Marketing verbal skill generation method and system based on large model cue word engineering

Abstract

The invention provides a marketing verbal skill generation method and system based on large model cue word engineering, and relates to the technical field of big data processing and artificial intelligence. According to the method provided by the invention, historical data of a target user is acquired, wherein the historical data comprises identity information, hobby and interest information and historical consumption information; determining a user portrait label of the target user according to the historical data; determining marketing strategy information of the target user based on the user portrait label of the target user, wherein the marketing strategy information comprises marketing product information, marketing channel information and verbal skill rhetorical mode information; generating a verbal skill prompt text according to the user portrait label of the target user and the marketing strategy information; and inputting the verbal skill prompt text into the natural language generation model, and outputting a marketing verbal skill text of the target user. According to the method, the marketing verbal skill text corresponding to the target user can be generated from multiple dimensions, the marketing verbal skill generation efficiency and the individuation degree can be improved, and the use requirements of the user in different use scenes can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing and artificial intelligence, and particularly to a method and system for generating marketing copy based on large model prompt engineering. Background Art

[0002] In the current marketing environment, the formulation of marketing copy has a crucial impact on marketing effects. Traditional ways of generating marketing copy often rely on the personal experience and creativity of marketers. This way is not only inefficient but also difficult to ensure the quality and consistency of the copy. With the development of big data and artificial intelligence technologies, more and more enterprises have begun to try to use these technologies to optimize the process of generating marketing copy.

[0003] In related technologies, marketing copy is usually generated by methods such as keyword replacement and template filling. Although these methods are simple, they lack personalization and innovation. Subsequently, machine learning-based models began to be introduced. Although they can generate relatively rich copy, the model training and maintenance costs are high, and it is difficult to adapt to rapidly changing marketing needs.

[0004] Therefore, there is an urgent need for a method and system for generating marketing copy based on large model prompt engineering to solve the problems of low efficiency and insufficient personalization in generating marketing copy and meet the usage requirements of users in different scenarios. Summary of the Invention

[0005] Embodiments of the present invention provide a method and system for generating marketing copy based on large model prompt engineering, which can improve the efficiency and personalization of generating marketing copy and meet the usage requirements of users in different usage scenarios.

[0006] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0007] In a first aspect, a method for generating marketing copy based on large model prompt engineering is provided. The method includes: obtaining historical data of a target user, where the historical data includes identity information, interest and hobby information, and historical consumption information; determining user portrait tags of the target user according to the historical data; determining marketing strategy information of the target user based on the user portrait tags of the target user, where the marketing strategy information includes marketing product information, marketing channel information, and speech rhetoric method information; generating a speech prompt text according to the user portrait tags and marketing strategy information of the target user; and inputting the speech prompt text into a natural language generation model to output a marketing copy text of the target user.

[0008] In a possible implementation of the first aspect, determining user portrait tags of a target user based on historical data includes: preprocessing the historical data to obtain feature information corresponding to the target user, where the preprocessing includes data cleaning, normalization, and feature extraction; obtaining user portrait tags of the target user based on the feature information corresponding to the target user.

[0009] In a possible implementation of the first aspect, before obtaining the historical data of the target user, the above method further includes: obtaining a basic data set, where the basic data set includes multiple user portrait tags and multiple marketing strategy information, and each marketing strategy information includes one or more different items among marketing product information, marketing channel information, and speech rhetoric method information; establishing a first database based on the basic data set, where the first database includes user portrait tags of multiple users and marketing strategy information corresponding to the user portrait tags of each user.

[0010] In a possible implementation of the first aspect, determining marketing strategy information of the target user based on the user portrait tags of the target user includes: obtaining marketing strategy information corresponding to the user portrait tags of the target user from the first database according to the user portrait tags of the target user.

[0011] In a possible implementation of the first aspect, generating a speech prompt text according to the user portrait tags and marketing strategy information of the target user includes: establishing a prompt word text template, where the prompt word text template includes multiple placeholder information and auxiliary information, and each placeholder information is respectively configured with a different position identifier, and the position identifier is used to represent user portrait tags, marketing product information, marketing channel information, or speech rhetoric method information; replacing the placeholder information included in the prompt word text template according to the user portrait tags, marketing product information, marketing channel information, or speech rhetoric method information of the target user to obtain a speech prompt text.

