Medical instrument digital supply chain AI intelligent data analysis method and device

Abstract

The embodiment of the invention provides a medical instrument digital supply chain AI intelligent data analysis method, which comprises the following steps: receiving a user model configuration request, analyzing to obtain configuration information, and completing configuration link between a large language model and a text-to-voice model; the method comprises the following steps: receiving a cue word template configuration request, analyzing to obtain template configuration information including template codes, template names, system settings and template contents, and constructing a cue word template; when a service analysis request is received, analyzing to obtain a template code and service data, matching a cue word template, replacing template content with the service data, generating cue words, inputting the cue words and system human settings into a large language model, obtaining a text analysis result, and immediately sending the text analysis result to a web platform dialog box for display; and inputting a text analysis result into a text-to-voice model to obtain a synchronous voice stream, and immediately sending the synchronous voice stream to a web platform for broadcasting. Intelligent data analysis is achieved through combination of the custom cue word and the AI model, and the method is efficient, flexible and easy to operate.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of data processing technology and related technical fields, and specifically, to a method and apparatus suitable for AI intelligent data analysis of a digital supply chain of medical devices. Background Art

[0002] With the continuous advancement of medical technology, the digital supply chain management system of medical devices is becoming increasingly important, but most current systems rely on manual statistics and relatively fixed data analysis methods. However, these methods have obvious limitations when processing large and complex data sets and cannot flexibly adapt to diverse analysis needs.

[0003] The main disadvantages of the prior art are mainly the following aspects:

[0004] The data analysis process is lengthy and error-prone due to reliance on a large amount of manual work, resulting in overall inefficiency;

[0005] Existing tools fail to deeply integrate artificial intelligence technology, fail to keep up with the latest industry trends, and fail to generate timely and effective analysis reports;

[0006] Supply chain managers need to have a strong background in statistics or data analysis, which limits the ability of ordinary users to quickly get started and apply the technology.

[0007] Existing analysis tools have limited functions and can only provide basic statistical data. They are unable to deeply explore the value of data and achieve comprehensive analysis and prediction.

[0008] Therefore, there is an urgent need for a method of AI intelligent data analysis in the digital supply chain of medical devices to solve the above problems. Summary of the invention

[0009] The embodiments described in this article provide a method and device for AI intelligent data analysis of a medical device digital supply chain to solve the problems existing in the prior art.

[0010] According to a first aspect of the present disclosure, a method for analyzing digital data of a digital supply chain of medical devices is provided, comprising:

[0011] Receive a model configuration request from a user, parse and obtain model configuration information, and complete configuration and linking with a selected large language model and text-to-speech model according to the model configuration information;

[0012] Receive a prompt word template configuration request from a user, parse and obtain template configuration information, construct a prompt word template based on the template configuration information, and store it in a database, wherein the template configuration information includes: template code, template name, system person setting and template content;

[0013] When receiving a request for initiating business analysis, the template code and business data are parsed and obtained, the template code is searched in the database, the corresponding prompt word template is matched, the template content and system persona are extracted from the prompt word template, and the template content is replaced with the business data to generate prompt words, the prompt words and the system persona are input into the selected large language model, the returned text analysis results are obtained, and the results are immediately sent to the dialog box of the web platform for display;

[0014] If the voice broadcast function has been enabled, the text analysis result is input into the selected text-to-speech model, the returned synchronous voice stream is obtained, and it is immediately sent to the web platform for broadcasting.

[0015] In some embodiments of the present disclosure, before the step of receiving the user's model configuration request and parsing and obtaining the model configuration option information, the following step is further included:

[0016] Configure necessary information, including access keys and security keys.

[0017] In some embodiments of the present disclosure, the template content includes: analyzed data, return format, requirements, and analysis results;

[0018] The analysis results include: risk assessment, advantage analysis, lue analysis, recommendations and data implications;

[0019] Replacing the template content with the business data, specifically, replacing the analyzed data in the template content with the business data.

[0020] In some embodiments of the present disclosure, the template content is replaced with the business data using a variable replacement method.

[0021] In some embodiments of the present disclosure, the selected text-to-speech model performs sentence segmentation processing on the text in the text analysis result based on symbols or character lengths as the sentence segmentation basis of the text;

[0022] Processing the text length of the text in the text analysis result. If the length of the text after the segmentation is too short, the segmented text is merged with the subsequent segmented text for broadcasting. If the length of the text after the segmentation is too long, it is divided into multiple shorter text segments and broadcasted multiple times.

[0023] The text analysis results are managed into a queue to be broadcasted, new sentence-segmented text that can be broadcasted is pushed into the queue to be broadcasted, the queue to be broadcasted is monitored, and the voice stream is output in batches in sequence, and the content in the voice stream is broadcasted according to the voice stream;

[0024] When a symbol is detected or a specific character length is reached, the content of that paragraph will be broadcast first. At the same time, the result return process will not be interrupted. After the content of the current paragraph is broadcast, the remaining content will continue to be broadcast according to the same rules until all the content is completed.

