Detection result normalization method, device and equipment based on large model

Abstract

The invention relates to a detection result normalization method, device and equipment based on a large model, and the method comprises the steps: inputting a hardware detection result into a normalization processing system, dividing the detection result into numeric type result data and text type result data through data routing, and carrying out the normalization processing according to the type of the detection result. The numeric type result data is subjected to numeric type processing of a normalization system to generate the most similar detection result description and the detection result meaning corresponding to the numeric value, and the text type result data is subjected to natural language type processing of the normalization system to generate natural language detection result description; and inputting the most similar detection result description, the detection result meaning and the detection result description into a large model for intelligent analysis, and generating a conditioning suggestion. According to the method, intelligent result analysis can be carried out by utilizing the large model, a numerical result normalization method is provided to adapt to the requirements of different detection data, the large model can also provide personalized diagnosis and conditioning suggestions according to the health data of the user, and the system can learn and improve from new detection data, so that the user experience is improved. The method adapts to new data and traditional Chinese medicine theory development.

Description

Technical Field

[0001] This application relates to the field of computer software technology, and particularly to a method, device, and equipment for normalizing detection results based on a large model. Background Art

[0002] With the development of technology, the combination of modern medicine and traditional Chinese medicine has become increasingly close. Especially in the application of intelligent detection equipment, it has significantly improved the objectivity and accuracy of traditional Chinese medicine diagnosis. Traditional Chinese medicine diagnosis methods include tongue diagnosis, pulse diagnosis, and face diagnosis, etc. By observing the tongue image, pulse condition, and face image of the patient, doctors can obtain rich diagnostic information. To better assist traditional Chinese medicine diagnosis, a variety of traditional Chinese medicine intelligent detection hardware devices have been developed in the market, such as tongue diagnosis instruments, pulse diagnosis instruments, and face diagnosis instruments. These devices provide objective diagnostic data through advanced sensing technology and data processing technology, greatly facilitating the process of traditional Chinese medicine diagnosis.

[0003] Although the emergence of traditional Chinese medicine intelligent detection hardware such as face diagnosis instruments, tongue diagnosis instruments, and pulse diagnosis instruments has greatly improved the objectivity and precision of traditional Chinese medicine diagnosis, in practical applications, there are still some significant problems and challenges. These problems mainly focus on the following aspects: inconsistent data formats, difficulties in data standardization and normalization, insufficient comprehensive analysis and decision support, and inadequate combination with large language models, which limit their effectiveness in practical applications. To overcome these problems, it is urgent to develop a system and method for normalizing the detection results of traditional Chinese medicine intelligent hardware based on a large model. By constructing an intermediate layer for numerical normalization, standard processing of the detection results of multiple devices can be achieved, improving the intelligence and accuracy of traditional Chinese medicine diagnosis. Summary of the Invention

[0004] This application provides a method for normalizing detection results based on a large model, which is characterized by including: Input the hardware detection results into the normalization processing system, and divide the detection results into numerical result data and text result data through data routing; According to the type of detection results, the numerical result data generates the most similar detection result description and the detection result meaning corresponding to the numerical value through the numerical type processing of the normalization system, and the text result data generates a natural language detection result description through the natural language type processing of the normalization system; Input the most similar detection result description, detection result meaning, and the detection result description into the large model for intelligent analysis to generate conditioning suggestions.

[0005] Optionally, the method for normalizing detection results based on a large model is characterized in that: The numerical type processing of the normalization system is to process the numerical result data using a RAG query module and a large model numerical processing module; The natural language type processing of the normalization system is to process the text result data using a predefined description template; The comprehensive processing of the processing results is to comprehensively describe and splice the processing results of the numerical result data and the text result data.

[0006] Optionally, the numerical type processing of the normalization system is to process the numerical result data using a RAG query module and a large model numerical processing module, including: Input the numerical result data into the RAG query module and the large model numerical processing module respectively; Use the RAG query module to perform preliminary normalization processing according to the detection results and construct a vector database, and perform similarity queries using the vector database according to the normalized values to obtain the most similar detection result description; Use the large model numerical processing module to perform preliminary normalization processing according to the detection results and construct a training dataset, and use the training dataset to input the large model for fine-tuning to obtain the meaning of the traditional Chinese medicine detection results corresponding to each numerical value.