[0012] In a possible implementation of the first aspect, before obtaining the historical data of the target user, the above method further includes: obtaining multiple user portrait tags and marketing speech examples corresponding to each user portrait tag; establishing a second database according to the marketing speech examples corresponding to each user portrait tag; replacing the placeholder information included in the prompt word text template according to the user portrait tags, marketing product information, marketing channel information, or speech rhetoric method information of the target user to obtain a speech prompt text, including: determining a marketing speech example from the second database according to the user portrait tags of the target user; replacing the placeholder information included in the prompt word text template according to the user portrait tags, marketing product information, marketing channel information, or speech rhetoric method information of the target user, and adding a marketing speech example to the prompt word text template to obtain a speech prompt text.

[0013] In a possible implementation of the first aspect, the above method further includes: obtaining a plurality of marketing script texts and the evaluation scores corresponding to each marketing script text, where the evaluation scores are used to characterize the semantic relevance, fluency, and conversion rate of the marketing script texts; iteratively training the natural language generation model based on the plurality of marketing script texts and the evaluation scores corresponding to each marketing script text to obtain a trained natural language generation model.

[0014] In a possible implementation of the first aspect, after inputting the script prompt text into the natural language generation model and outputting the marketing script text of the target user, the above method further includes: performing grammar checking and semantic correction on the marketing script text of the target user.

[0015] The beneficial effects of the present invention are as follows: The method provided by the present invention generates a script prompt text according to the user portrait tags and marketing strategy information of the target user. Since the marketing strategy information includes marketing product information, marketing channel information, or script rhetoric method information, therefore, it is possible to generate the marketing script text corresponding to the target user from multiple dimensions, which can improve the generation efficiency and personalization degree of the marketing script and meet the usage requirements of users in different usage scenarios. It can also be understood that: The method provided by the present invention can improve the generation efficiency and personalization degree of the marketing script. Through intelligent prompt word template construction and large model invocation, high-quality scripts that meet marketing requirements can be quickly generated. At the same time, since the method supports real-time update and optimization mechanisms, it can adapt to rapidly changing marketing requirements and maintain the continuous effectiveness and competitiveness of the system.

[0016] In a second aspect, an embodiment of the present invention provides a marketing script generation system based on large model prompt engineering. The above system includes: a data acquisition unit for acquiring the historical data of the target user, where the historical data includes identity information, interest and hobby information, and historical consumption information; a label determination unit for determining the user portrait tags of the target user according to the historical data; a strategy determination unit for determining the marketing strategy information of the target user based on the user portrait tags of the target user, where the marketing strategy information includes marketing product information, marketing channel information, and script rhetoric method information; a prompt text generation unit for generating a script prompt text according to the user portrait tags and marketing strategy information of the target user; and a marketing script generation unit for inputting the script prompt text into the natural language generation model and outputting the marketing script text of the target user.

[0017] In a third aspect, an electronic device is provided, which includes a memory and one or more processors; the memory is coupled to the processors; wherein, computer program code is stored in the memory, and the computer program code includes computer instructions. When the computer instructions are executed by the processors, the electronic device is caused to execute the method in any implementation manner of the first aspect.

[0018] In a fourth aspect, a computer-readable storage medium is provided, which includes computer instructions. When the computer instructions run on an electronic device, the electronic device is caused to execute the method in any implementation manner of the first aspect.

[0019] In a fifth aspect, a computer program product is provided. When the computer program product runs on a computer, the computer is caused to execute the method in any implementation manner of the first aspect.

[0020] It can be understood that for the beneficial effects that can be achieved by the system in the second aspect, the electronic device in the third aspect, the computer-readable storage medium in the fourth aspect, and the computer program product in the fifth aspect provided above, reference can be made to the beneficial effects in the first aspect and any possible design manner thereof, which will not be elaborated here. Description of the Drawings

[0021] Figure 1 It is a schematic hardware structure diagram of an electronic device shown in an embodiment of the present invention;

[0022] Figure 2 It is a flowchart of a marketing copy generation method based on large model prompt engineering shown in an embodiment of the present invention;

[0023] Figure 3 It is a flowchart of another marketing copy generation method based on large model prompt engineering shown in an embodiment of the present invention;

[0024] Figure 4 It is a flowchart of yet another marketing copy generation method based on large model prompt engineering shown in an embodiment of the present invention;

[0025] Figure 5 It is a flowchart of yet another marketing copy generation method based on large model prompt engineering shown in an embodiment of the present invention;

[0026] Figure 6 It is a schematic hardware structure diagram of a generation system shown in an embodiment of the present invention. Detailed Embodiments

[0027] The technical solutions in the embodiments of the present invention will be described below in conjunction with the accompanying drawings in the embodiments of the present invention. Among them, in the description of the present invention, unless otherwise specified, " / " indicates that the objects associated before and after are an "or" relationship. For example, A / B may represent A or B; the "or" in the present invention is merely a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. And, in the description of the present invention, unless otherwise specified, "a plurality of" means two or more than two. "At least one (piece)" or similar expressions below refer to any combination of these items, including any combination of single item (piece) or plural items (pieces).