[0025] In some embodiments of the present disclosure, through multiple rounds of conversations, the large language model combines industry best practices to perform in-depth analysis of the business data related to the medical device and derive the text analysis results.

[0026] In some embodiments of the present disclosure, after the step of exporting the text analysis result, the following steps are included:

[0027] The text analysis results of the conversation are exported as an analysis report, which includes: data meaning, risk assessment, analysis results of strengths and weaknesses, and improvement suggestions for the analysis results.

[0028] According to a second aspect of the present disclosure, a device for AI intelligent data analysis of a digital supply chain of medical devices is provided, comprising:

[0029] A model configuration module, used to receive a model configuration request from a user, parse and obtain model configuration information, and complete the configuration and linking with the selected large language model and text-to-speech model according to the model configuration information;

[0030] The template configuration module is used to receive a prompt word template configuration request from a user, parse and obtain template configuration information, construct a prompt word template based on the template configuration information, and store it in a database. The template configuration information includes: template code, template name, system person settings and template content;

[0031] The first processing module is used to parse and obtain the template code and business data when receiving a business analysis initiation request, query the template code in the database, match the corresponding prompt word template, extract the template content and system persona from the prompt word template, replace the template content with the business data, generate prompt words, input the prompt words and the system persona into the selected large language model, obtain the returned text analysis results, and send them to the dialog box of the web platform for display in real time;

[0032] The second processing module is used to input the text analysis result into the selected text-to-speech model if the voice broadcast function has been enabled, obtain the returned synchronous voice stream, and send it to the web platform for broadcast in real time.

[0033] According to a third aspect of the present disclosure, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the method in any one of the above embodiments are implemented.

[0034] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in any one of the above embodiments are implemented.

[0035] The method and device for AI intelligent data analysis of the digital supply chain of medical devices provided by the embodiments of the present disclosure receive user model configuration requests, parse and obtain configuration information, complete the configuration link between the large language model and the text-to-speech model; receive prompt word template configuration requests, parse and obtain template configuration information, including template code, template name, system person setting and template content, build and store prompt word templates; when receiving business analysis requests, parse and obtain template code and business data, match prompt word templates, replace template content with business data, generate prompt words, input prompt words and system person setting into the large language model, obtain text analysis results, and send them to the web platform dialog box for display immediately. If the voice broadcast function is enabled, input the text analysis results into the text-to-speech model, obtain the synchronous voice stream, and send it to the web platform for broadcast immediately. The automation of data format definition and transmission is realized, wherein the automation of transmission includes: reading template information, receiving business data to be analyzed, determining analysis result type information, and forwarding analysis results; the method provided by the embodiment of the present disclosure supports free switching of LLM model and TTS technology to improve the stability and risk resistance of the system; the method provided by the embodiment of the present disclosure supports real-time processing of text segments and synchronous voice broadcast to ensure that TTS voice broadcast and return text are as synchronized as possible. It can dynamically access AI generative models to realize automated data analysis and report generation. Users can set prompt word templates for different business modules according to needs, and the platform automatically generates corresponding analysis reports. The preset data format ensures that data transmission meets the standards and is convenient for AI model processing. Combined with industry best practices, the AI ​​model analyzes data in multiple dimensions and generates a comprehensive report including data meaning, risk assessment, etc. The analysis results help managers quickly understand the operating status, put forward improvement suggestions, and optimize management processes. Automated and intelligent data analysis enables managers to adjust operating strategies more timely and effectively and reduce costs. The AI ​​model is combined with industry best practices to greatly improve the accuracy and efficiency of analysis. Without complex statistical knowledge, supply chain managers can directly obtain optimization suggestions. Supports adjustment of analysis scope and depth according to actual needs, and flexibly responds to market changes. Reduces manual operations and potential errors, and improves decision-making capabilities. Provides text-synchronized voice broadcast to enhance information communication. Supports free switching of models, and allocates model advantages according to different business needs.

[0036] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described below. It should be noted that the drawings described below only relate to some embodiments of the present disclosure, but are not intended to limit the present disclosure, wherein:

[0038] Figure 1 It is a flowchart of a method for AI intelligent data analysis of a medical device digital supply chain provided by an embodiment of the present disclosure;

[0039] Figure 2 It is a structural schematic diagram of a device for AI intelligent data analysis of a medical device digital supply chain provided by an embodiment of the present disclosure;

[0040] Figure 3 is a schematic diagram of a prompt word template of a specific example provided by an embodiment of the present disclosure;

[0041] Figure 4 is a schematic diagram of a text analysis result of a specific example provided by an embodiment of the present disclosure;

[0042] Figure 5 It is a structural diagram of a computer device provided in an embodiment of the present disclosure.

