Intelligent marketing copywriting generation method and device, equipment and storage medium

Abstract

The invention discloses an intelligent marketing copywriting generation method and device, equipment and a storage medium, and relates to the technical field of intelligent marketing, and the method comprises the steps: analyzing a copywriting generation request input by a user, and obtaining a marketing feature parameter; generating a current matching explosive based on the marketing knowledge database and the marketing feature parameters; calling a copywriting generation model to generate a multi-source knowledge fusion context corresponding to the current matched explosive; and generating a target marketing copywriting based on the current matched explosive and the multi-source knowledge fusion context. According to the method, the key marketing parameters are obtained by analyzing the request to perform static search on the database formed by the marketing meta-knowledge, so that the most suitable current matching blasting text is obtained; and then a multi-source knowledge fusion context corresponding to the current matched explosive text is dynamically generated by using the document generation model, and the current matched explosive text is optimized based on the multi-source knowledge fusion context, so that a target marketing document better meeting marketing requirements is quickly generated, and the generation efficiency and quality of the marketing document are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent marketing technology, and in particular to an intelligent marketing copy generation method, device, equipment and storage medium. Background Art

[0002] With the booming development of large language models, using them to generate marketing copy for product marketing has become a trend.

[0003] However, the existing marketing copy generated based on the general large language model has a weak correlation with the brand and product, resulting in a low match between the generated marketing copy and the brand and product. Not only does it require multiple adjustments, but the marketing effect is also poor. That is, the existing intelligent marketing copy production method has the defects of insufficient knowledge integration and the inability to quickly generate marketing copy that meets the needs.

[0004] Therefore, how to improve the efficiency and quality of automatic generation of marketing copy has become an urgent problem to be solved. Summary of the Invention

[0005] The main purpose of this application is to provide an intelligent marketing copy generation method, device, equipment and storage medium, aiming to solve the technical problem of how to improve the automatic generation efficiency and generation quality of marketing copy.

[0006] To achieve the above objectives, this application proposes a method for generating intelligent marketing copy, which includes:

[0007] Parse the copywriting generation request input by the user to obtain marketing feature parameters;

[0008] Generate a current matching viral article based on the marketing knowledge database and the marketing feature parameters;

[0009] Calling the copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching viral article;

[0010] Generate target marketing copy based on the current matching hot article and the multi-source knowledge fusion context.

[0011] In one embodiment, the step of generating the current matching viral article based on the marketing knowledge database and the marketing characteristic parameters includes:

[0012] A mixed search is performed on the marketing knowledge database according to the marketing semantic vector and the marketing keyword to obtain the current matching hot articles recalled in multiple ways.

[0013] In one embodiment, before calling the copywriting generation model to generate the multi-source knowledge fusion context corresponding to the currently matching viral post, the method further includes:

[0014] Construct an initial sample dataset based on marketing meta-knowledge data; the marketing meta-knowledge data includes brand attribute data, product feature data, and viral article data;

[0015] Dynamically optimizing the initial sample data set to obtain an optimized sample data set;

[0016] A copywriting generation model is constructed based on the optimized sample data set.

[0017] In one embodiment, the step of dynamically optimizing the initial sample data set to obtain an optimized sample data set includes:

[0018] Performing a preset time-attenuated weighted operation on the historical hot posts and the real-time hot posts to generate a dynamic hot post weight;

[0019] The initial sample data set is optimized based on the dynamic hot article weight to obtain an optimized sample data set.

[0020] In one embodiment, the step of performing a preset time-decay weighted operation on the historical viral articles and the real-time viral articles to generate dynamic viral article weights includes:

[0021] Obtain the current publishing time interval corresponding to the historical hot article and the real-time hot article;

[0022] A preset time decay weighting operation is performed on the historical hot articles and the real-time hot articles based on the current publishing time interval to generate a dynamic hot article weight.

[0023] In one embodiment, the marketing meta-knowledge data further includes:

[0024] Audience profile data, including age, gender, and content interest tags;

[0025] Channel characteristic data, including the dissemination rules and content format requirements of different platforms;

[0026] Copywriting style data, including language style type and emotional tendency parameters.

[0027] In one embodiment, the step of generating a target marketing copy based on the current matching viral article and the multi-source knowledge fusion context includes:

[0028] Generate an initial marketing copy based on the current matching viral article and the multi-source knowledge fusion context;

[0029] Performing multi-level compliance checks on the copy generation request and the initial marketing copy;

[0030] The target marketing copy is determined based on the verification result and the initial marketing copy.

[0031] In addition, to achieve the above-mentioned purpose, the present application also proposes an intelligent marketing copy generation device, which includes:

[0032] The request parsing module is used to parse the copywriting generation request input by the user and obtain marketing feature parameters;

[0033] A static matching module, configured to generate a currently matching viral article based on a marketing knowledge database and the marketing characteristic parameters;

[0034] A dynamic optimization module is used to call the copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching explosive article;

[0035] A copywriting generation module is used to generate target marketing copywriting based on the current matching hot article and the multi-source knowledge fusion context.