[0028] In addition, in order to facilitate a clear description of the technical solutions in the embodiments of the present invention, in the embodiments of the present invention, terms such as "first" and "second" are used to distinguish identical or similar items with basically the same functions and roles. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and terms such as "first" and "second" do not necessarily limit being different from each other.

[0029] At the same time, in the embodiments of the present invention, words such as "exemplary" or "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present invention should not be construed as being superior or more advantageous than other embodiments or design solutions. Exactly speaking, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific way for easy understanding.

[0030] In the current marketing environment, the formulation of marketing language has a crucial impact on marketing effects. The traditional way of generating marketing language often relies on the personal experience and creativity of marketers. This way is not only inefficient but also difficult to ensure the quality and consistency of the language. With the development of big data and artificial intelligence technologies, more and more enterprises begin to try to use these technologies to optimize the process of generating marketing language.

[0031] In related technologies, marketing language generation usually adopts methods such as keyword replacement and template filling. Although these methods are simple, they lack personalization and innovation. Subsequently, models based on machine learning began to be introduced. Although they can generate relatively rich language, the model training and maintenance costs are high, and it is difficult to adapt to the rapidly changing marketing needs.

[0032] Therefore, there is an urgent need for a marketing language generation method and system based on large model prompt engineering to solve the problems of low efficiency and insufficient personalization in marketing language generation and meet the usage requirements of users in different scenarios.

[0033] In view of this, an embodiment of the present invention provides a method for generating marketing speech based on a large model prompt word project, the method comprising: obtaining historical data of a target user, the historical data comprising identity information, interest and hobby information, and historical consumption information; determining a user portrait tag of the target user based on the historical data; determining the marketing strategy information of the target user based on the user portrait tag of the target user, the marketing strategy information comprising marketing product information, marketing channel information, and speech rhetoric information; generating a speech prompt text based on the user portrait tag and the marketing strategy information of the target user; inputting the speech prompt text into a natural language generation model, and outputting the marketing speech text of the target user.

[0034] The method provided by the present invention generates a speech prompt text according to the user portrait label and marketing strategy information of the target user. Since the marketing strategy information includes marketing product information, marketing channel information or speech rhetoric information, it is possible to generate a marketing speech text corresponding to the target user from multiple dimensions, which can improve the efficiency and personalization of marketing speech generation and meet the user's usage needs in different usage scenarios. It can also be understood that the method provided by the present invention can improve the generation efficiency and personalization of marketing speech. Through intelligent prompt word template construction and large model calling, high-quality speech that meets marketing needs can be quickly generated. At the same time, since the method supports real-time update and optimization mechanisms, it can adapt to rapidly changing marketing needs and maintain the continuous effectiveness and competitiveness of the system.

[0035] In some embodiments, a marketing speech generation method based on a large model prompt word engineering provided by an embodiment of the present invention can be executed by a marketing speech generation system 100 based on a large model prompt word engineering (hereinafter referred to as the generation system 100). As an example, the generation system 100 can be any electronic device 200 with data processing capabilities, such as a general-purpose computer, a personal computer, a laptop, a switch or a tablet computer, etc. The specific implementation method of the purchase list generation system 100 is not limited here.

[0036] Figure 1 The hardware structure diagram of the electronic device provided by the embodiment of the present invention is shown. The electronic device 200 includes a processor 210, a memory 220 and a communication interface 230.

[0037] The processor 210 may include one or more processing cores. The processor 210 is connected to various parts within the electronic device 200 using various interfaces and lines. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 220, and by invoking data stored in the memory 220, it performs various functions of the electronic device 200 and processes data. Optionally, the processor 210 may be implemented in at least one hardware form such as a Central Processing Unit (CPU), a graphics processing unit (GPU), a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA).

[0038] The memory 220 may include a random access memory (RAM), and may also include a read-only memory (ROM). Optionally, the memory 220 includes a non-transitory computer-readable storage medium. The memory 220 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 220 may include a storage program area. Among them, the storage program area may store instructions for implementing the operating system, instructions for implementing at least one function (such as a data acquisition function, a text generation function, etc.), instructions for implementing the above various method embodiments, and the like.