[0043] In the drawings, reference numerals with the same last two digits correspond to the same elements. It should be noted that the elements in the drawings are schematic and not drawn to scale. DETAILED DESCRIPTION

[0044] Before introducing this application, the terms involved in this article are first explained:

[0045] Prompt words: User-defined or automatically generated instructions used to input specific requirements to the AI ​​model. When communicating with AI (such as a large language model), the instructions that need to be sent to the AI ​​can be a text description, such as "Please recommend me a best-selling book" when you are talking to the AI, or a parameter description in a certain format, such as asking the AI ​​to draw in a certain format, which requires describing the relevant drawing parameters.

[0046] AI generative model: refers to a model that generates natural language or structured data through artificial intelligence technology and can generate response content based on input prompt words.

[0047] LLM (Large language model) refers to a deep neural network model based on the Transformer structure that is trained using a large amount of text data. It can generate natural language text or understand the meaning of language text, and can handle a variety of natural language tasks, such as text classification, question answering, and dialogue. Common large language models include ChatGPT large language model, Wenxin Yiyan large language model, Spark Cognitive large language model, and Pangu large language model.

[0048] TTS (Text-to-Speech) is a technology that can convert text into speech. This technology is widely used in various devices and applications, such as smartphones, smart speakers, and car navigation systems, to provide users with voice output functions. The TTS model analyzes the input text and converts it into natural and fluent speech output, making it easier for people with visual impairments, the elderly, or foreign language learners to obtain information. Common TTS models include iFlytek TTS model, Azure AI TTS model, ChatTTS model, Seed-TTS model, EmotiVoiceTTS model, OpenVoiceTTS model, etc.

[0049] SSE (Server-Sent Events) is a technology that allows the server to actively push update information to the client, and is suitable for scenarios that require real-time interaction.

[0050] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0051] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative work also fall within the scope of protection of the present disclosure.

[0052] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase "embodiments" in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0053] The term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists, A and B exist at the same time, and B exists. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0054] Furthermore, in all embodiments of the present disclosure, terms such as “first” and “second” are used only to distinguish one component (or a part of a component) from another component (or another part of a component).

[0055] In the description of the present application, unless otherwise specified, "plurality" means more than two (including two), and similarly, "multiple groups" means more than two groups (including two).

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0057] The main shortcomings of existing technologies are as follows: due to reliance on a large amount of manual operations, the medical device supply chain data analysis process is lengthy and prone to errors, resulting in overall inefficiency; existing tools fail to deeply integrate artificial intelligence technology, cannot keep up with the latest industry trends, and cannot generate timely and effective analysis reports; medical device supply chain managers need to have a deep background in statistics or data analysis, which limits the rapid start-up and application of ordinary users; existing analysis tools have limited functions and can only provide basic statistical data, and cannot deeply explore the value of data and achieve comprehensive analysis and prediction.

[0058] The method for AI intelligent data analysis of the digital supply chain of medical devices provided in the embodiments of the present disclosure is applied to a server.

[0059] Based on the problems existing in the existing technology, Figure 1 is a flow chart of a method for AI intelligent data analysis of a medical device digital supply chain provided by an embodiment of the present disclosure, such as Figure 1 As shown in the figure, the specific process of the AI ​​intelligent data analysis method for the digital supply chain of medical devices includes:

[0060] S110, receiving a model configuration request from a user, parsing and acquiring model configuration information, and completing configuration and linking with a selected large language model and text-to-speech model according to the model configuration information.

[0061] In a specific implementation, the configuration information of the LLM model is parsed from the user's model configuration requirements, the identification information of the LLM model is extracted therefrom, the selected LLM model is enabled according to the identification information of the LLM model, and the configuration and linking with the selected LLM model are completed.

[0062] In a specific implementation, the configuration information of the TTS model is parsed from the model configuration requirements of the receiving user, the identification information of the TTS model is extracted therefrom, the selected TTS model is enabled according to the identification information of the TTS model, and the configuration and linking with the selected TTS model are completed.

[0063] Users can switch models freely and assign different services based on the advantages of different models.

[0064] In addition, before implementing the above steps, it also includes:

[0065] S100. Configure necessary information according to key configuration information, where the necessary information includes: an access key (ak) and a security key (sk).

[0066] In a specific implementation method, first, the web platform receives key configuration information provided by the user or the management end, which contains the credentials necessary to access and use the LLM and TTS models. The received key configuration information contains two core parts: the access key (ak) and the security key (sk). The access key (ak) is the credential for model access rights, which ensures that only authorized users or systems can access and use these models. The security key (sk) provides additional security protection, which is used to verify the authenticity of the request and ensure the security and integrity during the transmission and processing of data. The web platform will perform necessary configurations on the LLM model and TTS model based on this information. This includes but is not limited to setting model access rights, verifying the validity of requests, and ensuring the security of data transmission. By correctly configuring these keys, the system can ensure that the LLM and TTS models run efficiently in a secure environment while protecting user data from unauthorized access and leakage.