[0036] In addition, to achieve the above-mentioned purpose, the present application also proposes an intelligent marketing copy generation device, which includes: a memory, a processor, and a computer program stored in the memory and runnable on the processor, and the computer program is configured to implement the steps of the intelligent marketing copy generation method as described above.

[0037] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the intelligent marketing copy generation method as described above are implemented. The present application provides an intelligent marketing copy generation method, apparatus, device, and storage medium.

[0038] This application discloses an intelligent marketing copy generation method, apparatus, device, and storage medium. The method comprises: parsing a copy generation request input by a user to obtain marketing feature parameters; generating a current matching viral post based on a marketing knowledge database and the marketing feature parameters; invoking a copy generation model to generate a multi-source knowledge fusion context corresponding to the current matching viral post; and generating a target marketing copy based on the current matching viral post and the multi-source knowledge fusion context. The application first receives a copy generation request input by a user, then parses the request to obtain marketing feature parameters. Based on the marketing feature parameters, a search is then performed in a marketing knowledge database to generate a current matching viral post. The copy generation model is then invoked to generate a dynamic multi-source knowledge fusion context based on the current matching viral post. Finally, the target marketing copy is generated by combining the current matching viral post and the multi-source knowledge fusion context. Therefore, the application parses the request to obtain key marketing parameters, performs a static search in a database composed of marketing meta-knowledge, and obtains the most suitable current matching viral post. The copy generation model then dynamically generates a multi-source knowledge fusion context corresponding to the current matching viral post. Based on the multi-source knowledge fusion context, the current matching viral post is optimized to rapidly generate a target marketing copy that better meets marketing needs. Therefore, this application can improve the efficiency and quality of marketing copy generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0041] Figure 1 This is a schematic diagram of the first process of the first embodiment of the intelligent marketing copy generation method of this application;

[0042] Figure 2 This is a second flow chart of the first embodiment of the intelligent marketing copy generation method of this application;

[0043] Figure 3 This is a third flow chart of the first embodiment of the intelligent marketing copy generation method of this application;

[0044] Figure 4 This is a flowchart of the second embodiment of the intelligent marketing copy generation method of this application;

[0045] Figure 5 This is a process diagram of the second embodiment of the intelligent marketing copy generation method of this application;

[0046] Figure 6 This is a schematic diagram of the module structure of the intelligent marketing copy generation device according to an embodiment of the present application;

[0047] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the intelligent marketing copy generation method in the embodiment of this application.

[0048] The purpose, features and advantages of this application will be further explained with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0049] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0050] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0051] The main solution of this application is: to parse the copy generation request input by the user to obtain marketing feature parameters; to generate the current matching hot article based on the marketing knowledge database and marketing feature parameters, to call the copy generation model to generate the multi-source knowledge fusion context corresponding to the current matching hot article; to generate the target marketing copy based on the current matching hot article and the multi-source knowledge fusion context.

[0052] The existing technology has the problem of insufficient knowledge integration and inability to quickly generate marketing copy that meets the needs. It is necessary to effectively improve the efficiency and quality of automatic generation of marketing copy.

[0053] This application is aimed at marketing copy generation scenarios, especially for industries such as finance, e-commerce, and fast-moving consumer goods, and proposes a marketing copy generation method and system that integrates a marketing knowledge base, dynamic RAG retrieval, and multi-level verification to solve the above problems. This application first receives a copy generation request input by the user, and then parses the marketing feature parameters from the request. Then, based on the marketing feature parameters, it searches the marketing knowledge database to generate the current matching hot article. Then, it calls the copy generation model and lets the model generate a dynamic multi-source knowledge fusion context based on the current matching hot article. Finally, it combines the current matching hot article and the multi-source knowledge fusion context to generate the target marketing copy. Therefore, this application obtains key marketing parameters by parsing the request and performs a static search on the database composed of marketing meta-knowledge to obtain the most suitable current matching hot article; then, it uses the copy generation model to dynamically generate the multi-source knowledge fusion context corresponding to the current matching hot article, and optimizes the current matching hot article based on the multi-source knowledge fusion context to quickly generate a target marketing copy that better meets marketing needs. Therefore, this application can effectively improve the generation efficiency and quality of intelligent marketing copy.

[0054] It should be noted that the execution entity of this embodiment can be an intelligent marketing copy generation system, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an intelligent marketing copy generation device capable of performing the above functions, etc. This embodiment does not specifically limit this. The following uses the intelligent marketing copy generation device (hereinafter referred to as the generation device) as an example to illustrate this embodiment and the following embodiments.

[0055] Based on this, the embodiment of the present application provides a method for generating intelligent marketing copy. Figure 1 , Figure 1 This is a first flow chart of the first embodiment of the intelligent marketing copy generation method of this application.

