A method for generating doctor-patient conversation data based on IBS conditions

Abstract

The application discloses a doctor-patient conversation data generation method based on IBS diseases, and relates to the technical field of medical data processing, and comprises the following steps: S1, effective case data set construction; S2, natural language processing preprocessing; S3, conversation generation rule base construction; S4, doctor-patient conversation data simulation generation; S5, conversation data quality evaluation; S6, rule base iterative updating. The doctor-patient conversation data generation method based on IBS diseases integrates multiple calculation formulas such as case efficiency, TF-IDF, information correlation, perplexity, comprehensive score and optimization priority into the whole process from case screening to rule base updating, replaces traditional artificial subjective judgment, ensures the scientificity, consistency and repeatability of the processing results of each link, and greatly improves the quality stability of the generated data.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, specifically to a method for generating doctor-patient dialogue data based on IBS symptoms. Background Technology

[0002] Irritable bowel syndrome (IBS) is a common functional gastrointestinal disorder in clinical gastroenterology. Hospitals at all levels have accumulated a massive amount of historical case data on IBS, which is mostly in unstructured text format and contains complete clinical information from chief complaint to treatment recommendations. With the rapid development of medical artificial intelligence technology, the development of products such as intelligent consultation robots, IBS-assisted diagnostic models, and medical scenario dialogue systems requires a large amount of standardized doctor-patient dialogue data that conforms to clinical logic, fits the characteristics of IBS diagnosis and treatment, and serves as training samples.

[0003] Existing methods for generating medical dialogue data suffer from several shortcomings: First, they are mostly general-purpose methods, lacking customization for IBS-specific treatment processes and terminology, resulting in low clinical relevance of the generated data. Second, they lack a quantitative calculation system, relying on subjective human judgment in case screening, keyword extraction, and rule construction, leading to inconsistent and unscientific results. Third, the quality assessment of generated dialogue data lacks unified quantitative standards, making it difficult to guarantee data authenticity and accuracy. Fourth, the absence of a quantitative iterative update mechanism for the rule base prevents continuous optimization of the generated data quality. Furthermore, existing original IBS doctor-patient dialogue records in hospitals suffer from limited quantity, sensitive privacy information, and inconsistent formats, making them unsuitable for direct model training. Therefore, a method for generating IBS-specific doctor-patient dialogue data based on real cases and incorporating quantitative calculations is urgently needed. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for generating doctor-patient dialogue data based on IBS symptoms, thus solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for generating doctor-patient dialogue data based on IBS symptoms, comprising the following steps:

[0006] S1. Construction of effective case dataset: Obtain historical case data related to IBS stored in the hospital, filter the original case data to remove invalid cases, and after the effectiveness rate is quantified and calculated, obtain an effective case dataset with an effectiveness rate that meets the preset requirements.

[0007] S2. Natural Language Processing Preprocessing: Perform full-process natural language processing on the effective case dataset to extract and transform unstructured text information into structured information, and obtain structured key diagnostic and treatment information that can be used for subsequent rule construction;

[0008] S3. Construction of Dialogue Generation Rule Base: Based on the structured key diagnosis and treatment information, the doctor-patient consultation process is sorted out in combination with the clinical diagnosis and treatment logic of IBS, the correlation pattern of diagnosis and treatment information is explored and standardized dialogue templates and information matching rules are established, and integrated to form a doctor-patient dialogue generation rule base exclusive to IBS.

[0009] S4. Simulation and generation of doctor-patient dialogue data: Based on the dialogue generation rule base, select the diagnosis and treatment information unit to match the corresponding template, and use natural language generation technology to instantiate the template to generate doctor-patient dialogue text. Optimize the fluency of the text and establish an information association and traceability relationship with the original case.

[0010] S5. Dialogue Data Quality Assessment: Construct a multi-dimensional doctor-patient dialogue data quality assessment system, and use a combination of human and machine scoring methods to comprehensively and quantitatively evaluate the generated dialogue data. After removing unqualified data and standardizing and labeling qualified data, a standardized IBS disease doctor-patient dialogue dataset is obtained.