[0067] S120, receiving a prompt word template configuration request from a user, parsing and obtaining template configuration information, constructing a prompt word template based on the template configuration information, and storing it in a database for subsequent calling and use, wherein the template configuration information includes: template code, template name, system person setting, and template content.

[0068] In a specific implementation, the first conversation content is added to the template content to provide the user with a clear conversation guide.

[0069] The first preset content covers several key elements, including the data to be analyzed, the return format, requirements, and analysis results. The return format setting ensures that the data transmitted to the LLM model is both compliant with the standards and easy to process. The analysis result types include risk assessment, advantage analysis, lue-style analysis, suggestions, and interpretation of data meaning, which provide users with comprehensive and in-depth data insights.

[0070] Based on the template configuration information such as the template code, the template name, the system person setting and the template content, a prompt word template is constructed, and the prompt word template is used to define the data format.

[0071] Each business module corresponds to a prompt word template, and the prompt word template provides a standardized dialogue framework for different business scenarios.

[0072] Through the above configuration process, an efficient and accurate conversation experience is provided to users.

[0073] S130: When receiving a request to initiate business analysis, parse and obtain the template code and business data, query the template code in the database, match the corresponding prompt word template, extract the template content and system persona from the prompt word template, replace the template content with the business data, generate prompt words, input the prompt words and the system persona into the selected large language model, obtain the returned text analysis results, and immediately send them to the dialog box of the web platform for display.

[0074] In a specific implementation, the business data obtained is business data related to medical devices, such as basic product information, sales data, usage records, maintenance records, patient feedback, and reports from regulatory agencies, as well as data containing detailed information on medical devices, such as model, specification, manufacturer, production date, expiration date, use department, operator, etc., and its specific performance in clinical use, such as frequency of use, failure rate, maintenance records, etc. It may also be sales data, competitive situation, price trends and other information on the external medical device market.

[0075] In a specific implementation, when a business analysis initiation request is received, a prompt word template that matches it is intelligently selected according to the analysis type or business scenario specified in the request. The previously acquired business data related to medical devices are accurately replaced one by one with the analyzed data in the template content in the prompt word template. These business data may cover multiple dimensions such as sales statistics, usage records, patient feedback, and maintenance records of medical devices, which together constitute the basic materials for dialogue analysis. In the process of replacing data, the template content is replaced with the business data using the variable replacement method. Thereby, a complete, clear, and accurate prompt word is generated. The prompt word defines the data format to ensure that the data transmitted to the LLM model meets the standard and is convenient for model processing. Through it, the web platform can interact with users in a more natural and intuitive way, guiding users to gradually deepen the data analysis process, and finally reveal the valuable information hidden behind the data.

[0076] Users can set different prompt words for different business modules according to specific needs.

[0077] In addition, since the front-end web platform has established a stable link with the back-end LLM model through the server sent endpoint (SSE) technology, the generated prompt words are sent to the currently enabled LLM model through this channel.

[0078] The enabled LLM model will immediately start its powerful analytical capabilities to deeply process and interpret the prompt words. Combined with the best practices of the industry, the data will be analyzed in multiple dimensions to generate an analytical report containing data meaning, risk assessment, strengths and weaknesses;

[0079] The analysis results returned by the LLM model are received by the web platform in a streaming manner and are displayed word by word in a dialog box.

[0080] After the currently enabled LLM model receives the prompt word and completes the analysis, the LLM model will stream the text analysis results back to the front-end web platform.

[0081] After the front-end web platform receives the streaming data sent by the LLM model, it will be displayed in real time in the dialog box on the display screen, allowing users to clearly see each step of the analysis process and the final result. This method not only improves the real-time nature of data analysis, but also enhances the fluency and interactivity of the user experience.

[0082] S140: If the voice broadcast function has been enabled, the text analysis result is input into the selected text-to-speech model, the returned synchronous voice stream is obtained, and it is immediately sent to the web platform for broadcast.

[0083] In a specific implementation, a voice broadcast switch button is provided on the web platform interface, and the user controls the on and off state of the voice broadcast function by clicking the button;

[0084] Specifically, it is necessary to determine whether the voice broadcast function has been set to an enabled state;

[0085] If the voice broadcast function is set to be turned on, that is, when the user clicks the analysis button of the relevant template, the web platform inputs the text analysis results into the selected text-to-speech model;

[0086] The selected text-to-speech model performs sentence segmentation processing on the text of the text analysis result based on the symbol or character length as the sentence segmentation basis of the text;

[0087] Processing the text analysis result according to the text length. If the length of the text after the segmentation is too short, the segmented text is combined with the subsequent segmented text for broadcasting. If the length of the text after the segmentation is too long, it is divided into multiple shorter text segments and broadcasted multiple times.