[0056] In this embodiment, the intelligent marketing copy generation method includes steps S10 to S40:

[0057] Step S10: parsing the copywriting generation request input by the user to obtain marketing feature parameters;

[0058] It's easy to understand that the copywriting request input by the user can be an instruction to the system to generate marketing copy, including marketing requirements such as product information, target audience, and marketing scenarios. The marketing feature parameters can be key data extracted from the user request to describe the marketing requirements, such as product keywords, target audience characteristics, and marketing campaign themes. These parameters can be used for matching and adjusting subsequent copywriting generation.

[0059] Step S20: generating a currently matching viral article based on the marketing knowledge database and the marketing characteristic parameters;

[0060] It's important to understand that existing intelligent marketing copywriting methods suffer from knowledge fragmentation and rigid knowledge integration. Specifically, existing general-purpose models lack a deep understanding of a company's knowledge, particularly regarding brand, product, and user knowledge. This results in a lack of integration between the generated copywriting and content like brand identity and product selling points. Therefore, in this embodiment, the marketing knowledge database can be a database that stores various marketing-related knowledge, including brand information, product features, historical viral articles, industry regulations, and other content, to provide data support for subsequent copywriting generation.

[0061] The above-mentioned currently matching hot articles can be historical hot articles that match the current marketing needs to a certain extent after the generating device performs a static search on the marketing knowledge database based on marketing characteristic parameters. These hot articles can be used as a reference basis for the subsequent generation of target marketing copy.

[0062] Step S30: Calling a copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching viral article;

[0063] Step S40: Generate a target marketing copy based on the current matching viral article and the multi-source knowledge fusion context.

[0064] It's understood that the copywriting generation model described above can be a trained artificial intelligence model capable of generating textual content based on input information. In this solution, it is used to generate a multi-source knowledge fusion context corresponding to the currently matching viral post. This multi-source knowledge fusion context integrates knowledge from multiple textual contexts, including brand knowledge, product knowledge, and viral post knowledge, to provide rich information support for generating targeted marketing copy.

[0065] It is easy to understand that the final generation device can optimize the current matching hot articles based on the multi-source knowledge fusion context and generate copy that meets the user's marketing needs, that is, the above-mentioned target marketing copy, which comprehensively considers various marketing factors and knowledge.

[0066] In a feasible implementation manner, the marketing feature parameters include marketing semantic vectors and marketing keywords; Figure 2 , Figure 2 This is a second flow chart of the first embodiment of the intelligent marketing copy generation method of this application. In this embodiment, step S20 may include step S21:

[0067] Step S21 , performing a mixed search on the marketing knowledge database according to the marketing semantic vector and the marketing keyword to obtain the current matching hot articles recalled in multiple ways.

[0068] It's easy to understand that to improve the accuracy and efficiency of static retrieval databases, this embodiment can perform multi-way recall based on marketing feature parameters. In this case, the marketing semantic vector can be generated by vectorizing the marketing-related text information in the copy generation request. This feature vector can be used to represent the semantic features of the text, making it easier for the generating device to calculate and match, thereby improving retrieval granularity and efficiency.

[0069] These marketing keywords can be important marketing-related terms included in the copywriting request, such as product names, brand names, and marketing campaign keywords. They can be used to accurately locate and filter information in the marketing knowledge database, improving search accuracy. The classic BM25 algorithm can be used as a marketing keyword search method in this process.

[0070] Therefore, in this embodiment, the generating device can perform a traversal search for hot articles in the marketing knowledge database based on the comprehensive use of two retrieval methods, namely, a dual-path hot article search based on marketing semantic vectors and a marketing keyword-based retrieval, and then quickly obtain the current matching hot articles that best match the marketing feature parameters from the marketing knowledge database multiple times to improve the comprehensiveness and accuracy of the retrieval.

[0071] In this implementation, the method of obtaining the current matching hot articles from the marketing knowledge database is further optimized. Through hybrid retrieval and multi-way recall, the current matching hot articles that match the marketing needs are obtained more comprehensively and accurately from the marketing knowledge database, solving the problem of incomplete and inaccurate retrieval results that may be caused by a single retrieval method, thereby providing a better reference for subsequent copy generation and improving the quality of copy generation.

[0072] In one possible implementation, refer to Figure 3 , Figure 3 This is a third flow chart of the first embodiment of the intelligent marketing copy generation method of this application. In this embodiment, step S40 may include steps A1 to A3:

[0073] Step A1: generating an initial marketing copy based on the currently matching viral article and the multi-source knowledge fusion context;

[0074] Step A2: Perform multi-level compliance verification on the copy generation request and the initial marketing copy;

[0075] Step A3: determining the target marketing copy based on the verification result and the initial marketing copy.

[0076] It should be noted that the existing marketing copy generated based on the general large model has compliance risks. The root cause is that the traditional rule engine relies only on keyword interception, which has low interception efficiency and lacks the integration and fusion of the rule library of the marketing industry and the corporate compliance specification library. It is unable to understand the contextual semantics of the user input request or the generated marketing copy, and there is a risk of missing illegal words in the generated copy. Therefore, this embodiment can perform compliance checks and verifications on the copy generation request input by the user and the initial marketing copy generated based on the current matching hot article and multi-source knowledge fusion context at multiple levels (such as banned words, semantic violations, multimodal content security, industry specifications, etc.), that is, perform the above-mentioned multi-level compliance verification to ensure that the copy complies with laws and regulations, industry specifications and the internal standards of the enterprise, and adjust the initial marketing copy or directly adopt it according to the verification results to obtain the final compliant target marketing copy.

