Search

CN-121996784-A - Automatic mining method for marketing text and money explosion formula based on big data analysis

CN121996784ACN 121996784 ACN121996784 ACN 121996784ACN-121996784-A

Abstract

The invention discloses an automatic mining method of a marketing text explosion formula based on big data analysis, which belongs to the field of intelligent digital marketing generation, after clustering, determining emotion commonality category and evaluating user response intensity, and further mining association rules of high response emotion types and structural modes. Aiming at the platform difference refinement rule subset to determine a platform exclusive creation rule, the crowd data are fused to generate a document template suitable for various crowds, finally, the document is optimized by matching promotion information and adjusting parameters, the problems that the traditional document creation depends on manual experience and is difficult to quantify and optimize are solved, and the data-driven accurate document generation is realized.

Inventors

  • XU ZHE

Assignees

  • 大力奇迹(杭州)科技有限责任公司

Dates

Publication Date
20260508
Application Date
20260127

Claims (10)

  1. 1. The automatic mining method for the marketing text explosion formula based on big data analysis is characterized by comprising the following steps of: Extracting historical text data and user interaction records, and analyzing the text data to obtain emotion expression vectors and structural design modes; Clustering emotion expression vectors to determine emotion expression commonality categories; associating the emotion expression commonality category with the user interaction record, evaluating the response intensity and judging the emotion expression type with high response intensity; aiming at the emotion expression type with high response intensity, extracting an associated structure design mode, and comparing repeated frequencies to obtain an associated rule of emotion and structure; If the association rule contains the platform difference feature, separating a platform rule subset and clustering and refining to determine a platform exclusive document creation rule; fusing the exclusive rule of the platform and crowd characteristic data, generating a rule set, acquiring a user psychological change index, and acquiring a document template suitable for various crowds; and matching new promotion information according to the document template, and if the matching degree is low, adjusting parameters to output an optimized document.
  2. 2. The method of claim 1, wherein the process of extracting historical document data and user interaction records, parsing the document data to obtain emotion expression vectors and structural design patterns comprises: Extracting historical document data and user interaction records; Word segmentation processing is carried out on the text data to separate words and sentence pattern units, so as to obtain an emotion related word set and a non emotion related word set; based on the emotion dictionary, matching emotion related words and giving emotion intensity values, and classifying the emotion related words as key emotion expression elements if the emotion intensity values are higher than a threshold value to form emotion expression vectors; Analyzing the sentence pattern unit structure, extracting a trunk structure and modification components, and judging whether the structure meets a common design mode or not so as to determine the core characteristics of the structural design mode; And associating the emotion expression vector with a structural design mode, and extracting features after vector mapping fusion to obtain the document feature description.
  3. 3. The method of claim 1, wherein clustering emotion expression vectors to determine emotion expression commonality categories comprises: Calculating the distance between the emotion expression vectors, and classifying the distance into the same group if the distance is smaller than a threshold value to obtain a preliminary emotion expression group; calculating a clustering center for the grouping result, extracting a representative emotion characteristic value, and determining emotion expression core characteristics; acquiring a characteristic value distribution range, comparing similarities among groups, and merging the groups if the distribution ranges overlap and exceed a threshold value to obtain optimized emotion expression class division; and (3) distributing unique identification labels for the optimized categories, obtaining emotion expression commonality category results, and judging typical classification modes.
  4. 4. The method of claim 1, wherein associating emotion expression commonality categories with user interaction records, evaluating response intensity and determining high response intensity emotion expression types comprises: Acquiring interaction frequency data under emotion expression categories from a user interaction record database, and classifying and sorting to obtain an emotion category association mapping table; Based on the mapping table, adopting an emotion analysis tool to quantitatively evaluate response intensity of emotion expression categories under different user preferences, and marking the emotion expression categories as high response emotion types if the emotion expression categories are higher than a threshold value to form a high response emotion type set; Acquiring user feedback data corresponding to the high-response emotion type set from the user interaction record, comparing and analyzing the matching degree, and judging the emotion expression combination with the most representation; Based on the most representative combination, user behavior characteristic data are obtained, and the emotion expression recommendation scheme suitable for different user preferences is determined through deep mining.
  5. 5. The method of claim 1, wherein extracting the association structure design pattern for the high response intensity emotion expression type, and comparing the repetition frequency to obtain the emotion to structure association rule comprises: Extracting structural design patterns associated with the emotion expression types with high response intensity from the document data, and counting repeated occurrence frequencies of the patterns to obtain preliminary corresponding relation data; screening the preliminary corresponding relation data, judging strong association if the repetition frequency of the structural design mode is higher than a threshold value, and obtaining a screened association mode set; Comparing the matching degree of the emotion expression type and the structural design mode, and determining a main structural design mode after sequencing to form a classification mapping table; And acquiring association rules of emotion expression types and structural design modes based on the classification mapping table, and storing and judging the optimal combination relation.
  6. 6. The method of claim 1, wherein if the association rule includes a platform difference feature, separating a subset of the platform rules and clustering the subset to refine the subset, the determining the platform-specific document creation rule comprising: extracting a data set containing platform difference characteristics from the association rules, classifying, screening and separating out rule subsets of a specific platform, and performing identification processing to obtain a rule grouping set; Adopting a clustering tool to refine the rule grouping set, dividing platform characteristics in a multi-dimensional manner, and determining the rule characteristics of creating the text in the subset; If the characteristics of the document creation rule accord with the threshold value, extracting kernel rule elements in the subset, judging the matching degree with the exclusive attribute of the platform, and obtaining a preliminary selection set of the exclusive document rule of the platform; and comparing and verifying the initial set, acquiring a document rule highly related to the platform difference characteristic, and integrating the verification result to determine a final platform exclusive document creation rule set.
  7. 7. An automated marketing document payoff formula mining system based on big data analysis, for implementing the method of any of claims 1-6, the system comprising: the data extraction and analysis module is used for extracting historical document data and user interaction records, and analyzing the document data to obtain emotion expression vectors and structural design modes; the emotion clustering module is used for clustering emotion expression vectors to determine emotion expression commonality categories; The response evaluation module is used for associating the emotion expression commonality category with the user interaction record, evaluating the response intensity and judging the emotion expression type with high response intensity; The association rule mining module is used for extracting an association structure design mode aiming at the emotion expression type with high response intensity, and obtaining association rules of emotion and structure by comparing repeated frequencies; the platform adaptation module is used for separating a platform rule subset and clustering and refining when the association rule contains the platform difference characteristics, and determining a platform exclusive document creation rule; The template generation module is used for fusing the exclusive rule of the platform and the crowd characteristic data, generating a rule set and acquiring a psychological change index of a user to obtain a document template suitable for various crowds; and the document output module is used for matching new promotion information according to the document template, adjusting parameters when the matching degree is low, and outputting an optimized document.
  8. 8. A computer terminal device, comprising: One or more processors; a memory coupled to the processor for storing one or more programs; When executed by the one or more processors, causes the one or more processors to implement the steps of the method of any of claims 1-6.
  9. 9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-6.
  10. 10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-6.