[0088] The text analysis results are managed into a queue to be broadcasted, new sentence-segmented text that can be broadcasted is pushed into the queue to be broadcasted, the queue to be broadcasted is monitored, and the voice stream is output in batches in sequence, and the content in the voice stream is broadcasted according to the voice stream;

[0089] When a symbol is detected or a specific character length is reached, the content of that paragraph will be broadcast first. At the same time, the result return process will not be interrupted. After the content of the current paragraph is broadcast, the remaining content will continue to be broadcast according to the same rules until all the content is completed.

[0090] The synchronous voice stream returned by the SST model is obtained and sent to the web platform for broadcasting in real time to ensure that the TTS voice broadcast and the returned text are synchronized as much as possible.

[0091] During the broadcast process, users can set the voice broadcast switch to control the voice broadcast in real time to meet different usage needs. Synchronous voice broadcast with text gives users a deeper impression.

[0092] During the broadcast process, users can control the broadcast at any time through the voice broadcast switch.

[0093] In addition, after implementing the above steps, it also includes:

[0094] S150. Through multiple rounds of dialogues and in combination with industry best practices, conduct in-depth analysis of the business data related to the medical device and export the analysis result data.

[0095] In a specific implementation method, through multiple rounds of dialogue, combined with market conditions and other conditions, the data is deeply analyzed;

[0096] After the analysis is completed, the analysis results of this conversation are exported as an analysis report, and the analysis includes: data meaning, risks, advantages, disadvantages and suggestions;

[0097] Each business module can perform data analysis;

[0098] The user clicks the analysis button on the interface to initiate a business analysis request, triggering the analysis process. The Web platform automatically imports the preset prompt word template and business data, automatically analyzes the industry data, conducts in-depth analysis through multiple rounds of conversations, and can export the analysis result data.

[0099] By combining AI models with industry best practices, the accuracy and efficiency of analysis can be greatly improved.

[0100] Without the need for complex statistical knowledge, supply chain managers can directly obtain actionable optimization suggestions.

[0101] Support adjustment of analysis scope and depth according to actual needs, and flexibly respond to changing market environments.

[0102] In addition, after implementing the above steps, it also includes:

[0103] S160. Export the analysis results of this conversation into an analysis report, wherein the analysis report includes: data meaning, risk assessment, analysis results of strengths and weaknesses, and improvement suggestions for the analysis results.

[0104] The AI ​​intelligent data analysis method for the digital supply chain of medical devices provided in the embodiment of the present disclosure receives a user model configuration request, parses and obtains configuration information, completes the configuration link between the large language model and the text-to-speech model; receives a prompt word template configuration request, parses and obtains template configuration information, including template code, template name, system person setting and template content, and constructs and stores the prompt word template; when receiving a business analysis request, parses and obtains template code and business data, matches the prompt word template, replaces the template content with business data, generates prompt words, inputs the prompt words and system person setting into the large language model, obtains the text analysis results, and sends them to the web platform dialog box for display immediately. If the voice broadcast function is enabled, the text analysis results are input into the text-to-speech model, the synchronous voice stream is obtained, and it is sent to the web platform for broadcast immediately. The combination of custom prompt words and AI models makes data analysis and broadcasting efficient and flexible, and easy to operate.

[0105] On the basis of the above embodiments, the present disclosure also provides a device for AI intelligent data analysis of a medical device digital supply chain, such as Figure 2 As shown, the device comprises:

[0106] The model configuration module 210 is used to receive a model configuration request from a user, parse and obtain model configuration information, and complete the configuration and linking with the selected large language model and text-to-speech model according to the model configuration information;

[0107] The template configuration module 220 is used to receive a prompt word template configuration request from a user, parse and obtain template configuration information, construct a prompt word template based on the template configuration information, and store it in a database. The template configuration information includes: template code, template name, system person settings and template content;

[0108] The first processing module 230 is used to parse and obtain the template code and business data when receiving a business analysis initiation request, query the template code in the database, match the corresponding prompt word template, extract the template content and system persona from the prompt word template, replace the template content with the business data, generate prompt words, input the prompt words and the system persona into the selected large language model, obtain the returned text analysis results, and send them to the dialog box of the web platform for display in real time;

[0109] The second processing module 240 is used to input the text analysis result into the selected text-to-speech model if the voice broadcast function has been enabled, obtain the returned synchronous voice stream, and send it to the web platform for broadcast in real time.

[0110] In a specific embodiment, it also includes:

[0111] The key configuration module 200 is used to configure necessary information, where the necessary information includes an access key and a security key.

[0112] In a specific implementation, the template content includes: analyzed data, return format, requirements and analysis results;

[0113] The analysis results include: risk assessment, advantage analysis, lue analysis, recommendations and data implications;

[0114] Replacing the template content with the business data, specifically, replacing the analyzed data in the template content with the business data.

[0115] In a specific implementation manner, the template content is replaced with the business data using a variable replacement method.