[0039] The communication interface 230 is used to communicate with other devices, equipment, or communication networks, such as data storage devices, image processing devices, or Ethernet, a radio access network (RAN), a wireless local area network (WLAN), etc.

[0040] In terms of physical implementation, the above-mentioned various devices (such as the processor 210, the memory 220, and the communication interface 230) may respectively be devices within the same device (such as a laptop computer). Or, at least two of them may be provided in the same device, that is, as different devices within a device, similar to the deployment method of devices or components in a distributed system.

[0041] It can be understood that the structure illustrated in this embodiment does not constitute a specific limitation on the electronic device 200. In other embodiments of the present invention, the electronic device 200 may include more or fewer components than shown, or combine certain components, or split certain components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0042] The following describes, in conjunction with the accompanying drawings of the specification, a marketing copywriting generation method provided by an embodiment of the present invention based on large model prompt engineering.

[0043] Figure 2 FIG. is a flowchart of a marketing copywriting generation method provided by an embodiment of the present invention based on large model prompt engineering. Optionally, this method may be executed by Figure 1 the illustrated electronic device 200, that is, executed by the generation system 100. This method may include the following steps:

[0044] S1. Obtain the historical data of the target user, where the historical data includes identity information, interest and hobby information, and historical consumption information.

[0045] Among them, the historical data may be obtained from a customer management system, the social media of the target user, or a social survey questionnaire of the target user. It should be noted that the historical data may also be data obtained from other sources. The present invention embodiment does not particularly limit the specific acquisition method and source of the historical data.

[0046] S2. Determine the user portrait label of the target user according to the historical data.

[0047] In some embodiments, referring to Figure 3 , the above S2 specifically includes:

[0048] S21. Perform preprocessing on the historical data to obtain the feature information corresponding to the target user. The preprocessing includes data cleaning, normalization, and feature extraction.

[0049] S22. Obtain the user portrait label of the target user according to the feature information corresponding to the target user.

[0050] Specifically, data cleaning is to remove duplicate data, fill in missing values, and correct incorrect data in the historical data to ensure the integrity and accuracy of the data. Normalization is to convert data from different sources and different formats into a unified format and standard for subsequent processing and analysis. Feature extraction is to extract feature information related to marketing copywriting generation from the historical data, such as identity information, interest and hobby information, and historical consumption information.

[0051] S3. Determine the marketing strategy information of the target user based on the user portrait tags of the target user. The marketing strategy information includes marketing product information, marketing channel information, and speech rhetoric method information.

[0052] In some embodiments, before the above S1, the method further includes: obtaining a basic data set, which includes multiple user portrait tags and multiple marketing strategy information. Among them, each marketing strategy information includes one or more different items of marketing product information, marketing channel information, and speech rhetoric method information; establishing a first database based on the basic data set, and the first database includes the user portrait tags of multiple users and the marketing strategy information corresponding to the user portrait tags of each user.

[0053] Further, the above S3 specifically includes the following steps:

[0054] Obtain the marketing strategy information corresponding to the user portrait tags of the target user from the first database according to the user portrait tags of the target user.

[0055] As can be seen from the above, the method provided by the embodiments of the present invention can determine the marketing strategy information corresponding to the target user based on the user portrait tags. Among them, the marketing strategy information includes marketing product information, marketing channel information, and speech rhetoric method information, which can make the obtained information more suitable for the target user, improve the efficiency and personalization of marketing speech generation, and meet the usage requirements of users in different usage scenarios.

[0056] S4. Generate a speech prompt text according to the user portrait tags and marketing strategy information of the target user.

[0057] In some embodiments, see Figure 4 , the above S4 specifically includes:

[0058] S41. Establish a prompt text template, which includes multiple placeholder information and auxiliary information.

[0059] Among them, each placeholder information is respectively configured with a different position identifier, and the position identifier is used to represent the user portrait tag, marketing product information, marketing channel information, or speech rhetoric method information; the auxiliary information is used to assist the natural language generation model to identify the intention of the prompt text.