[0007] Optionally, the method for normalizing the detection results based on a large model is characterized in that: The system summarizes and classifies the detection results of various intelligent detection devices; Use a large model for intelligent analysis to generate a comprehensive report and conditioning suggestions.

[0008] Optionally, the natural language type processing of the normalization system is to process the text result data using a predefined description template, including: Fill the detection results into the predefined description template to generate a description of the condition.

[0009] Optionally, the comprehensive processing of the processing results is to comprehensively describe and splice the processing results of the numerical result data and the text result data, including: The comprehensive description is to splice the most similar detection result descriptions together; The splicing process is to splice the comprehensive description, the meaning of the traditional Chinese medicine detection results, and the description of the condition.

[0010] Optionally, the method for normalizing the detection results based on a large model is characterized in that: The construction of the vector database and the fine-tuning of the large model are both offline training.

[0011] The present application also provides a device for normalizing detection results based on a large model, characterized in that the device includes: a memory, a processor, and a computer program stored on the memory and running on the processor, and the computer program is configured to implement the steps of the method for normalizing detection results based on a large model as described in any one of claims 1 to 7.

[0012] The present application also provides a device for normalizing detection results based on a large model, characterized in that the device includes: a face diagnosis instrument, a tongue diagnosis instrument, and a pulse diagnosis instrument, and the device provides objective detection data through advanced sensing technology and data processing technology.

[0013] The present application also provides an electronic device, characterized in that it includes: a memory and a processor, and a computer program is stored in the memory, wherein when the processor executes the computer program, the method for normalizing detection results based on a large model as described in any one of claims 1 to 7 is implemented.

[0014] The beneficial effects of the present application are as follows: The use of a large model improves the accuracy of the detection results of traditional Chinese medicine intelligent hardware. The system can use the large model for intelligent result analysis, and provides a normalization method for numerical result data to meet the needs of different detection data. The large model can also provide personalized diagnosis and conditioning suggestions based on the user's health data, and the system can learn and improve from new detection data to adapt to new data and the development of traditional Chinese medicine theory. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following briefly introduces the drawings required in the description of the embodiments or the prior art. Obviously, the following drawings are only the embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0016] Figure 1 Shows a flowchart of a method for normalizing detection results based on a large model disclosed in the present application; Figure 2 Shows an overall flow architecture diagram of a traditional Chinese medicine large model combined with a data normalization algorithm disclosed in the present application; Figure 3 Shows a real-time processing architecture diagram of a traditional Chinese medicine large model disclosed in the present application; Figure 4 Shows a process diagram of data offline RAG processing disclosed in the present application; Figure 5 Shows a process diagram of data offline fine-tuning training disclosed in the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Various exemplary embodiments, features, and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0018] Among them, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality" means two or more unless otherwise specifically defined.

[0019] The term "exemplary" used herein means "serving as an example, embodiment, or illustration". Any embodiment described herein as "exemplary" does not have to be construed as superior or better than other embodiments.

[0020] In addition, for a better description of the present application, numerous specific details are given in the following specific implementation manners. Those skilled in the art should understand that the present application can also be implemented without some specific details. In some instances, methods, means, elements, and circuits well-known to those skilled in the art are not described in detail so as to highlight the gist of the present application.

[0021] The present application is a method for detecting data normalization. In this method, the system can obtain detection data through a traditional Chinese medicine intelligent detection hardware device worn by the user, and then the system classifies the detection data, generates corresponding results according to the data type and performs comprehensive processing, and finally inputs them into a large model for intelligent analysis to output conditioning suggestions.

[0022] As Figure 1 shown, it is a method for normalizing detection results based on a large model according to an embodiment of the present application, specifically including the following steps: S100, the hardware detection results are input into the normalization processing system, and the detection results are divided into numerical result data and text result data through data routing.

[0023] In this step, user detection data is obtained from various traditional Chinese medicine intelligent detection hardware devices and input into the data routing of the data normalization processing system, and is divided into data type and text type result data according to the data type.

[0024] S200, according to the type of detection results, the numerical result data generates the most similar detection result description and the detection result meaning corresponding to the numerical value through the numerical type processing of the normalization system, and the text result data generates a natural language detection result description through the natural language type processing of the normalization system.

[0025] In this step, the RAG query module and the large model numerical processing module in the data normalization processing system are used to process the numerical result data respectively. By using the vector database to query the most similar results and the fine-tuned large model to obtain the meaning of the traditional Chinese medicine detection results corresponding to the numerical values, the processing results of the numerical result data are generated.