[0077] For example, this embodiment's multi-level compliance verification may include: Level 1 verification: Regular expression-based banned word blocking; Level 2 verification: BERT-based semantic violation detection; Level 3 verification: Multimodal content security assessment based on a marketing knowledge database; Level 4 verification: Cross-validation of industry standards against user-entered copy generation requests; and Real-time policy verification: Establishing a dynamic blacklist and whitelist mechanism to synchronize regulatory policy updates in real time. Therefore, this embodiment, based on a comprehensive multi-path compliance verification strategy, achieves better compliance review results and generates safer intelligent marketing copy.

[0078] This embodiment provides an intelligent marketing copy generation method, comprising: parsing a copy generation request input by a user to obtain marketing feature parameters; the marketing feature parameters include marketing semantic vectors and marketing keywords; performing a hybrid search of a marketing knowledge database based on the marketing semantic vectors and marketing keywords to obtain a multi-way recall of a currently matching viral post; invoking a copy generation model to generate a multi-source knowledge fusion context corresponding to the currently matching viral post; generating a target marketing copy based on the currently matching viral post and the multi-source knowledge fusion context; generating an initial marketing copy based on the currently matching viral post and the multi-source knowledge fusion context; performing a multi-level compliance check on the copy generation request and the initial marketing copy; and determining a target marketing copy based on the check results and the initial marketing copy. This embodiment first receives a copy generation request input by a user, then parses the request to obtain marketing feature parameters. Based on the marketing feature parameters, a search is performed in the marketing knowledge database to generate a currently matching viral post. The copy generation model is then invoked to generate a dynamic multi-source knowledge fusion context based on the currently matching viral post. Finally, the target marketing copy is generated by combining the currently matching viral post and the multi-source knowledge fusion context. Therefore, this embodiment parses the request to obtain key marketing parameters, performs a static search on the database of marketing meta-knowledge, and obtains the most suitable matching viral article. It then uses the copywriting generation model to dynamically generate a multi-source knowledge fusion context corresponding to the currently matching viral article. Based on this multi-source knowledge fusion context, it optimizes the currently matching viral article and quickly generates a target marketing copy that better meets marketing needs. Therefore, this embodiment can improve the efficiency and quality of marketing copywriting generation.

[0079] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction and will not be repeated later.

[0080] Based on the first embodiment, please refer to Figure 4 , Figure 4 This is a flow chart of the second embodiment of the intelligent marketing copy generation method of this application. In this embodiment, before step S30, the intelligent marketing copy generation method further includes steps B1 to B3:

[0081] Step B1: construct an initial sample dataset based on marketing meta-knowledge data; the marketing meta-knowledge data includes brand attribute data, product feature data, and viral article data;

[0082] It is understandable that in order to generate marketing copy that is highly consistent with the brand, it is necessary not only to consider the brand tone and product parameters of the enterprise, but also to consider the integration of external trend data such as explosive articles, hot spots and other information. Existing solutions lack effective structured methods and effective integration methods, and the generated content marketing effect is poor. Therefore, in order to solve the problem of low quality of existing model training data, in this embodiment, relevant meta-knowledge of marketing can be constructed to construct and optimize the copy generation model to achieve the effect of improving the quality of the copy. Therefore, the above-mentioned marketing meta-knowledge data can be marketing basic knowledge and general knowledge data to improve the relevance of subsequent intelligent marketing copy to the brand or industry, which may include brand attribute data, such as brand name, brand story, brand tone and brand audience; product feature data, such as product name, product introduction, product function, product selling point, etc.; explosive article data, that is, historical explosive articles and real-time explosive article related information, etc.

[0083] It should be understood that the above-mentioned marketing meta-knowledge data can also be integrated into the retrieval enhancement process, that is, the brand knowledge base, product knowledge base and hot article knowledge base included in the marketing knowledge database in the above-mentioned embodiment 1, and its specific data fields can be consistent with the brand attribute data, product feature data and hot article data contained in the above-mentioned marketing meta-knowledge data. In addition, the existing method of generating intelligent marketing copy based on a large language model has the defect of controlling the granularity too coarsely, and the generated intelligent marketing copy lacks fine-grained style constraints (such as language style, emotional tendency, etc.), and the generated content does not match the user's intention or has a large deviation. Therefore, in a feasible implementation method, in this embodiment, the marketing meta-knowledge data also includes:

[0084] Audience profile data, including age, gender, and content interest tags;

[0085] Channel characteristic data, including the dissemination rules and content format requirements of different platforms;

[0086] Copywriting style data, including language style type and emotional tendency parameters.