[0011] S6. Rule base iterative update: Based on the dialogue data quality assessment results and the newly added historical case data of IBS symptoms, the optimization priority of each module of the dialogue generation rule base is quantitatively determined, and the rule base is modified and supplemented with information according to the priority to achieve iterative optimization and performance improvement of the rule base.

[0012] Preferably, in step S1, the efficiency calculation formula is as follows:

[0013] ;

[0014] in For the number of effective cases, This represents the total number of original cases.

[0015] Preferably, the historical case data related to IBS symptoms includes at least one of the following: patient basic information, IBS symptom description, medical history collection record, physical examination results, auxiliary examination data, diagnostic conclusion, and treatment recommendations. The data screening process involves removing case data that lacks key diagnostic information, contains logical contradictions, or has abnormal formatting. The preset threshold is 80% to 95%. When the calculated case effectiveness rate is lower than this threshold, supplementary historical case data of the same type of IBS is collected and re-screened.

[0016] Preferably, the natural language processing preprocessing in S2 includes text cleaning, word segmentation, part-of-speech tagging, named entity recognition, keyword extraction, and text structure mapping. The keyword extraction uses the TF-IDF formula to specifically extract IBS-specific diagnostic and treatment terms. The TF-IDF formula is:

[0017] ;

[0018] in, For the term In cases Number of times it appears in For cases Total number of occurrences of all terms in the text Total number of cases For included terms The number of cases, the text structuring transforms unstructured case text into structured data in key-value pair form.

[0019] Preferably, the construction process of the IBS patient-doctor dialogue generation rule base includes: based on structured key diagnostic and treatment information, sorting out the typical process of clinical patient-doctor consultation for IBS; extracting the doctor's question types, patient's response logic, and the correlation of diagnostic and treatment information in the consultation process, and applying the correlation formula... Calculate the correlation between medical information, whereby... This represents the number of cases where information A and information B both appear. For the number of cases containing information A, The number of cases containing information B; combined with natural language expression standards, establish doctor's script templates, patient response templates, and information matching rules, and integrate them to form a rule base.

[0020] Preferably, the natural language generation technology includes at least one of rule-based generation technology, template-based generation technology, and pre-trained language model-based generation technology, wherein the pre-trained language model is a Chinese language model adapted to medical text, and the fluency of the generated dialogue text is optimized using a perplexity formula.

[0021] Preferably, the process of simulating and generating doctor-patient dialogue data in S3 is as follows: randomly selecting data from structured key diagnostic and treatment information according to a formula. Randomly or selectively select diagnostic and treatment information units, among which This refers to the total number of structured diagnostic and treatment information units. For the first The probability of extracting each information unit; matching doctor's script templates and patient response templates in the dialogue generation rule base; instantiating and filling the templates using natural language generation technology to generate multi-round, clinically logical IBS symptom doctor-patient dialogue texts, while preserving the information association between the dialogue and the original case.

[0022] Preferably, the process also includes a step of quality assessment of the generated IBS patient-doctor dialogue data: constructing a quality assessment index system for patient-doctor dialogue data, scoring the dialogue data through a combination of manual scoring and automatic machine calculation, and applying a comprehensive scoring formula. Calculate the overall score of the dialogue data, where To evaluate the total number of indicators, For the first The weight coefficients of the indicators and satisfy the following conditions: , For the first The scores of each indicator are calculated; dialogue data with scores below a preset threshold are removed, and qualified dialogue data are labeled to obtain a standardized IBS disease doctor-patient dialogue dataset.

[0023] Preferably, the quality assessment index system includes five primary indicators: information completeness, clinical logic rationality, language fluency, dialogue fit, and terminology accuracy. Each primary indicator has 3 to 5 secondary indicators. The weight coefficients of each primary indicator are allocated according to W1:W2:W3:W4:W5=0.25:0.30:0.15:0.15:0.15. The comprehensive score adopts a percentage system, and the preset passing threshold is 70 to 85 points.