[0026] For the text result data, the predefined description template in the data normalization processing system is used to fill the text content of the detection result into the template to generate the processing result of the text result data.

[0027] S300, input the description of the most similar detection result, the meaning of the detection result, and the description of the detection result into the large model for intelligent analysis to generate conditioning suggestions.

[0028] In this step, the most similar results in the processing results of the numerical result data are spliced into a comprehensive description, and then the comprehensive description, the meaning of the traditional Chinese medicine detection results corresponding to the numerical values, and the processing results of the text result data are spliced and input to generate a comprehensive description containing the above content, which is input to the traditional Chinese medicine large model for comprehensive analysis and output conditioning suggestions.

[0029] The method for normalizing the detection results based on the large model in this embodiment can improve the accuracy of the detection results of traditional Chinese medicine intelligent hardware through the use of the large model, adapt to the needs of different detection data according to the normalization methods of the numerical result data in the RAG query module and the large model numerical processing module, and the system can use the large model for intelligent result analysis.

[0030] The overall technical architecture of the above technical solution is as Figure 3 shown in the real-time processing architecture diagram of a traditional Chinese medicine large model disclosed in this application, which specifically includes the following content: Input the tongue diagnosis results, face diagnosis results, hand diagnosis results, pulse diagnosis results, interrogation results, and constitution detection results obtained by the traditional Chinese medicine intelligent detection hardware device into the data normalization processing system.

[0031] In the system, first classify the above detection results through data routing processing, where the data type processing is input to the RAG query module and the large model numerical processing module, and the natural language type directly passes through the predefined description template to generate the corresponding processing results. The processing results are summarized, naturally language spliced, and Prompt optimized to form a comprehensive description.

[0032] Finally, the comprehensive description generated by the data normalization processing system is input to the traditional Chinese medicine large model for intelligent analysis, and the large model outputs conditioning suggestions to the user.

[0033] Specifically, the numerical type processing of the normalization system refers to using the RAG query module and the large model numerical processing module to process the numerical result data.

[0034] The RAG query module uses a vector database for RAG (Retrieval-Augmented Generation). RAG is a technology that combines retrieval and generation, suitable for retrieving relevant information from large-scale datasets and generating relevant answers. For numerical result data, similar queries can be performed through the vector database to find the closest description of the detection results, and then these results are input into a large model for analysis.

[0035] The large model numerical processing module fine-tunes the large model to enable it to understand and process numerical result data, and combines the TCM detection result descriptions corresponding to each numerical value for training.

[0036] Specifically, the natural language type processing of the normalization system refers to processing text result data using predefined description templates.

[0037] Using predefined description templates means filling the detection results into the predefined templates.

[0038] Specifically, the comprehensive processing of the processing results refers to the comprehensive description and splicing processing of the processing results of numerical result data and text result data.

[0039] As described in the above steps, the RAG query module includes the following steps: Specifically, perform preliminary normalization processing on the detection results and construct a vector database. First, obtain the original detection results of the pulse diagnosis instrument. For example, the detection values of each meridian are between -100 and 100. Then perform preliminary normalization processing to normalize each detection value to between 0 and 1. The normalization formula is as follows: X′=(X−Xmin) / (Xmax−Xmin), where X is the original detection result, X is the normalized result, Xmin = -100, and Xmax = 100.

[0040] Next, construct a vector database. First, define the detection result description template, that is, define the detection result description corresponding to each detection value according to TCM theory. For example: Detection result description corresponding to the liver meridian numerical value: X>30X>30X>30: "Excessive liver fire, which may be caused by rapid liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc." 30≤X≤30 - 30 \leq X \leq 30−30≤X≤30: "The liver meridian is normal." X<−30X<-30X<−30: "Low liver function, which may be caused by long-term fatigue, liver damage, malnutrition, etc." Description of the detection results corresponding to the stomach meridian values: X > 30 X > 30 X > 30: "Excessive stomach fire, which may be caused by excessive gastric acid secretion, easy stomachache, gastric ulcer, gastroesophageal reflux, etc." −30 ≤ X ≤ 30 -30 \leq X \leq 30−30≤X≤30: "The stomach meridian is normal." X < −30 X < -30 X < −30: "Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." Vectorize the values of each meridian and the corresponding detection result descriptions according to the above content, and store them in a vector database to complete the construction of the vector database.