[0060] The following illustrates the process of establishing a prompt text template through an example. First, the generation system 100 deeply analyzes the marketing objectives and customer groups to clarify the requirements and constraints for generating the sales pitch. Based on the results of the requirements analysis, the generation system 100 designs a set of prompt templates that include necessary placeholders (placeholder information). The template should cover multiple dimensions such as customer characteristics, product information, channel information, rhetorical devices, etc., to ensure the personalization and diversity of the generated sales pitch. The generation system 100 determines the meaning and value range of each placeholder (corresponding to different position identifiers) to ensure the accuracy and consistency of the subsequent filled data. The definition of the placeholder should be based on the actual business requirements and data characteristics to avoid ambiguous or repetitive definitions. Finally, the generation system 100 validates and optimizes the template through actual data filling and sales pitch generation tests. During the test process, user feedback and marketing effect data can be collected to continuously improve and perfect the template.

[0061] Exemplarily, the following is a prompt text template provided by an embodiment of the present invention, including:

[0062] prompt_template = """

[0063] Recommend our {product_info} to customers with {customer_features} on the {channel_info} channel for {activity_info}. Please write a marketing sales pitch using the rhetorical device of {rhetorical_device}:

[0064] Among them, {customer_features} is the placeholder information with the position identifier of the user portrait label, {product_info} is the placeholder information with the position identifier of the marketing product information, {channel_info} is the placeholder information with the position identifier of the marketing channel information, {activity_info} is the placeholder information with the position identifier of the marketing product information, {rhetorical_device} is the placeholder information with the position identifier of the rhetorical device, and the other text is auxiliary information.

[0065] S42. Replace the placeholder information included in the prompt text template according to the user portrait label, marketing product information, marketing channel information, or rhetorical device information of the target user to obtain a sales pitch prompt text.

[0066] Optionally, before the above S1, the method further includes: obtaining multiple user portrait labels and marketing sales pitch examples corresponding to each user portrait label; establishing a second database according to the marketing sales pitch examples corresponding to each user portrait label;

[0067] Combined with the above embodiments, seeFigure 5 , the above S42 specifically includes:

[0068] S421. Determine marketing speech examples from the second database according to the user portrait tag of the target user.

[0069] S422. Replace the placeholder information included in the prompt word text template according to the user portrait label, marketing product information, marketing channel information or rhetoric information of the target user, and add a marketing speech example to the prompt word text template to obtain a speech prompt text.

[0070] Exemplarily, the following is another prompt word text template provided by an embodiment of the present invention, which includes:

[0071] prompt_template="""

[0072] Take {xxx} as an example, recommend our {product_info} to the customers of {customer_features}, and conduct {activity_info} on the {channel_info} channel. Please use the rhetorical device of {rhetorical_device} to write marketing words:

[0073] From the above, it can be seen that the method provided by the embodiment of the present invention establishes a second database in advance, and adds marketing speech examples corresponding to the user portrait tags of the target users to the speech prompt text, so that the natural language generation model can refer to the marketing speech examples to output the corresponding marketing speech text, which can improve the efficiency and personalization of marketing speech generation and meet the user's usage needs in different usage scenarios.

[0074] S5. Input the sales talk prompt text into the natural language generation model and output the marketing sales talk text for the target users.

[0075] In one possible implementation, the natural language generation model may be a GPT series model or a BERT series model. It should be noted that it may also be a model of other series, and the embodiment of the present invention does not impose any particular limitation on this.

[0076] As can be seen from the above S1 - S5, the method provided by the embodiments of the present invention generates a speech prompt text based on the user portrait tags and marketing strategy information of the target user. Since the marketing strategy information includes marketing product information, marketing channel information, or speech rhetoric method information, therefore, it is possible to generate a marketing speech text corresponding to the target user from multiple dimensions, which can improve the generation efficiency and personalization degree of the marketing speech, and meet the usage requirements of users in different usage scenarios. It can also be understood that: the method provided by the present invention can improve the generation efficiency and personalization degree of the marketing speech. Through intelligent prompt word template construction and large model invocation, high - quality speech that meets marketing requirements can be quickly generated. At the same time, since this method supports real - time update and optimization mechanisms, it can adapt to rapidly changing marketing requirements and maintain the continuous effectiveness and competitiveness of the system.

[0077] By using the method proposed by the present invention, the generation efficiency and personalization degree of the marketing speech can be significantly improved. Through intelligent prompt word template construction and large model invocation, high - quality speech that meets marketing requirements can be quickly generated. At the same time, since this method supports real - time update and optimization mechanisms, it can adapt to rapidly changing marketing requirements and maintain the continuous effectiveness and competitiveness of the system. Specifically, it is reflected in the following three aspects: 1. Significantly improve the generation efficiency and quality of the marketing speech, and reduce the cost of manually writing the speech; 2. Enhance the pertinence and effectiveness of the speech by precisely customizing and optimizing the speech, and improve the marketing conversion rate; 3. Provide a flexible and extensible speech generation framework to meet the needs of different marketing scenarios and provide strong support for front - line marketers.

