characteristics, creating a user family profile based on the user characteristics and the traffic corresponding to the user characteristics, and analyzing the characteristic attributes of the platform family users based on the family profile to obtain family user characteristic attribute data; Extracting several features related to household user consumption from household user feature attribute data to obtain a household user consumption feature set, and calculating the platform household user consumption index based on the household user consumption feature set; Extracting several features related to the family user style from the family user feature attribute data to obtain a family user style feature set; Analyze the commodities to be advertised to obtain the target group characteristics information; According to the target group characteristic information of the advertisement to be placed and the consumption index of the platform's household users, the basic expected value of the communication income that the advertisement to be placed can obtain from placing it on the advertising platform is calculated by fitting; Filter out the target group users to be advertised on the platform, extract several features related to user style from the user feature attribute data of the target group users, and obtain a target group user style feature set; Obtaining the style of the advertisement to be placed, and performing comprehensive matching calculation based on the style of the advertisement to be placed, the style of the platform target group users, and the style of the platform users, and determining the revenue correction value of the advertisement to be placed on the advertisement placement platform based on the calculation result; Combined with the revenue correction and the basic expected value of communication revenue, a comprehensive calculation is performed to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; Determine whether the comprehensive communication income reaches the expected value of the advertising. If so, the advertising is considered qualified; if not, the advertising is considered unqualified.

2. According to claim 1, the intelligent recommendation method based on family circle portrait is characterized in that The step of analyzing the characteristic attributes of the platform family users according to the family user traffic data comprises the following steps: Extracting a number of household user features from household user traffic data to obtain a feature data set; Extracting and integrating family user features with the same attributes in the family user feature data set, and assigning weights to the family user features according to the traffic proportions corresponding to the family user features, to obtain several family user feature classification sets; Establishing a relationship between several family user feature classification sets according to the relationship between family user features in several family user feature classification sets; A knowledge graph of household user characteristic attributes is constructed based on the relationship between several household user characteristic classification sets, characteristic data sets and several household user characteristic classification sets.

3. According to claim 1, the intelligent recommendation method based on family circle portrait is characterized in that The calculation of the platform household user consumption index according to the household user consumption characteristic set comprises the following steps: Extracting a plurality of household user consumption feature classification sets from a plurality of household user feature classification sets, wherein the household user consumption feature classification sets include at least one household user consumption amount class and a plurality of household user consumption correction feature classification sets; According to the household user consumption correction feature classification set, several household user consumption amount correction weights are generated, and the weighted average is calculated based on the household user consumption amount correction weights and the household user consumption amount categories to obtain the platform household user consumption index.

4. According to claim 1, the intelligent recommendation method based on family circle portrait is characterized in that The method of fitting and calculating the basic expected value of the communication income that the advertisement to be placed can obtain on the advertisement placement platform according to the characteristic information of the target group of the advertisement to be placed and the consumption index of the platform household users specifically includes the following steps: Extracting and integrating target group features with the same attributes from the target group feature information to be placed advertisements, and obtaining a plurality of target group feature classification sets; According to the matching degree calculation formula, the matching degree between the target group feature classification set and the family user feature classification set is calculated to obtain the advertising target matching degree; Comprehensively calculate the basic expected value of advertising dissemination revenue based on the advertising target matching degree and the platform's household user consumption indicators; The matching degree calculation formula is: In the formula, P is the matching degree, ω i is the influence weight of the feature classification set of the i-th target group, m i is the number of target group features in the i-th target group feature classification set, n i is the number of target group features in the household user feature classification set corresponding to the i-th target group feature classification set, k i is the number of common features in the feature classification set of the i-th target group and the corresponding feature classification set of family users, α i is the similarity between the feature classification set of the ith target group and its corresponding feature classification set of family users, is the weight of the jth identical feature in the i-th target group feature classification set, is the weight of the j-th identical feature in the household user feature classification set corresponding to the i-th target group feature classification set.