[0116] In a specific implementation, the selected text-to-speech model performs sentence segmentation processing on the text in the text analysis result based on symbols or character lengths as the sentence segmentation basis of the text;

[0117] Processing the text length of the text in the text analysis result. If the length of the text after the segmentation is too short, the segmented text is merged with the subsequent segmented text for broadcasting. If the length of the text after the segmentation is too long, it is divided into multiple shorter text segments and broadcasted multiple times.

[0118] The text analysis results are managed into a queue to be broadcasted, new sentence-segmented text that can be broadcasted is pushed into the queue to be broadcasted, the queue to be broadcasted is monitored, and the voice stream is output in batches in sequence, and the content in the voice stream is broadcasted according to the voice stream;

[0119] When a symbol is detected or a specific character length is reached, the content of that paragraph will be broadcast first. At the same time, the result return process will not be interrupted. After the content of the current paragraph is broadcast, the remaining content will continue to be broadcast according to the same rules until all the content is completed.

[0120] In a specific embodiment, it also includes:

[0121] The third processing module 250 is used to perform in-depth analysis of the business data related to the medical device through multiple rounds of dialogue, the large language model combined with industry best practices, and export the text analysis results.

[0122] In a specific embodiment, it also includes:

[0123] The export module 260 is used to export the text analysis results of the conversation into an analysis report, which includes: data meaning, risk assessment, analysis results of advantages and disadvantages, and improvement suggestions for the analysis results.

[0124] The device for AI intelligent data analysis of the digital supply chain of medical devices provided in the embodiments of the present disclosure realizes intelligent data analysis by combining custom prompt words with AI models, which is efficient, flexible and easy to operate.

[0125] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0126] The following is a specific example:

[0127] First, the user selects the Wenxinyiyan LLM model and the iFlytek TTS model, and completes the configuration and linking with the Wenxinyiyan LLM model and the iFlytek TTS model respectively to ensure that the system can call these models for subsequent processing.

[0128] For each business module (such as business modules A, B, and C), the user builds a prompt word template containing template code, template name, system persona, and template content. These prompt word templates are written into the database for subsequent calls. For example, for business module A, a prompt word template A is built. Prompt word template A includes template number "cusReturnMoney", template name "Financial Statement-Unreturned Soft Information", and system persona "You are a professional analyst in the supply chain industry of medical devices". The first preset template content is "Below is a set of unreturned payment information of my company. Please analyze what these data can illustrate based on the data. Requirements: 1. No more than 200 words, 2. You need to list the key data first, then analyze and summarize, 3. You don’t need to say improvement strategies, but you should say it when I ask, 4. The first sentence at the beginning of the next reply needs to say: "I am the AI ​​analysis assistant of Yishu Chain", 5. The text needs to be typeset for easy reading. The following are specific values, time period %s, corresponding unreturned amount %s".

[0129] When the user clicks the "Analysis" button of business module A on the web platform interface, a series of automated processing flows will be triggered.

[0130] First, according to the business module A clicked by the user, the prompt word template A corresponding to the template code "cusReturnMoney" is automatically matched. Then, the medical supply chain business data associated with business module A is read. These data may come from a database, API interface or other data source. Then, using the variable replacement method, the "time period %" and "uncollected amount %s" in the template content are replaced with specific information in the medical supply chain business data. The result after processing is the month (existing month on the x-axis): January, February, March, amount (amount corresponding to the y-axis): 1000w, 2000w, 3000w, and a complete prompt word containing actual business data is generated, as follows: The following is a set of uncollected information of my company. Please analyze what these data can explain in combination with the data. Requirements: 1. It cannot exceed 200 words, 2. You need to list the key data first, then analyze and summarize, 3. You don’t need to say the improvement strategy, but you take the initiative to say it when I ask, 4. The first sentence at the beginning of the next reply needs to say: "I am the AI ​​analysis assistant of the medical chain", 5. The text needs to be typeset for easy reading. Below are the specific values. The time periods are January, February, and March, and the corresponding uncollected amounts are 1000w, 2000w, and 30000w.

[0131] Then, once the prompt word is generated, it will be immediately sent to the enabled Wenxinyiyan LLM model. The LLM model is a natural language processing model based on deep learning that can understand and generate natural language text. In this scenario, the Wenxinyiyan LLM model will analyze and process the received prompt word and generate text analysis results, such as Figure 4 shown.

[0132] Finally, the front end establishes a real-time link with the back end through SSE technology. The Wenxinyiyan LLM model will send the text analysis results to the web platform in a streaming manner through the SSE link, and display the analysis results word by word in the dialog box of the web platform interface.

[0133] If the user has turned on the voice broadcast function, the web platform will send the text analysis results to the currently enabled iFlytek TTS model. The iFlytek TTS model will segment the text analysis results and synchronously output the voice stream to the web platform, which will broadcast it in real time.