[0041] Finally, perform a similarity query using the vector database based on the normalized values to obtain the most similar detection result description. According to the input normalized values, use the vector database to perform a similarity query to find the closest detection result description.

[0042] The following is a specific example of using the RAG query module: Original data: Pulse diagnosis instrument (numerical result data): Liver meridian: 50, Stomach meridian: -40 Normalized data: Liver meridian: 0.75; Stomach meridian: 0.3 Generation of detection result descriptions: Liver meridian: 0.75 -> "Excessive liver fire, which may be caused by faster liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc." Stomach meridian: 0.3 -> "Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." Vector database query: Input: {"Liver meridian": 0.75, "Stomach meridian": 0.3} Output: "Excessive liver fire, which may be caused by faster liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc. Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." As described in the above steps, the large model numerical processing module includes the following steps: Specifically, perform preliminary normalization processing based on the detection results and construct a training dataset. First, obtain the original detection results of the pulse diagnosis instrument, for example, the detection values of each meridian are between -100 and 100. Then perform preliminary normalization processing on the original detection results for subsequent use.

[0043] Next, construct a training dataset. First, define a detection result description template, that is, define the detection result description corresponding to each detection value according to traditional Chinese medicine theory. For example: Description of the detection results corresponding to the liver meridian values: X > 30 X > 30 X > 30: "Excessive liver fire, which may be caused by rapid liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc." −30 ≤ X ≤ 30 -30 \leq X \leq 30−30≤X≤30: "The liver meridian is normal." X < −30 X < -30 X < −30: "Low liver function, which may be caused by long-term fatigue, liver damage, malnutrition, etc." Description of the detection results corresponding to the stomach meridian values: X > 30 X > 30 X > 30: "Excessive stomach fire, which may be caused by excessive gastric acid secretion, easy stomachache, gastric ulcer, gastroesophageal reflux, etc." −30 ≤ X ≤ 30 -30 \leq X \leq 30−30≤X≤30: "The stomach meridian is normal." X < −30 X < -30 X < −30: "Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." Generate training data according to the above content, and construct a training data set with the normalized values and corresponding natural language descriptions. For example: Input data: {"liver meridian": 0.75, "stomach meridian": 0.3} Output description: "Excessive liver fire, which may be caused by rapid liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc. Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." Finally, fine-tune the large model, and input the constructed training data set into the large model for fine-tuning, so that the large model can understand the meaning of the traditional Chinese medicine detection results corresponding to each value.

[0044] The following is a specific example of using the numerical processing module of the large model: Original data: Pulse detector (numerical result data): Liver meridian: 50, Stomach meridian: -40 Normalized data: Liver meridian: 0.75; Stomach meridian: 0.3 Generation of detection result description: Liver meridian: 0.75 -> "Excessive liver fire, which may be caused by rapid liver metabolism, taking medicine, excessive exercise, ingesting foods that require liver metabolism, liver inflammation, lack of sleep, etc." Stomach meridian: 0.3 -> "Low stomach function, which may be caused by insufficient gastric acid secretion, indigestion, stomach cold, etc." Training data set: Input: {"liver meridian": 0.75, "stomach meridian": 0.3} Output: "Excessive liver fire may be caused by factors such as rapid liver metabolism, taking medications, excessive exercise, consuming foods that require liver metabolism, liver inflammation, and lack of sleep. Low stomach function may be caused by insufficient gastric acid secretion, indigestion, and stomach cold." As described in the above steps, the natural language type is to use the detection results to fill in the predefined description template to generate a description of the condition.