[0087] It is understood that the above-mentioned audience profile data can be a data set describing the characteristics of the target audience, including age, gender, content interests, and income level labels, etc., which is used to accurately locate the target audience and make the generated copy more in line with the audience's needs; channel feature data can be data related to different marketing platforms, including the dissemination rules of each platform (such as release time, promotion method, etc.) and content form requirements (such as word limit, format requirements, etc.), to ensure that the copy can be effectively disseminated across different channels; copy style data can be data about the copy language style type (including but not limited to colloquialism, professional authority, emotional resonance, suspense reasoning, irony humor, poetic aesthetics, storytelling, comparative evaluation, etc.), emotional tendency parameters (such as positive, negative, etc.) and content form (including but not limited to seeding, tutorials, etc.), used to control the language style, emotional expression and content form of the copy. Therefore, this embodiment can further expand the marketing meta-knowledge data so that the copy generation process can consider more marketing factors and improve the relevance and adaptability of the copy. More comprehensive marketing meta-knowledge data helps to generate marketing copy that is more in line with the preferences of the target audience, adapts to different marketing channels, and has a specific style, which can further improve marketing effectiveness.

[0088] Step B2, dynamically optimizing the initial sample data set to obtain an optimized sample data set;

[0089] Step B3: constructing a copywriting generation model based on the optimized sample data set.

[0090] It will be readily understood that this embodiment constructs an initial data set for training the copywriting generation model based on the aforementioned marketing meta-knowledge data. This data set contains samples of various marketing-related information, namely, the aforementioned initial sample dataset. This initial sample dataset is then adjusted and improved according to specific rules and algorithms to better meet the requirements of model training, thereby obtaining an optimized sample dataset. Compared to the initial sample dataset, this dynamically optimized sample dataset enables the model to better learn marketing-related knowledge, effectively improving the quality and practicality of the marketing copy output by the copywriting generation model.

[0091] In a feasible implementation, the viral article data includes historical viral articles and real-time viral articles. In this embodiment, step B2 may include steps B21 to B22:

[0092] Step B21: Perform a preset time-decay weighted calculation on the historical hot posts and the real-time hot posts to generate a dynamic hot post weight;

[0093] Step B22: Optimize the initial sample data set based on the dynamic hot article weight to obtain an optimized sample data set.

[0094] It should be noted that the static knowledge base of existing large language models cannot effectively capture the characteristics and trending events of social media viral posts. There are problems with knowledge base updates being delayed and unable to capture sudden trends in real time. As a result, the generated intelligent marketing copy lacks integration with dynamic hot spots and lacks the explosive power of dissemination. Therefore, in this embodiment, the viral post data includes but is not limited to the content structure characteristics, keyword characteristics, popular tags, titles, and release time of historical and real-time viral posts. Keywords are primarily keywords extracted based on TF-IDF over the past N days (N is between 1 and 30 days); popular tags are primarily top-10 tags ranked based on page views and interaction volume, with N ranging from 1 to 30. In this embodiment, a preset time-decay weighted operation can be performed on historical and real-time viral posts in the viral post data, and the weights of historical and real-time viral posts in the viral post data can be dynamically adjusted based on real-time performance. The initial sample data set is then optimized based on the adjusted dynamic viral post weights, so that the generation device can better capture hot trends based on the dynamically optimized knowledge base and generate more disseminating marketing copy.

[0095] It is understandable that, when performing the preset time decay weighting calculation, this embodiment can perform a weighted calculation on historical and real-time hot articles according to the preset time decay rules to reflect the timeliness of hot articles, where the hot articles closer to the current time have a higher weight. Therefore, in a feasible embodiment, the hot article data includes historical and real-time hot articles; in this embodiment, step B21 can include steps B211 to B212:

[0096] Step B211, obtaining the current publishing time interval corresponding to the historical viral article and the real-time viral article;

[0097] Step B212: Perform a preset time-attenuation weighted operation on the historical hot articles and the real-time hot articles based on the current publishing time interval to generate a dynamic hot article weight.

[0098] It is easy to understand that the above-mentioned current release time interval can be the time difference between the release time of historical or real-time hot articles and the current time, which is used to measure the newness of the hot article. After the generating device reads the current release time interval corresponding to the historical and real-time hot articles from the relevant data records, it can perform a preset time decay weighted operation based on the current release time interval. This preset time decay weighted operation uses the improved Newton's cooling law. The corresponding formula is as follows:

[0099] Wd=e^(-λ·Δt)×(1+θ·e^(-k·Δt)); (1)

[0100] Where Wd is the dynamic explosive article weight corresponding to the current historical / real-time explosive article; Δt is the difference between the current time of the historical / real-time explosive article and the time when the content was released (in days); λ is the basic decay rate (which can take values ​​from 0.05 to 0.6); k is the real-time decay acceleration factor (which can take values ​​from 0.1 to 1.0); and θ is the real-time content enhancement coefficient (which can take values ​​from 0.5 to 0.9).