[0024] Preferably, the method further includes the step of iteratively updating the IBS patient-doctor dialogue generation rule base: collecting the quality assessment results of the generated dialogue data, calculating the optimization priority of each rule base module, and combining the optimization priority with newly added historical IBS case data to supplement and correct the dialogue templates, response logic and information matching rules in the rule base.

[0025] Beneficial effects

[0026] This invention provides a method for generating doctor-patient dialogue data based on IBS symptoms. Compared with existing technologies, it has the following advantages:

[0027] (1) The method for generating doctor-patient dialogue data based on IBS symptoms integrates multiple calculation formulas such as case effectiveness rate, TF-IDF, information correlation, confusion, comprehensive score, and optimization priority into the entire process from case screening to rule base update, replacing traditional manual subjective judgment, ensuring the scientificity, consistency and repeatability of the processing results at each stage, and greatly improving the quality and stability of the generated data.

[0028] (2) This method for generating doctor-patient dialogue data based on IBS symptoms, through customized design for IBS symptoms, sorts out the IBS-specific clinical consultation process, extracts core keywords for IBS diagnosis and treatment, establishes matching rules according to the correlation of diagnosis and treatment information, and generates dialogue data that is highly consistent with the actual clinical diagnosis and treatment of IBS, thus solving the problem of low clinical consistency of data generated by general methods.

[0029] (3) The method for generating doctor-patient dialogue data based on IBS symptoms uses a quantitative quality assessment index system and fluency optimization formula to screen and optimize the generated dialogue data in all dimensions, ensuring that the data information is complete, the clinical logic is reasonable, the language is fluent, the terminology is accurate, and standardized annotation is completed. It can be directly used for scenarios such as medical NLP model training and intelligent consultation system development.

[0030] (4) The method for generating doctor-patient dialogue data based on IBS symptoms establishes an iterative update mechanism for the rule base based on the optimization priority formula, combines the quality assessment results and new case data, and optimizes the rule base in a targeted manner to form a closed loop, ensuring that the quality of the generated data continues to improve with iteration, and adapts to the updates of IBS clinical diagnosis and treatment standards and the development needs of medical artificial intelligence. Attached Figure Description

[0031] Figure 1 The flowchart illustrates a method for generating doctor-patient dialogue data based on IBS symptoms, as provided by this invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] This invention provides a technical solution: a method for generating doctor-patient dialogue data based on IBS symptoms, specifically including the following steps:

[0034] S1: Construction of an effective case dataset

[0035] Obtain historical case data related to IBS symptoms stored in the hospital. The data includes at least one of the following: basic patient information (age, gender, physical condition, etc.), description of IBS symptoms (abdominal pain, bloating, changes in bowel habits, etc.), medical history records (past medical history, allergy history, triggers, etc.), physical examination results, auxiliary examination data (stool routine, colonoscopy, etc.), diagnostic conclusions (IBS type, severity of disease, etc.), and treatment recommendations (medication, diet, lifestyle adjustments, etc.).

[0036] The original case data was filtered to remove cases lacking core IBS symptoms, diagnostic conclusions, or other key diagnostic information; cases with logical contradictions between symptoms and diagnoses; and cases with formatting issues such as garbled or incomplete text. The data was then processed according to the effectiveness formula. Calculate the case effectiveness rate, among which The number of valid cases after screening. The original total number of cases is represented by , where 'valid' refers to the number of valid cases after filtering, and 'total' is the original total number of cases. The preset case validity threshold is 80%~95%. When the calculated validity rate is lower than this threshold, supplementary historical case data of the same type of IBS are collected and filtered again until the validity rate reaches the preset threshold, thus obtaining a valid case dataset.

[0037] S2: Natural Language Processing Preprocessing

[0038] The valid case dataset undergoes full-process natural language processing preprocessing, sequentially performing text cleaning, word segmentation, part-of-speech tagging, named entity recognition, keyword extraction, and text structuring. The core steps are keyword extraction and text structuring.

[0039] Keyword extraction: Based on the characteristics of IBS diagnosis and treatment, the TF-IDF formula was used.