[0045] First, data collection and preprocessing are carried out. For example, the original detection results of the tongue diagnosis instrument are obtained: such as the color and thickness of the tongue coating. Then, a natural language description is generated according to the predefined description template, that is, the detection results are filled into the predefined template. For example: Template: "The thickness of the patient's tongue coating is {thickness}, and the color is {color}." Original data: Tongue coating thickness: 3mm, color: light white Description: "The thickness of the patient's tongue coating is 3mm, and the color is light white." The following is a specific example of using the predefined description template: Original data: Tongue diagnosis instrument (natural language type): Tongue coating thickness: 3mm, color: light white Natural language description: "The thickness of the patient's tongue coating is 3mm, and the color is light white." As described in the above steps, the comprehensive description refers to splicing the descriptions of the most similar detection results together. For example: The description of the numerical result data is: Liver meridian: 0.75 -> "Excessive liver fire may be caused by factors such as rapid liver metabolism, taking medications, excessive exercise, consuming foods that require liver metabolism, liver inflammation, and lack of sleep." Stomach meridian: 0.3 -> "Low stomach function may be caused by insufficient gastric acid secretion, indigestion, and stomach cold." The comprehensive description is: "The normalized value of the liver meridian is 0.75 (excessive liver fire), and the normalized value of the stomach meridian is 0.3 (low stomach function)." The splicing process is to splice the comprehensive description, the meaning of the traditional Chinese medicine detection results, and the description of the condition together. For example: The description of the numerical result data is: "The normalized value of the liver meridian is 0.75 (excessive liver fire), and the normalized value of the stomach meridian is 0.3 (low stomach function)." The natural language type description is: "The thickness of the patient's tongue coating is 3mm, and the color is light white." The splicing input is: "Pulse diagnosis result: The normalized value of the liver meridian is 0.75 (excessive liver fire), and the normalized value of the stomach meridian is 0.3 (low stomach function). The thickness of the patient's tongue coating is 3mm, and the color is light white." Specifically, the system can summarize and classify the detection results of various intelligent detection devices, and perform intelligent analysis using a large model to generate a comprehensive report and conditioning suggestions. The various intelligent detection devices refer to traditional Chinese medicine intelligent detection hardware devices such as face diagnosis devices, tongue diagnosis devices, and pulse diagnosis devices.

[0046] Specifically, both the construction of the vector database and the fine-tuning of the large model are offline training. As Figure 2 shown, the overall process architecture diagram of the traditional Chinese medicine large model combined with the data normalization algorithm specifically includes the following content: The real-time deduction architecture is: collecting detection data through traditional Chinese medicine intelligent detection hardware, then sending it to the detection result normalization intermediate layer for data processing, and finally sending the processed result to the traditional Chinese medicine large model to output conditioning suggestions.

[0047] The offline training architecture is: searching and cleaning the traditional Chinese medicine detection results, and then sending them to the detection result normalization intermediate layer for training to establish a vector database and fine-tune the large model.

[0048] As Figure 4 shown, the data offline RAG processing process diagram specifically includes the following content: Input and store data value detection results such as tongue diagnosis results, face diagnosis results, hand diagnosis results, pulse diagnosis results, interrogation results, and constitution detection results in the vector database, interpret and analyze the detection results, and vectorize the numerical values of each meridian and the corresponding detection result descriptions to form a vector database.

[0049] As Figure 5 shown, the data offline fine-tuning training process diagram specifically includes the following content: Clean the numerical result data such as tongue diagnosis results, face diagnosis results, hand diagnosis results, pulse diagnosis results, interrogation results, and constitution detection results and the result interpretation data set to improve the accuracy and consistency of the data and ensure the data analysis results. Finally, input the cleaned data into the traditional Chinese medicine large model to complete the offline fine-tuning training of the data.

[0050] Based on the same inventive concept, the present application also provides a device for normalizing detection results based on a large model. The device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. The computer program is configured to implement the steps of any one of the methods for normalizing detection results based on a large model.

[0051] The above method can be implemented by the following device for normalizing detection results based on a large model. The device includes: a face diagnosis device, a tongue diagnosis device, and a pulse diagnosis device. The device provides objective detection data through advanced sensing technology and data processing technology.

[0052] Specifically, for a face diagnosis instrument, face diagnosis is one of the important methods in traditional Chinese medicine diagnosis. By observing changes in the color, luster, texture, etc. of the face, doctors can judge the health status of patients. The face diagnosis instrument uses a high-resolution camera and image processing technology to comprehensively scan and analyze the face, and extract characteristic data such as facial color, texture, and shape. Modern face diagnosis instruments usually come with a spectral analysis function, which can accurately detect the pigment distribution and blood flow conditions in different parts of the face, thus providing reliable data support for traditional Chinese medicine diagnosis.

[0053] For a tongue diagnosis instrument, tongue diagnosis is an important part of traditional Chinese medicine diagnosis. By observing the color, thickness, moisture level of the tongue coating, and the shape of the tongue body, doctors can infer the visceral functions and lesion sites of patients. The tongue diagnosis instrument uses a high-definition camera and image recognition technology to capture and analyze tongue images, and extract characteristics such as the color, thickness, distribution of the tongue coating, and the shape of the tongue body. Some tongue diagnosis instruments are also equipped with 3D imaging technology, which can more comprehensively present the three-dimensional structure of the tongue body and provide more detailed reference data for doctors.