5. According to claim 1, the intelligent recommendation method based on family circle portrait is characterized in that The said performing comprehensive matching calculation according to the style of the advertisement to be placed, the style of the family users of the platform target group and the style of the family users of the platform, and determining the revenue correction value of the advertisement to be placed on the advertisement placement platform according to the calculation result specifically includes: Propose a number of family user feature classification sets related to style to form a platform family user style feature set; Extract the features in the family user feature classification set that are the same as the target group feature classification set, and combine them into a platform target group attribute class set; Extracting style features related to the platform target group attribute classification set based on the relationship between several family user feature classification sets, and combining them into several platform target group family user style feature class sets; Extracting style features of the advertisements to be placed and combining them into a set of style feature classes of the advertisements to be placed; According to the matching degree calculation formula, the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform target group and the matching degree between the style feature class set of the advertisement to be placed and the style feature class set of the family users of the platform are calculated respectively; According to the correction value calculation formula, calculate the revenue correction value of the advertisement to be placed on the advertisement placement platform; The correction value calculation formula is: γ=P1+ρP2 Where γ is the correction value, P1 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform target group, P2 is the matching degree between the style feature set of the advertisement to be delivered and the style feature set of the family users of the platform, and ρ is the conversion rate of family users of the platform.

6. According to claim 1, the intelligent recommendation method based on family circle portrait is characterized in that The advertising delivery analysis method based on preference data also includes an advertising style adjustment, and the advertising style adjustment includes: According to the correction value calculation formula, the style characteristics of the advertisement to be placed are adjusted, wherein the adjustment of the style characteristics of the advertisement to be placed includes adding and / or deleting the style characteristics of the advertisement, and adjusting the weight of each style characteristic in the style characteristic class set of the advertisement to be placed; Calculating the style feature data of the advertisement to be placed when the revenue correction value of the advertisement to be placed on the advertisement placement platform is the largest, and obtaining the optimal style feature data of the advertisement to be placed; The advertisement to be delivered is re-created according to the best style feature data of the advertisement to be delivered.

7. According to claim 1, the recommendation system of the intelligent recommendation method based on family circle portrait is characterized in that include: A platform analysis module, the platform analysis module is used to analyze the characteristic attributes of the platform family users according to the family user traffic data, obtain the family user characteristic attribute data, extract several characteristics related to the family user consumption from the family user characteristic attribute data, obtain the family user consumption feature set, calculate the platform family user consumption index according to the family user consumption feature set, extract several characteristics related to the family user style from the family user characteristic attribute data, and obtain the family user style feature set; An advertisement analysis module, which is used to analyze the commodity subject to be advertised to obtain target group characteristic information and the style of the advertisement to be placed; A fitting analysis module, which is used to fit and calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index, determine the revised value of the revenue of the advertisement to be placed on the advertisement placement platform, and perform a comprehensive calculation based on the revised value and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertisement placement platform and determine whether the comprehensive communication revenue reaches the expected value of the advertisement placement; An advertisement adjustment module, wherein the advertisement adjustment module is used to adjust the advertisement style.

8. According to claim 7, a recommendation system of an intelligent recommendation method based on family circle portrait, characterized in that Also includes: Platform analysis module A family user characteristic attribute analysis unit, the family user characteristic attribute analysis unit is used to analyze the characteristic attributes of the platform family users according to the family user traffic data; A household user consumption analysis unit, the household user consumption analysis unit is used to calculate the platform household user consumption index according to the household user consumption feature set; A family user style analysis unit is used to extract several features related to the family user style from the family user feature attribute data to obtain a family user style feature set.

9. The recommendation system according to claim 7, characterized in that Also includes: The fitting analysis module comprises: A basic revenue analysis unit, which is used to calculate the basic expected value of the communication revenue that the advertisement to be placed can obtain on the advertisement placement platform according to the target group characteristic information of the advertisement to be placed and the platform household user consumption index; A revenue correction unit, the revenue correction unit is used to determine a revenue correction value of the advertisement to be placed on the advertisement placement platform; A comprehensive revenue analysis unit, which is used to perform a comprehensive calculation based on the revenue correction and the basic expected value of the communication revenue to obtain the comprehensive communication revenue of the advertisement to be placed on the advertising platform; A judgment unit is used to judge whether the comprehensive communication income reaches the expected value of the advertisement delivery.

Images