[0101] In this embodiment, the sample dataset can be more rationally optimized by accurately calculating the dynamic weights of historical and real-time viral articles, improving the copywriting generation model's ability to grasp the timeliness of viral articles. Furthermore, the dynamic retrieval capabilities of the marketing knowledge database constructed based on the optimized sample dataset can also be enhanced, further improving the real-time performance of the generation device's combined static and dynamic multi-channel recall of content. Therefore, this embodiment can migrate viral article structures based on online search and knowledge base search, and achieve knowledge fusion and migration through the time factor, improving the real-time performance of the generated target marketing copy.

[0102] In summary, refer to Figure 5 The process of generating intelligent marketing copy in this embodiment is explained. Figure 5 This is a process diagram of the second embodiment of the intelligent marketing copy generation method of this application, as shown in FIG. Figure 5 As shown, the intelligent marketing copy generation process of this embodiment may include four steps:

[0103] 1) Marketing meta-knowledge driven:

[0104] To address the low quality of existing model training data, this embodiment constructs marketing meta-knowledge data to build and optimize the copywriting generation model and improve the model's control granularity. The marketing meta-knowledge constructed in this embodiment may include extracting core brand elements, including brand knowledge about the brand story, visual identification system, and target audience portraits; establishing a functional attribute matrix and a list of differentiated selling points, summarizing product knowledge about product technical parameters; collecting characteristics of viral social media posts, including viral knowledge about dissemination paths, interaction indicators, and content templates; analyzing target audience behavior data to obtain audience knowledge about demographic characteristics and content preference tags; and formulating copywriting specifications, covering channel knowledge and copywriting knowledge about language style, narrative structure, and distribution channel adaptation rules.

[0105] 2) Model fine-tuning and enhancement:

[0106] This embodiment can first construct an initial sample data set based on marketing meta-knowledge. At the same time, in order to improve the sensitivity and timeliness of response to hot events, a preset time decay weighted operation is performed based on the real-time nature of historical hot articles and real-time hot articles. The initial sample data set is optimized based on the dynamic hot article weights generated after the operation to obtain an optimized sample data set. The model is then trained and strengthened based on the optimized sample data set to obtain an iterative copywriting generation model.

[0107] 3) Marketing Knowledge Base:

[0108] This embodiment can also build a domain-specific corpus, integrate the marketing meta-knowledge corresponding to brand documents, product manuals and compliance guidelines to build a marketing knowledge database, and subsequently perform real-time matching of explosive articles based on the marketing knowledge database.

[0109] 4) Intelligent marketing copy generation:

[0110] The generation device can first perform a hierarchical compliance check on the copy generation request input by the user (i.e., the aforementioned multi-level compliance check). It then dynamically searches and enhances the marketing knowledge database based on the marketing feature parameters obtained after parsing the copy generation request. Specifically, during the search process, it can retrieve multiple matching hot articles through keyword search and marketing vector search based on the dynamic weights of real-time and historical hot articles.

[0111] Then, based on the copywriting generation model, the dynamic multi-source knowledge fusion context corresponding to the currently matching viral article is obtained. During this process, content optimization can be performed based on the dynamic weights of real-time viral articles and historical viral articles.

[0112] Finally, based on the current matching viral post and the multi-source knowledge fusion context, dynamic cross-modal information fusion is achieved to generate a highly real-time initial marketing copy. This initial marketing copy is also subject to layered compliance verification. If and only if both the copy generation request and the layered compliance verification of the initial marketing copy pass, the generated initial marketing copy is determined as the target marketing copy and fed back to the user.

[0113] It should be noted that compared with the traditional method of generating marketing copy based on a large language model, the target marketing copy generated by this embodiment has an improvement of more than 15% in content acceptance and more than 10% in the utilization rate of diversified knowledge.

[0114] In summary, to address the low quality of existing model training data, this embodiment constructs marketing meta-knowledge data to improve the relevance and adaptability of model-generated copy. This allows for the generation of marketing copy that better aligns with target audience preferences, adapts to different marketing channels, and possesses a specific style, further enhancing marketing effectiveness. Furthermore, this embodiment can migrate viral content structures based on online and knowledge base searches, and utilizes time factors to achieve knowledge fusion and migration, improving the real-time nature of generated target marketing copy.

[0115] This embodiment discloses constructing an initial sample dataset based on marketing meta-knowledge data; the marketing meta-knowledge data includes brand attribute data, product feature data, and viral article data; the viral article data includes historical viral articles and real-time viral articles; obtaining the current release time intervals corresponding to the historical viral articles and real-time viral articles; performing a preset time-decay weighting operation on the historical viral articles and real-time viral articles based on the current release time interval to generate dynamic viral article weights; and optimizing the initial sample dataset based on the dynamic viral article weights to obtain an optimized sample dataset. A copywriting generation model is constructed based on the optimized sample dataset. To address the low quality of training data for existing models, this embodiment constructs marketing meta-knowledge data to improve the relevance and adaptability of model-generated copywriting. This allows for the generation of marketing copywriting that better suits the preferences of the target audience, adapts to different marketing channels, and possesses a specific style, further enhancing marketing effectiveness. Furthermore, this embodiment can migrate viral article structures based on online and knowledge base searches, and integrate and migrate knowledge using a time factor to improve the real-time nature of generated target marketing copy.