[0134] The user sees a dialog box like Figure 4 The text analysis results are displayed. Enter a round of conversation information in the dialog box: "Can you tell me what changes have taken place in the market environment during this period?" After receiving the conversation information, the system sends it to the Wenxin Yiyanda language model. The Wenxin Yiyanda language model generates text information "Of course. According to the market data I obtained, during time period 1, the medical device industry was affected by policy adjustments, and some hospitals and clinics had tight capital chains, resulting in longer payment cycles. In time period 2, with the gradual implementation of policies and the adaptation of the market, the payment situation improved." and sends it to the web platform and displays it in the dialog box on the web platform interface.

[0135] The user sees the reply content displayed in the dialog box, and continues to enter a round of conversation information in the dialog box: "Based on this data, should our company adjust its collection strategy?" After receiving the conversation information, the system sends it to the Wenxin Yiyanda language model. The Wenxin Yiyanda language model generates a text message: "Although the previous requirement was not to talk about improving the strategy, since you asked, I can give you some suggestions. Based on the market conditions and your data, I think the company can consider granting a certain grace period to some customers with good credit during the policy adjustment period to maintain good customer relationships; at the same time, for customers with poor credit or long-term arrears, we can strengthen collection efforts to ensure the recovery of funds." It is sent to the web platform and displayed in the dialog box on the web platform interface.

[0136] The user sees the reply content displayed in the dialog box and continues to enter a round of conversation information in the dialog box: "These suggestions are very useful, thank you. Can you export the analysis results of this conversation?" After receiving the conversation information, the system sends it to the Wenxin Yiyanda language model. The Wenxin Yiyanda language model generates a text message "Of course. I will export the complete content of this conversation and my analysis results into a PDF file for your subsequent use or sharing. Please wait a moment, I will generate and send this file for you." and sends it to the web platform and displays it in the dialog box on the web platform interface.

[0137] Through multiple rounds of dialogue, combined with market conditions and other conditions, the data is deeply analyzed, and the analysis results of this dialogue can be exported. Users not only get an in-depth analysis of the uncollected data, but also a comprehensive interpretation combined with market conditions and practical suggestions for improvement. Finally, users can also export analysis results or analysis reports for subsequent use or sharing with team members.

[0138] The entire process is automated and real-time, and can provide high-quality analysis results, greatly improving user experience and work efficiency.

[0139] Data analysis can be performed on each business module. The above embodiment only takes one business module A as an example, but the present invention is not limited to the scenario of performing data analysis on business module A. Performing data analysis on other business modules using the method of the embodiment of the present disclosure is also within the scope of protection of the present disclosure.

[0140] The present application also provides a computer device. Figure 5 , Figure 5 This is a basic structural block diagram of the computer device in this embodiment.

[0141] The computer device includes a memory 510 and a processor 520 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device with components 510-520, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0142] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.

[0143] The memory 510 includes at least one type of readable storage medium, and the readable storage medium includes a non-volatile memory or a volatile memory, for example, a flash memory, a hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc., and the RAM may include a static RAM or a dynamic RAM. In some embodiments, the memory 510 may be an internal storage unit of a computer device, for example, a hard disk or a memory of the computer device. In other embodiments, the memory 310 may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card or a flash card (FlashCard) equipped on the computer device. Of course, the memory 510 may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory 510 is generally used to store the operating system and various application software installed on the computer device, such as the program code of the above method. In addition, the memory 510 may also be used to temporarily store various data that have been output or are to be output.

[0144] The processor 520 is generally used to perform the overall operation of the computer device. In this embodiment, the memory 510 is used to store program codes or instructions, the program code includes computer operation instructions, and the processor 520 is used to execute the program codes or instructions stored in the memory 510 or process data, such as running the program code of the above method.

[0145] In this article, the bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus system can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0146] Another embodiment of the present application also provides a computer-readable medium, which may be a computer-readable signal medium or a computer-readable medium. A processor in a computer reads a computer-readable program code stored in the computer-readable medium, so that the processor can execute the functional actions specified in each step or a combination of steps in the above method; and generate a device for implementing the functional actions specified in each block or a combination of blocks in the block diagram.

[0147] Computer-readable media include but are not limited to electronic, magnetic, optical, electromagnetic, infrared memory or semiconductor systems, devices or apparatuses, or any appropriate combination of the foregoing, the memory is used to store program codes or instructions, the program codes include computer operating instructions, and the processor is used to execute the program codes or instructions of the above methods stored in the memory.

[0148] For the definitions of memory and processor, please refer to the description of the aforementioned computer device embodiment and will not be repeated here.

[0149] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication link shown or discussed can be through some interfaces, the indirect coupling or communication link of devices or units can be electrical, mechanical or other forms.

[0150] Each functional unit or module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0151] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program code.