Description

Automatic mining method for marketing text and money explosion formula based on big data analysis Technical Field The invention belongs to the technical field of intelligent digital marketing generation, and particularly relates to an automatic mining method for a marketing text explosion formula based on big data analysis. Background In the current digital marketing field, document creation is an important bridge connecting brands and consumers, and the quality of the document creation directly influences marketing effects and market competitiveness. However, the conventional document creation method mainly depends on personal experience and intuitive judgment of the creator. This approach exposes significant limitations in facing the needs of massive users and diverse platforms and people. The user preference which is difficult to systematically grasp and change rapidly is judged manually, so that the user preference cannot adapt to the different situation characteristics timely. For example, in a promotional scenario on an e-commerce platform, while experience shows that warm, intimate mood may be more attractive to click than formal mood, this perception cannot be quantified and its effective collocation rules with a particular sentence pattern structure are always undefined, resulting in a high uncertainty in the effect of the literature, affecting the overall performance of the marketing campaign. The more advanced technical difficulty is that the inherent association and commonality rules between key elements such as emotion expression and structural design are difficult to accurately extract from massive success and failure cases. The attractive force of the text is the result of the combined action of multidimensional factors such as emotion, intonation, structure and the like, and the factors are mutually interwoven and mutually influenced. The existing methods lack effective technical means to disassemble these complex relationships, for example, how the fine variation of emotion expression ultimately affects the resonance effect of structural design through intonation intermediation cannot be quantitatively analyzed. The lack of analytical capability makes the authoring process difficult to surpass individual experience, and cannot form a reusable and verifiable standardized authoring rule, so that enterprises have low efficiency and unstable quality in content output. Thus, the digital marketing field has long faced with the core technical challenges of how to convert unstructured document content into a computable, analyzable data model, and automatically mine generic authoring formulas therefrom. The solution of the problem is not only to break through the bottleneck of the natural language processing technology in deep semantic association mining, but also to establish a set of intelligent generation mechanism capable of dynamically adapting to the characteristics of the platform and the crowd characteristics, thereby fundamentally improving the accuracy and the scale capability of the document creation. Disclosure of Invention In order to solve the technical problems, the invention provides an automatic mining method for a marketing text payoff formula based on big data analysis, which aims to solve the problems in the prior art. In order to achieve the above object, the present invention provides an automatic mining method for a marketing text payoff formula based on big data analysis, comprising the following steps: Extracting historical text data and user interaction records, and analyzing the text data to obtain emotion expression vectors and structural design modes; Clustering emotion expression vectors to determine emotion expression commonality categories; associating the emotion expression commonality category with the user interaction record, evaluating the response intensity and judging the emotion expression type with high response intensity; aiming at the emotion expression type with high response intensity, extracting an associated structure design mode, and comparing repeated frequencies to obtain an associated rule of emotion and structure; If the association rule contains the platform difference feature, separating a platform rule subset and clustering and refining to determine a platform exclusive document creation rule; fusing the exclusive rule of the platform and crowd characteristic data, generating a rule set, acquiring a user psychological change index, and acquiring a document template suitable for various crowds; and matching new promotion information according to the document template, and if the matching degree is low, adjusting parameters to output an optimized document. In a second aspect, the invention also provides an automatic mining system for the marketing text explosion formula based on big data analysis, which is used for implementing an automatic mining method for the marketing text explosion formula based on big data analysis, and the system comprises t