[0116] It should be noted that the above examples are only used to understand this application and do not constitute a limitation on the intelligent marketing copy generation method of this application. More simple transformations based on this technical concept are all within the scope of protection of this application.

[0117] This application also provides an intelligent marketing copy generation device, please refer to Figure 6 , Figure 6 This is a schematic diagram of the module structure of the intelligent marketing copy generation device according to an embodiment of the present application. In this embodiment, the intelligent marketing copy generation device includes:

[0118] Request parsing module 601, used to parse the copywriting generation request input by the user to obtain marketing feature parameters;

[0119] Static matching module 602, for generating current matching viral articles based on the marketing knowledge database and the marketing feature parameters;

[0120] Dynamic optimization module 603, configured to call a copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching viral article;

[0121] The copywriting generation module 604 is configured to generate a target marketing copywriting based on the current matching viral article and the multi-source knowledge fusion context.

[0122] As an implementable method, in this embodiment, the marketing feature parameters include marketing semantic vectors and marketing keywords; the static matching module 602 is also used to perform a mixed search on the marketing knowledge database based on the marketing semantic vectors and the marketing keywords to obtain the current matching hot articles of multi-channel recall.

[0123] As an implementable method, in this embodiment, the dynamic optimization module 603 is also used to construct an initial sample data set based on marketing meta-knowledge data, and the marketing meta-knowledge data includes brand attribute data, product feature data and hot article data; dynamically optimize the initial sample data set to obtain an optimized sample data set; and construct a copy generation model based on the optimized sample data set.

[0124] As an implementable method, in this embodiment, the hot article data includes historical hot articles and real-time hot articles; the dynamic optimization module 603 is further used to perform a preset time-attenuated weighted operation on the historical hot articles and the real-time hot articles to generate dynamic hot article weights, and optimize the initial sample data set based on the dynamic hot article weights.

[0125] As an implementable method, in this embodiment, the dynamic optimization module 603 is further used to obtain the current publishing time interval corresponding to the historical hot articles and the real-time hot articles, and perform a preset time decay weighting operation based on the time interval to generate a dynamic hot article weight.

[0126] As an implementable method, in this embodiment, the marketing meta-knowledge data also includes: audience portrait data (including age, gender and content interest tags), channel feature data (including communication rules and content form requirements of different platforms), and copywriting style data (including language style type and emotional tendency parameters).

[0127] As an implementable method, in this embodiment, the copy generation module 604 is also used to generate initial marketing copy based on the current matching hot article and multi-source knowledge fusion context, and perform multi-level compliance verification on the copy generation request and the initial marketing copy, and determine the target marketing copy based on the verification results.

[0128] The intelligent marketing copy generation device provided in this application adopts the intelligent marketing copy generation method in the above-mentioned embodiment, which can solve the technical problems of single static data matching and poor adaptability to dynamic scenarios in traditional marketing copy generation. Compared with the existing technology, the device provided in this application achieves high-matching and highly compliant marketing copy generation through multi-source knowledge fusion and dynamic weight optimization. Its beneficial effects are the same as those of the above-mentioned method embodiment, and the other technical features of the device are the same as those disclosed in the above-mentioned method, which will not be repeated here.

[0129] The present application provides an intelligent marketing copy generation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the intelligent marketing copy generation method in the above-mentioned embodiment one.

[0130] Reference below Figure 7 , which shows a schematic diagram of the structure of an intelligent marketing copy generation device suitable for implementing the embodiments of the present application. The intelligent marketing copy generation device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (such as in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The intelligent marketing copy generation device shown is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0131] like Figure 7 As shown, the intelligent marketing copy generation device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the intelligent marketing copy generation device are also stored in the random access memory 1004. The processing device 1001, the read-only memory 1002 and the random access memory 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the intelligent marketing copy generation device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows an intelligent marketing copy generation device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have instead.

[0132] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include an intelligent marketing copy generation program product, which includes an intelligent marketing copy generation program carried on a computer-readable medium, and the intelligent marketing copy generation program contains program code for executing the method shown in the flowchart. In such an embodiment, the intelligent marketing copy generation program can be downloaded and installed from the network through a communication device, or installed from the storage device 1003, or installed from the read-only memory 1002. When the intelligent marketing copy generation program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are performed.

[0133] The intelligent marketing copy generation device provided in this application utilizes the intelligent marketing copy generation method of the aforementioned embodiment to solve the technical problems associated with intelligent marketing copy generation. Compared to the prior art, the beneficial effects of the intelligent marketing copy generation device provided in this application are the same as those of the intelligent marketing copy generation method provided in the aforementioned embodiment. Other technical features of the intelligent marketing copy generation device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0134] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0135] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0136] The present application provides a storage medium having computer-readable program instructions (i.e., an intelligent marketing copy generation program) stored thereon, and the computer-readable program instructions are used to execute the intelligent marketing copy generation method in the above-mentioned embodiment.