[0078] In a possible implementation manner, after the above S5, the method provided by the embodiments of the present invention further includes: performing grammar checking and semantic correction on the marketing speech text of the target user.

[0079] Specifically, grammar checking: Perform grammar checking on the generated marketing speech text to ensure the accuracy and fluency of the speech. Grammar checking can include aspects such as spelling checking, punctuation checking, and sentence structure checking. Semantic correction: Correct and optimize the semantics of the marketing speech text to ensure that the speech is clearly expressed, accurate, and meets the marketing objectives. Semantic correction can include aspects such as removing redundant information, adjusting sentence structure, and optimizing language expression.

[0080] The method provided by the embodiments of the present invention further includes: performing personalized customization on the marketing speech text of the target user.

[0081] Specifically, the personalized customization is: According to the marketing objectives and customer needs, perform personalized customization on the marketing speech text. The personalized customization can include aspects such as adjusting the tone and style of the marketing speech text and adding specific elements to better meet customer needs and marketing objectives.

[0082] As can be seen from the above, the method provided by the present invention can effectively improve the effectiveness and fluency of the sales talk by performing grammar checking, semantic correction and personalization on the sales talk text, and meet the usage requirements of users in different usage scenarios.

[0083] In a possible implementation manner, the method provided by the embodiments of the present invention further includes: obtaining a plurality of sales talk texts and the evaluation scores corresponding to each sales talk text, where the evaluation scores are used to characterize the semantic relevance, fluency and conversion rate of the sales talk text; and iteratively training a natural language generation model based on the plurality of sales talk texts and the evaluation scores corresponding to each sales talk text to obtain a trained natural language generation model.

[0084] Specifically, the generation system 100 first establishes an evaluation index system including multiple dimensions such as semantic relevance, fluency and conversion rate. Then, the plurality of sales talk texts are scored and sorted according to the evaluation indexes to obtain the evaluation scores corresponding to each sales talk text. The evaluation scores are used to characterize user feedback and marketing effects, and the natural language generation model is iteratively trained based on the plurality of sales talk texts and the evaluation scores corresponding to each sales talk text to improve the pertinence and effectiveness of the sales talk.

[0085] It should be noted that the natural language generation model is trained and optimized to improve the quality and efficiency of the generated sales talk. During the training process, transfer learning and incremental learning methods can be adopted to accelerate the model training speed and improve the model performance.

[0086] In a possible implementation manner, the generation system 100 provided by the embodiments of the present invention exchanges data and collaborates with other systems through an API interface.

[0087] The present invention integrates the generation system 100 into the intelligent marketing system through the API interface, realizing the automatic generation and intelligent recommendation of sales talk. At the same time, a simple and intuitive user interface and rich API interfaces are provided, facilitating marketers to quickly obtain the required sales talk and perform personalization. In addition, the system security protection measures are strengthened to ensure the security and privacy of data. This system integration solution enables the method of the present invention to be easily applied to actual marketing scenarios, providing more accurate and effective support for marketing activities.

[0088] The above mainly introduced the solution of the embodiment of the present invention from the perspective of methods. It can be understood that in order to implement the above functions, the generation system 100 includes at least one of the corresponding hardware structures and software modules for executing each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments disclosed herein, the embodiments of the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving the hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the embodiments of the present invention.

[0089] The embodiments of the present invention can divide the generation system 100 into functional units according to the above method examples. For example, each function of the generation system 100 can be divided into corresponding functional units, or two or more functions can be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. It should be noted that the division of units in the embodiments of the present invention is illustrative, only a logical function division, and there may be other division methods in actual implementation.

[0090] Exemplarily, Figure 6 FIG. shows a schematic hardware structure diagram of a generation system provided by an embodiment of the present invention. The generation system 100 includes: a data acquisition unit 110, configured to acquire historical data of a target user, where the historical data includes identity information, interest and hobby information, and historical consumption information; a label determination unit 120, configured to determine a user portrait label of the target user according to the historical data; a strategy determination unit 130, configured to determine marketing strategy information of the target user based on the user portrait label of the target user, where the marketing strategy information includes marketing product information, marketing channel information, and rhetorical device information; a prompt text generation unit 140, configured to generate a rhetorical prompt text according to the user portrait label and marketing strategy information of the target user; and a marketing rhetoric generation unit 150, configured to input the rhetorical prompt text into a natural language generation model and output a marketing rhetoric text of the target user.