[0152] Unless the context clearly indicates otherwise, the singular form of the words used herein and in the appended claims includes the plural and vice versa. Thus, when referring to the singular, the plural form of the corresponding term is generally included. Similarly, the words "comprise" and "include" are to be interpreted as inclusive rather than exclusive. Likewise, the terms "include" and "or" should be interpreted as inclusive unless such interpretation is expressly prohibited herein. Where the term "example" is used herein, particularly when it is located after a group of terms, the "example" is merely exemplary and illustrative and should not be considered exclusive or comprehensive.

[0153] Further aspects and scopes of adaptability become apparent from the description provided herein. It should be understood that various aspects of the present application can be implemented individually or in combination with one or more other aspects. It should also be understood that the description and specific embodiments herein are intended for purposes of illustration only and are not intended to limit the scope of the present application.

[0154] Several embodiments of the present disclosure are described in detail above, but it is obvious that those skilled in the art can make various modifications and variations to the embodiments of the present disclosure without departing from the spirit and scope of the present disclosure. The protection scope of the present disclosure is defined by the attached claims.

Claims

1. A method for AI intelligent data analysis of digital supply chain of medical devices, characterized in that: include: Receive a model configuration request from a user, parse and obtain model configuration information, and complete configuration and linking with a selected large language model and text-to-speech model according to the model configuration information; Receive a prompt word template configuration request from a user, parse and obtain template configuration information, construct a prompt word template based on the template configuration information, and store it in a database, wherein the template configuration information includes: template code, template name, system person setting and template content; When receiving a request for initiating business analysis, the template code and business data are parsed and obtained, the template code is searched in the database, the corresponding prompt word template is matched, the template content and system persona are extracted from the prompt word template, and the template content is replaced with the business data to generate prompt words, the prompt words and the system persona are input into the selected large language model, the returned text analysis results are obtained, and the results are immediately sent to the dialog box of the web platform for display; If the voice broadcast function has been enabled, the text analysis result is input into the selected text-to-speech model, the returned synchronous voice stream is obtained, and it is immediately sent to the web platform for broadcasting.

2. The method according to claim 1, characterized in that Before the step of receiving the model configuration request from the user and parsing and obtaining the model configuration option information, the method further includes: Configure necessary information, including access keys and security keys.

3. The method according to claim 1, characterized in that: The template content includes: analyzed data, return format, requirements and analysis results; The analysis results include: risk assessment, advantage analysis, lue analysis, recommendations and data implications; Replacing the template content with the business data, specifically, replacing the analyzed data in the template content with the business data.

4. The method according to claim 1, characterized in that: The template content is replaced with the business data using a variable replacement method.

5. The method according to claim 1, characterized in that Also includes: The selected text-to-speech model performs sentence segmentation processing on the text in the text analysis result based on symbols or character lengths as the sentence segmentation basis of the text; Processing the text length of the text in the text analysis result. If the length of the text after the segmentation is too short, the segmented text is merged with the subsequent segmented text for broadcasting. If the length of the text after the segmentation is too long, it is divided into multiple shorter text segments and broadcasted multiple times. The text analysis results are managed into a queue to be broadcasted, new sentence-segmented text that can be broadcasted is pushed into the queue to be broadcasted, the queue to be broadcasted is monitored, and the voice stream is output in batches in sequence, and the content in the voice stream is broadcasted according to the voice stream; When a symbol is detected or a specific character length is reached, the content of that paragraph will be broadcast first. At the same time, the result return process will not be interrupted. After the content of the current paragraph is broadcast, the remaining content will continue to be broadcast according to the same rules until all the content is completed.

6. The method according to claim 1, characterized in that Through multiple rounds of conversations, the large language model combines industry best practices to conduct in-depth analysis of the business data related to the medical device and derive the text analysis results.

7. The method according to claim 6, characterized in that After the step of exporting the text analysis result, the following steps are included: The text analysis results of the conversation are exported as an analysis report, which includes: data meaning, risk assessment, analysis results of strengths and weaknesses, and improvement suggestions for the analysis results.

8. A device for AI intelligent data analysis of digital supply chain of medical devices, characterized in that: include: A model configuration module, used to receive a model configuration request from a user, parse and obtain model configuration information, and complete the configuration and linking with the selected large language model and text-to-speech model according to the model configuration information; The template configuration module is used to receive a prompt word template configuration request from a user, parse and obtain template configuration information, construct a prompt word template based on the template configuration information, and store it in a database. The template configuration information includes: template code, template name, system person settings and template content; The first processing module is used to parse and obtain the template code and business data when receiving a business analysis initiation request, query the template code in the database, match the corresponding prompt word template, extract the template content and system persona from the prompt word template, replace the template content with the business data, generate prompt words, input the prompt words and the system persona into the selected large language model, obtain the returned text analysis results, and send them to the dialog box of the web platform for display in real time; The second processing module is used to input the text analysis result into the selected text-to-speech model if the voice broadcast function has been enabled, obtain the returned synchronous voice stream, and send it to the web platform for broadcast in real time.

9. A computer device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images