[0137] The storage medium provided in this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination thereof. More specific examples of storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the storage medium can be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0138] The above-mentioned storage medium may be included in the intelligent marketing copy generation device; or it may exist independently without being assembled into the intelligent marketing copy generation device.

[0139] The above-mentioned storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the intelligent marketing copy generation device, the intelligent marketing copy generation device enables: intelligent marketing copy generation.

[0140] The intelligent marketing copy generation program code for performing the operations of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0141] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and intelligent marketing copy generation program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a portion of code, which contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the boxes can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, as well as the combination of boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified functions or operations, or can be implemented with a combination of dedicated hardware and computer instructions.

[0142] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0143] The readable storage medium provided in this application is a storage medium, which stores computer-readable program instructions (i.e., an intelligent marketing copy generation program) for executing the above-mentioned intelligent marketing copy generation method, which can solve the technical problems of single static data matching and poor adaptability to dynamic scenarios in traditional marketing copy generation. Compared with the prior art, the device provided in this application achieves high-matching and highly compliant marketing copy generation through multi-source knowledge fusion and dynamic weight optimization. Its beneficial effects are the same as those in the above-mentioned method embodiment, and other technical features in the device are the same as those disclosed in the above-mentioned method, which will not be repeated here.

[0144] The above are only some embodiments of the present application and are not intended to limit the patent scope of the present application. All equivalent structural transformations made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for generating intelligent marketing copy, characterized in that: The method comprises: Parse the copywriting generation request input by the user to obtain marketing feature parameters; Generate a current matching viral article based on the marketing knowledge database and the marketing feature parameters; Calling the copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching viral article; Generate target marketing copy based on the current matching hot article and the multi-source knowledge fusion context.

2. The method for generating intelligent marketing copy according to claim 1, wherein: The marketing feature parameters include marketing semantic vectors and marketing keywords; The step of generating the current matching viral article based on the marketing knowledge database and the marketing characteristic parameters includes: A mixed search is performed on the marketing knowledge database according to the marketing semantic vector and the marketing keyword to obtain the current matching hot articles recalled in multiple ways.

3. The method for generating intelligent marketing copy according to claim 1, wherein: Before calling the copywriting generation model to generate the multi-source knowledge fusion context corresponding to the currently matching viral article, the method further includes: Construct an initial sample dataset based on marketing meta-knowledge data; the marketing meta-knowledge data includes brand attribute data, product feature data, and viral article data; Dynamically optimizing the initial sample data set to obtain an optimized sample data set; A copywriting generation model is constructed based on the optimized sample data set.

4. The method for generating intelligent marketing copy according to claim 3, wherein: The hot article data includes historical hot articles and real-time hot articles; The step of dynamically optimizing the initial sample data set to obtain an optimized sample data set includes: Performing a preset time-attenuated weighted operation on the historical hot posts and the real-time hot posts to generate a dynamic hot post weight; The initial sample data set is optimized based on the dynamic hot article weight to obtain an optimized sample data set.

5. The method for generating intelligent marketing copy according to claim 4, wherein: The step of performing a preset time-attenuated weighted operation on the historical viral articles and the real-time viral articles to generate dynamic viral article weights includes: Obtain the current publishing time interval corresponding to the historical hot article and the real-time hot article; A preset time decay weighting operation is performed on the historical hot articles and the real-time hot articles based on the current publishing time interval to generate a dynamic hot article weight.

6. The method for generating intelligent marketing copy according to claim 3, wherein: The marketing meta-knowledge data also includes: Audience profile data, including age, gender, and content interest tags; Channel characteristic data, including the dissemination rules and content format requirements of different platforms; Copywriting style data, including language style type and emotional tendency parameters.

7. The method for generating intelligent marketing copy according to claim 1, wherein: The step of generating a target marketing copy based on the current matching viral article and the multi-source knowledge fusion context includes: Generate an initial marketing copy based on the current matching viral article and the multi-source knowledge fusion context; Performing multi-level compliance checks on the copy generation request and the initial marketing copy; The target marketing copy is determined based on the verification result and the initial marketing copy.

8. An intelligent marketing copy generation device, characterized in that: The intelligent marketing copy generating device includes: The request parsing module is used to parse the copywriting generation request input by the user and obtain marketing feature parameters; A static matching module, configured to generate a currently matching viral article based on a marketing knowledge database and the marketing characteristic parameters; A dynamic optimization module is used to call the copywriting generation model to generate a multi-source knowledge fusion context corresponding to the currently matching explosive article; A copywriting generation module is used to generate target marketing copywriting based on the current matching hot article and the multi-source knowledge fusion context.

9. An intelligent marketing copy generation device, characterized in that: The intelligent marketing copy generation device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the intelligent marketing copy generation method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the intelligent marketing copy generation method according to any one of claims 1 to 7 are implemented.

Images