[0091] Optionally, the label determination unit 120 is specifically configured to: preprocess the historical data to obtain feature information corresponding to the target user, where the preprocessing includes data cleaning, normalization, and feature extraction; and obtain the user portrait label of the target user according to the feature information corresponding to the target user.

[0092] Optionally, the data acquisition unit 110 is further configured to acquire a basic data set, which includes multiple user portrait tags and multiple marketing strategy information. Each marketing strategy information includes one or more different items among marketing product information, marketing channel information, and speech rhetoric method information. The strategy determination unit 130 is further configured to establish a first database based on the basic data set. The first database includes user portrait tags of multiple users and marketing strategy information corresponding to the user portrait tags of each user.

[0093] Optionally, the strategy determination unit 130 is specifically configured to obtain, from the first database, the marketing strategy information corresponding to the user portrait tag of the target user according to the user portrait tag of the target user.

[0094] Optionally, the prompt text generation unit 140 is specifically configured to: establish a prompt text template, which includes multiple placeholder information and auxiliary information. Each placeholder information is respectively configured with a different position identifier, and the position identifier is used to represent user portrait tags, marketing product information, marketing channel information, or speech rhetoric method information. The auxiliary information is used to assist the natural language generation model in identifying the intention of the prompt text. Replace the placeholder information included in the prompt text template according to the user portrait tag, marketing product information, marketing channel information, or speech rhetoric method information of the target user to obtain a speech prompt text.

[0095] Optionally, the data acquisition unit 110 is further configured to acquire multiple user portrait tags and marketing speech examples corresponding to each user portrait tag. The prompt text generation unit 140 is further configured to establish a second database according to the marketing speech examples corresponding to each user portrait tag. The prompt text generation unit 140 is further configured to: determine a marketing speech example from the second database according to the user portrait tag of the target user; replace the placeholder information included in the prompt word text template according to the user portrait tag, marketing product information, marketing channel information, or speech rhetoric method information of the target user, and add a marketing speech example to the prompt word text template to obtain a speech prompt text.

[0096] Optionally, the above system further includes a model training unit 160, which is configured to: acquire multiple marketing speech texts and evaluation scores corresponding to each marketing speech text. The evaluation score is used to represent the semantic relevance, fluency, and conversion rate of the marketing speech text. Iteratively train the natural language generation model based on the multiple marketing speech texts and the evaluation scores corresponding to each marketing speech text to obtain a trained natural language generation model.

[0097] Optionally, the above system further includes a text optimization unit 170, which is configured to: perform grammar checking and semantic correction on the marketing speech text of the target user.

[0098] It should be understood that for the specific descriptions of the above optional manners, reference may be made to the foregoing method embodiments, which will not be elaborated herein. In addition, for the explanations and beneficial effects descriptions of any of the above-provided generation systems 100, reference may be made to the corresponding method embodiments above, which will not be elaborated herein.

[0099] An embodiment of the present invention further provides a computer-readable storage medium, in which at least one computer instruction is stored, and the at least one computer instruction is loaded and executed by a processor to implement the methods of the above various embodiments. For the explanations and beneficial effects descriptions of the relevant content in any of the above-provided computer-readable storage media, reference may be made to the corresponding embodiments above, which will not be elaborated herein.

[0100] An embodiment of the present invention further provides a chip. A control circuit for implementing the functions of the above generation system 100 and one or more ports are integrated in the chip. Optionally, the functions supported by the chip may refer to the above, which will not be elaborated herein.

[0101] Those of ordinary skill in the art can understand that all or part of the steps for implementing the above embodiments can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a random access memory, etc. The above processing unit or processor can be a central processing unit, a general-purpose processor, an application specific integrated circuit (ASIC), a digital signal processor (DSP), a field programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.

[0102] Embodiments of the present invention also provide a computer program product containing instructions. When the instructions run on a computer, the computer is caused to execute any one of the methods in the above embodiments. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of the present invention are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from a website, a computer, a server, or a data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that the computer can access or a data storage device such as a server or a data center integrating one or more available media. The available medium may be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as an SSD), etc.

[0103] It should be noted that the devices for storing computer instructions or computer programs provided in the embodiments of the present invention, such as but not limited to, the above-mentioned memory, computer-readable storage medium, and communication chip, etc., are all non-transitory. Those skilled in the art should be able to realize that in the above one or more examples, the functions described in the embodiments of the present invention can be implemented by hardware, software, firmware, or any combination thereof. When implemented using software, these functions may be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. The computer-readable storage medium includes computer storage media and communication media, where the communication media includes any medium facilitating the transmission of a computer program from one place to another. The storage medium may be any available medium that a general-purpose or special-purpose computer can access.