[0054] For a pulse diagnosis instrument, pulse diagnosis is another important method in traditional Chinese medicine diagnosis. By perceiving the frequency, strength, rhythm, and shape of the pulse, doctors can understand the cardiovascular health status and other visceral problems of patients. The pulse diagnosis instrument uses a pressure sensor and waveform analysis technology to accurately measure pulse waveform and pulse pressure data. Modern pulse diagnosis instruments usually have a multi-channel detection function, which can simultaneously monitor multiple pulse positions and provide more comprehensive pulse information. By analyzing the pulse waveform, the pulse diagnosis instrument can identify minute changes in the pulse condition and provide an objective basis for traditional Chinese medicine diagnosis.

[0055] Meanwhile, the present application also provides an electronic device, including: a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, it implements the method for normalizing the detection results based on a large model in any of the foregoing embodiments.

[0056] The above has described the embodiments of the present application. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technologies in the market, or to enable other ordinary skill in the art in the technical field to understand the disclosed embodiments herein.

Claims

1. A method for normalizing detection results based on a large model, characterized in that: include: The hardware detection results are input into the normalization processing system, and the detection results are divided into numerical result data and text result data through data routing; According to the test result type, the numerical result data is processed by the normalization system to generate the most similar test result description and the meaning of the test result corresponding to the numerical value, and the text result data is processed by the normalization system to generate a natural language test result description; The most similar test result description, the meaning of the test result and the test result description are input into the big model for intelligent analysis to generate conditioning suggestions.

2. The method for normalizing detection results based on a large model as claimed in claim 1, characterized in that: The numerical type processing of the normalization system is to use the RAG query module and the large model numerical processing module to process the numerical result data; The natural language type processing of the normalization system is to process the text type result data using a predefined description template; The comprehensive processing of the processing results is to comprehensively describe and splice the processing results of the numerical result data and the text result data.

3. The method for normalizing detection results based on a large model as claimed in claim 2, characterized in that: The numerical type processing of the normalization system is to use the RAG query module and the large model numerical processing module to process the numerical result data, including: Input the numerical result data into the RAG query module and the large model numerical processing module respectively; Use the RAG query module to perform preliminary normalization processing based on the detection results and build a vector database. Use the vector database to perform similarity queries based on the normalized values ​​to obtain the most similar detection result description; Use the large model numerical processing module to perform preliminary normalization processing based on the test results and build a training data set. Use the training data set to input the large model for fine-tuning to obtain the meaning of the TCM test results corresponding to each value.

4. The method for normalizing detection results based on a large model as claimed in claim 1, characterized in that: The system summarizes and classifies the test results of various intelligent testing equipment; Use big models for intelligent analysis and generate comprehensive reports and conditioning recommendations.

5. The method for normalizing detection results based on a large model as claimed in claim 2, characterized in that: The natural language type processing of the normalization system is to process the text type result data using a predefined description template, including: Use the test results to fill in the predefined description template to generate a description of the condition.

6. The method for normalizing detection results based on a large model as claimed in claim 2, characterized in that: The comprehensive processing of the processing results is to comprehensively describe and splice the processing results of the numerical result data and the text result data, including: The comprehensive description is to splice together the most similar detection result descriptions; The splicing process is to splice the comprehensive description, the meaning of the TCM test results and the description of the condition.

7. The method for normalizing detection results based on a large model as claimed in claim 3, characterized in that: The construction of the vector database and the fine-tuning of the large model are both offline training.

8. A device for normalizing detection results based on a large model, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the computer program is configured to implement the steps of the method for normalizing detection results based on a large model as described in any one of claims 1 to 7.

9. A device for normalizing detection results based on a large model, characterized in that: The device comprises a facial diagnosis instrument, a tongue diagnosis instrument and a pulse diagnosis instrument, and provides objective detection data through advanced sensing technology and data processing technology.

10. An electronic device, characterized in that: include: A memory and a processor, wherein a computer program is stored in the memory, wherein the processor implements the method for normalizing detection results based on a large model as described in any one of claims 1 to 7 when executing the computer program.

Images