[0040] Targeted extraction of IBS-specific diagnostic and treatment terminology, including: IBS diagnosis and treatment terminology In cases Number of times it appears in For cases Total number of occurrences of all terms in the text The total number of cases in the valid case dataset. Include terms in the valid case dataset The number of cases. The TF-IDF value of each term is calculated using this formula, and the top 80% of terms are selected as core diagnostic keywords for IBS, ensuring that the extracted keywords are highly relevant to IBS diagnosis and treatment.

[0041] Text structuring mapping: Transform unstructured case texts into structured data in the form of key-value pairs of "diagnosis and treatment information type - specific content", such as "abdominal pain frequency - 3 times a week", "bow bowel habits - alternating diarrhea and constipation", "diagnosis and treatment recommendations - take antispasmodic drugs + light diet", extract key diagnosis and treatment information, and provide structured data support for the subsequent construction of rule base.

[0042] S3: Construction of a rule base for generating doctor-patient dialogues related to IBS symptoms

[0043] Based on structured key diagnostic and treatment information, combined with IBS clinical treatment guidelines and actual consultation processes, a dedicated rule base for generating doctor-patient dialogues is constructed. The specific process is as follows:

[0044] This paper outlines the typical clinical consultation process for IBS patients, following the clinical logic of "chief complaint inquiry → present illness history collection → past medical history confirmation → symptom detail verification → examination result inquiry → diagnosis conclusion communication → treatment recommendation notification".

[0045] Extract the relationships between doctor's question types (closed-ended questions, open-ended questions), patient's response logic (symptom description, result confirmation, question feedback), and medical information in the consultation process, and apply the correlation formula. Calculate the correlation between any two pieces of medical records, where, This represents the number of cases in which information A and information B both appear in the valid case dataset. For the number of cases containing information A, The number of cases containing information B; the correlation degree R ranges from 0 to 1, and the closer the value is to 1, the higher the clinical correlation between the two pieces of information. Information matching rules are established according to the correlation degree from high to low.

[0046] By combining medical natural language expression standards, doctor dialogue templates are established for different question types (such as "Where do you mainly feel unwell?" "How often do you experience abdominal pain?"), and patient response templates are established for different diagnosis and treatment information (such as "My stomach often bloates and hurts, about 3 times a week" "I get diarrhea when I eat spicy food"). The doctor dialogue templates, patient response templates, and information matching rules sorted by relevance are integrated to form an IBS disease doctor-patient dialogue generation rule library.

[0047] S4: Simulation and generation of doctor-patient dialogue data for IBS symptoms

[0048] Based on the constructed rule base, and combined with natural language generation technology, doctor-patient dialogue data is simulated and generated. The process is as follows:

[0049] Random sampling formula from structured key diagnostic and treatment information Randomly select or select treatment information units based on needs, among which This refers to the total number of structured diagnostic and treatment information units. For the first The probability of randomly selecting each information unit ensures that the selected information units cover the core dimensions of IBS diagnosis and treatment.

[0050] Based on the extracted diagnosis and treatment information units, the corresponding doctor's script templates and patient response templates in the rule base are matched, following the matching rules from high to low information relevance to ensure the clinical logical coherence of the dialogue.

[0051] Natural language generation techniques based on rules, templates, or pre-trained medical language models are employed to instantiate and populate templates, generating multi-turn doctor-patient dialogue texts for IBS symptoms; simultaneously, a perplexity formula is used. The generated dialogue text is optimized for fluency, among which... The total number of words in the dialogue text. For vocabulary In context The generation probability is adjusted by modifying model parameters to reduce perplexity and improve text fluency; after generation, the information relationship between the dialogue text and the original case is preserved to ensure data traceability.

[0052] S5: Dialogue Data Quality Assessment

[0053] To ensure the quality of the generated dialogue data, a quantitative quality assessment index system was constructed, and a comprehensive quality assessment was conducted through a combination of manual scoring and automatic machine calculation.

[0054] The evaluation indicator system is constructed, which includes five primary indicators: information completeness, clinical logic rationality, language fluency, dialogue relevance, and terminology accuracy. Each primary indicator has 3 to 5 secondary indicators. For example, information completeness has secondary indicators such as "coverage of core symptoms" and "mention of examination results", and clinical logic rationality has secondary indicators such as "information relevance" and "compliance of consultation process".