[0104] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A marketing copywriting generation method based on large model prompt engineering, characterized in that The method includes: Obtaining historical data of a target user, where the historical data includes identity information, hobby information, and historical consumption information; Determining user portrait tags of the target user according to the historical data; Determining marketing strategy information of the target user based on the user portrait tags of the target user, where the marketing strategy information includes marketing product information, marketing channel information, and rhetorical device information; Generating a speech prompt text according to the user portrait tags and marketing strategy information of the target user; Inputting the speech prompt text into a natural language generation model to output a marketing speech text of the target user.

2. The method according to claim 1, wherein The determining the user portrait tags of the target user according to the historical data includes: Performing preprocessing on the historical data to obtain feature information corresponding to the target user, where the preprocessing includes data cleaning, normalization, and feature extraction; Obtaining the user portrait tags of the target user according to the feature information corresponding to the target user.

3. The method according to claim 2, characterized in that, Before obtaining the historical data of the target user, the method further includes: Obtaining a basic data set, where the basic data set includes multiple user portrait tags and multiple marketing strategy information, and among them, each marketing strategy information includes one or more different items of marketing product information, marketing channel information, and rhetorical device information; Establishing a first database based on the basic data set, where the first database includes user portrait tags of multiple users and marketing strategy information corresponding to the user portrait tags of each user.

4. The method according to claim 3, characterized in that, The determining the marketing strategy information of the target user based on the user portrait tags of the target user includes: Obtaining the marketing strategy information corresponding to the user portrait tags of the target user from the first database according to the user portrait tags of the target user.

5. The method according to claim 4, wherein The generating a speech prompt text according to the user portrait tags and marketing strategy information of the target user includes: Establishing a prompt text template, where the prompt text template includes multiple placeholder information and auxiliary information, and among them, each placeholder information is respectively configured with a different position identifier, and the position identifier is used to represent a user portrait tag, marketing product information, marketing channel information, or rhetorical device information; the auxiliary information is used to assist the natural language generation model to recognize the intention of the prompt text; Replacing the placeholder information included in the prompt text template according to the user portrait tags, marketing product information, marketing channel information, or rhetorical device information of the target user to obtain a speech prompt text.

6. The method according to claim 5, wherein Before obtaining the historical data of the target user, the method further includes: Obtaining multiple user portrait tags and marketing speech examples corresponding to each user portrait tag; Establishing a second database according to the marketing speech examples corresponding to each user portrait tag; The replacing the placeholder information included in the prompt text template according to the user portrait tags, marketing product information, marketing channel information, or rhetorical device information of the target user to obtain a speech prompt text includes: Determining a marketing speech example from the second database according to the user portrait tags of the target user; The placeholder information included in the prompt word text template is replaced according to the user portrait label, marketing product information, marketing channel information or speech rhetoric information of the target user, and the marketing speech example is added to the prompt word text template to obtain the speech prompt text.

7. The method according to claim 6, wherein The method further comprises: Acquire multiple marketing speech texts and an evaluation score corresponding to each marketing speech text, wherein the evaluation score is used to characterize the semantic relevance, fluency and transaction rate of the marketing speech text; The natural language generation model is iteratively trained based on multiple marketing speech texts and the evaluation score corresponding to each marketing speech text to obtain a trained natural language generation model.

8. The method according to claim 7, wherein After inputting the speech prompt text into a natural language generation model and outputting the marketing speech text of the target user, the method further includes: Perform grammar checking and semantic correction on the marketing text of the target user.

9. A marketing copy generation system based on large model prompt engineering, characterized in that, The system comprises: A data acquisition unit, used to acquire historical data of a target user, wherein the historical data includes identity information, hobby information, and historical consumption information; A label determination unit, used to determine a user portrait label of the target user according to the historical data; A strategy determination unit, configured to determine marketing strategy information of the target user based on the user portrait tag of the target user, wherein the marketing strategy information includes marketing product information, marketing channel information, and rhetoric information; A prompt text generating unit, used to generate a prompt text according to the user portrait label and marketing strategy information of the target user; The marketing speech generation unit is used to input the speech prompt text into the natural language generation model and output the marketing speech text of the target user.

10. An electronic device, characterized in that, include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the marketing speech generation method based on large model prompt word engineering as described in any one of claims 1-8.

Images