[0055] Quantitative scoring calculation: After scoring each secondary indicator, the scores are summarized into the primary indicator and calculated according to the comprehensive scoring formula. Calculate the overall score of the dialogue data, where The total number of primary indicators (in this embodiment) =5), For the first The weight coefficients of the indicators and satisfy the following conditions: , For the first The scores for each indicator. The weights of each primary indicator are allocated according to W1:W2:W3:W4:W5=0.25:0.30:0.15:0.15:0.15, highlighting the core importance of information completeness and clinical logical rationality;

[0056] Data filtering and annotation: The preset comprehensive score qualification threshold is 70~85 points. Dialogue data with scores below this threshold are removed. The qualified dialogue data are standardized and annotated (such as annotating dialogue rounds, diagnosis and treatment information types, IBS subtypes, etc.) to obtain a standardized IBS disease doctor-patient dialogue dataset.

[0057] S6: Iterative updates to the dialogue generation rule base

[0058] To achieve continuous optimization of the quality of generated data, an iterative update mechanism for the rule base based on quantitative calculation is established:

[0059] Collect the quality assessment results of the generated dialogue data, extract the score data of each indicator, and optimize them according to the priority formula. The optimization priority of each module (doctor's script template, patient's response template, and information matching rules) in the calculation rule base is determined. This represents the actual score of the evaluation indicator corresponding to the module. The full score for the indicator is 100 points. The threshold for the indicator to pass is... This represents the correlation between this module and the core IBS diagnosis and treatment. A higher Q value indicates a higher optimization priority for this module.

[0060] Based on the optimization priority from high to low, and combined with the professional optimization suggestions from clinical medical staff, the script templates and response logic in the rule base were revised and supplemented; at the same time, the newly added historical case data of IBS were processed according to steps 1-2 and updated to the structured diagnosis and treatment information database, and the correlation value in the information matching rules was corrected.

[0061] After the update is completed, the dialogue data is regenerated and its quality is evaluated to verify the update effect, forming a closed-loop iterative mechanism of "generation-evaluation-update-regeneration" to continuously improve the quality and diversity of the generated dialogue data.

[0062] Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0063] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0064] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for generating doctor-patient dialogue data based on IBS symptoms, characterized in that, Includes the following steps: S1. Construction of effective case dataset: Obtain historical case data related to IBS stored in the hospital, filter the original case data to remove invalid cases, and after the effectiveness rate is quantified and calculated, obtain an effective case dataset with an effectiveness rate that meets the preset requirements. S2. Natural Language Processing Preprocessing: Perform full-process natural language processing on the effective case dataset to extract and transform unstructured text information into structured information, and obtain structured key diagnostic and treatment information that can be used for subsequent rule construction; S3. Construction of Dialogue Generation Rule Base: Based on the structured key diagnosis and treatment information, the doctor-patient consultation process is sorted out in combination with the clinical diagnosis and treatment logic of IBS, the correlation pattern of diagnosis and treatment information is explored and standardized dialogue templates and information matching rules are established, and integrated to form a doctor-patient dialogue generation rule base exclusive to IBS. S4. Simulation and generation of doctor-patient dialogue data: Based on the dialogue generation rule base, select the diagnosis and treatment information unit to match the corresponding template, and use natural language generation technology to instantiate the template to generate doctor-patient dialogue text. Optimize the fluency of the text and establish an information association and traceability relationship with the original case. S5. Dialogue Data Quality Assessment: Construct a multi-dimensional doctor-patient dialogue data quality assessment system, and use a combination of human and machine scoring methods to comprehensively and quantitatively evaluate the generated dialogue data. After removing unqualified data and standardizing and labeling qualified data, a standardized IBS disease doctor-patient dialogue dataset is obtained. S6. Rule base iterative update: Based on the dialogue data quality assessment results and the newly added historical case data of IBS symptoms, the optimization priority of each module of the dialogue generation rule base is quantitatively determined, and the rule base is modified and supplemented with information according to the priority to achieve iterative optimization and performance improvement of the rule base.

2. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: In S1, the formula for calculating efficiency is: ； in For the number of effective cases, This represents the total number of original cases.

3. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: The historical case data related to IBS symptoms includes at least one of the following: patient basic information, IBS symptom description, medical history collection record, physical examination results, auxiliary examination data, diagnostic conclusion, and treatment recommendations. The data screening process involves removing case data that lacks key diagnostic information, contains logical contradictions, or has abnormal formatting. The preset threshold is 80% to 95%. When the calculated case effectiveness rate is lower than this threshold, supplementary historical case data of the same type of IBS is collected and re-screened.

4. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: The natural language processing preprocessing in S2 includes text cleaning, word segmentation, part-of-speech tagging, named entity recognition, keyword extraction, and text structure mapping. The keyword extraction uses the TF-IDF formula to specifically extract IBS-specific diagnostic and treatment terms. The TF-IDF formula is: ； in, For the term In cases Number of times it appears in For cases Total number of occurrences of all terms in the text Total number of cases For included terms The number of cases, the text structuring transforms unstructured case text into structured data in key-value pair form.

5. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: The construction process of the IBS patient-doctor dialogue generation rule base includes: based on structured key diagnostic and treatment information, sorting out the typical process of clinical patient-doctor consultation for IBS; extracting the doctor's question types, patient's response logic, and the correlation of diagnostic and treatment information in the consultation process, and applying the correlation formula... Calculate the correlation between medical information, whereby... This represents the number of cases where information A and information B both appear. For the number of cases containing information A, The number of cases containing information B; combined with natural language expression standards, establish doctor's script templates, patient response templates, and information matching rules, and integrate them to form a rule base.

6. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: The natural language generation technology includes at least one of rule-based generation technology, template-based generation technology, and pre-trained language model-based generation technology. The pre-trained language model is a Chinese language model adapted to medical text, and the fluency of the generated dialogue text is optimized using a perplexity formula.

7. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: The process of simulating and generating doctor-patient dialogue data in S3 is as follows: Data is randomly extracted from structured key diagnostic and treatment information using a formula. Randomly or selectively select diagnostic and treatment information units, among which This refers to the total number of structured diagnostic and treatment information units. For the first The probability of extracting each information unit; matching doctor's script templates and patient response templates in the dialogue generation rule base; instantiating and filling the templates using natural language generation technology to generate multi-round, clinically logical IBS symptom doctor-patient dialogue texts, while preserving the information association between the dialogue and the original case.

8. The method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: It also includes the step of quality assessment of the generated IBS patient-doctor dialogue data: constructing a quality assessment index system for patient-doctor dialogue data, scoring the dialogue data through a combination of manual scoring and automatic machine calculation, and applying a comprehensive scoring formula. Calculate the overall score of the dialogue data, where To evaluate the total number of indicators, For the first The weight coefficients of the indicators and satisfy the following conditions: , For the first The scores of each indicator are calculated; dialogue data with scores below a preset threshold are removed, and qualified dialogue data are labeled to obtain a standardized IBS disease doctor-patient dialogue dataset.

9. A method for generating doctor-patient dialogue data based on IBS symptoms according to claim 8, characterized in that: The quality assessment index system includes five primary indicators: information completeness, clinical logic rationality, language fluency, dialogue fit, and terminology accuracy. Each primary indicator has 3 to 5 secondary indicators. The weight coefficients of each primary indicator are allocated according to W1:W2:W3:W4:W5=0.25:0.30:0.15:0.15:0.

15. The comprehensive score adopts a percentage system, and the preset passing threshold is 70 to 85 points.

10. A method for generating doctor-patient dialogue data based on IBS symptoms according to claim 1, characterized in that: It also includes the step of iteratively updating the IBS patient-doctor dialogue generation rule base: collecting the quality assessment results of the generated dialogue data, calculating the optimization priority of each rule base module, and combining the optimization priority with newly added historical IBS case data to supplement and correct the dialogue templates, response logic and information matching rules in the rule base.

Images