Traditional Chinese medicine menstrual prescription clinical teaching evaluation method and system based on information entropy, and medium

Abstract

The invention discloses a traditional Chinese medicine menstrual prescription clinical teaching evaluation method based on information entropy. The method comprises the following steps: constructing a conditional probability knowledge base; initializing a diagnosis space and a feature space, determining each diagnosis feature in the diagnosis space and each symptom problem distribution initial prior probability in the feature space based on a conditional probability knowledge base, and calculating to obtain an initial information entropy; acquiring an input symptom feature, calling a corresponding conditional probability according to a conditional probability knowledge base, executing Bayesian updating, and calculating a diagnostic posterior probability; after each Bayesian update, calculating and diagnosing spatial information entropy, and determining entropy decrement according to the initial information entropy; recording the diagnosis probability distribution and entropy decrement; and traversing the candidate symptom features which are not input, calculating expected information gains of the candidate symptom features, and outputting recommendation questions according to the expected information gains. According to the invention, objective quantitative evaluation of student thinking paths is realized, and dynamic optimization and guidance of inquiry strategies are realized through information gain calculation.

Description

Technical Field

[0001] This invention relates to the field of clinical teaching technology in traditional Chinese medicine, and in particular to a clinical teaching evaluation method, system, and medium for classical Chinese medicine formulas based on information entropy. Background Technology

[0002] In TCM clinical teaching, such as the teaching of syndrome differentiation and treatment centered on the *Treatise on Febrile and Miscellaneous Diseases*, traditionally it has relied on apprenticeship-style instruction or classroom case discussions, emphasizing the accumulation of experience. However, the teaching effectiveness often depends on the individual level of the teacher, and the evaluation lacks a unified and objective quantitative standard. For example, different teachers may give different evaluations of the same student's questioning process, resulting in unstable and non-reproducible results.

[0003] With the development of information technology, some computer-aided instruction (CAI) systems and online examination software have been applied to TCM teaching. These systems typically present answers in the form of case study question banks or multiple-choice questions. After students answer, the system compares their answers and provides feedback on whether they are correct or not. However, this interactive path is rigid, unable to simulate the openness of real clinical consultation, and even less able to conduct process-oriented analysis of students' thinking processes. In other words, its assessment can only judge whether the result is right or wrong, but cannot reflect the quality of the reasoning path.

[0004] In the field of artificial intelligence, some expert systems or assisted diagnostic systems can output diagnostic conclusions through knowledge graphs or rule chains to improve the diagnostic efficiency of clinicians, rather than serving teaching scenarios. Due to the complexity of the reasoning process and the lack of visual representation, it often exhibits a black-box effect in teaching, failing to help students understand diagnostic logic and lacking assessment and guidance of students' thinking paths. Existing teaching methods rely on subjective experience for assessment, lack objective quantitative standards, remain at the level of result comparison, struggle to analyze the thought process, and the reasoning is unexplainable and lacks real-time guidance, thus failing to meet the needs of objectively assessing and intelligently teaching the clinical reasoning process of TCM students. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method, system, and medium for clinical teaching evaluation of TCM classic formulas based on information entropy. Based on a conditional probability knowledge base, and combined with Bayesian updates and information entropy calculation, it achieves an objective quantitative evaluation of students' reasoning paths and dynamically optimizes and guides consultation strategies through information gain calculation.

[0006] To address the aforementioned technical problems, the first aspect of this invention discloses a clinical teaching evaluation method for TCM classic formulas based on information entropy, the method comprising: A conditional probability knowledge base is constructed based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonym entries into standard symptom identifiers, and the diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts. Initialize the diagnostic space and feature space, and determine the initial prior probability for each diagnostic feature in the diagnostic space and each symptom question in the feature space based on the conditional probability knowledge base, and calculate the initial information entropy; The system acquires the input symptom features, calls the corresponding conditional probabilities from the conditional probability knowledge base, performs a Bayesian update, and calculates the posterior probability of diagnosis. After each Bayesian update, the diagnostic space information entropy is calculated, and the entropy reduction is determined based on the initial information entropy. The entropy reduction is then bound to the input symptom features to form a consultation path. The diagnostic probability distribution and entropy reduction are recorded, and the diagnostic distribution and entropy value changes are output. Iterate through the candidate symptom features that have not been input, calculate the expected information gain of the candidate symptom features, and output the recommendation question based on the expected information gain.

[0007] In some implementations, a conditional probability knowledge base is constructed, including: Obtain the diagnostic mapping rule table and the symptom dictionary, wherein the symptom dictionary normalizes the synonyms or variant expressions in ancient books into unified symptom identifiers and assigns a unique index to each symptom identifier; The texts of ancient Chinese medicine books are analyzed, and when diagnostic keywords and symptom keywords co-occur in the same text or in adjacent contexts, a correspondence between diagnostic markers and symptom markers is established; the ancient Chinese medicine books mentioned include at least the "Treatise on Febrile and Miscellaneous Diseases"; Determine the initial prior probability of the co-occurrence scenario; The diagnostic identifier, symptom identifier, and corresponding probability value are stored in the form of triples to form the conditional probability knowledge base.

[0008] In some implementations, determining the initial prior probability of the co-occurrence scenario includes: The passages in ancient Chinese medicine books are classified according to their importance to diagnosis, into core definition passages, formula-symptom correspondence passages, differential diagnosis or concurrent symptoms passages, and prognosis or contraindication passages. Symptoms appearing in the core definition clauses are set as the first probability value, symptoms appearing in the corresponding clauses of the prescription are set as the second probability value, symptoms appearing in the identification or concurrent syndrome clauses are set as the third probability value, symptoms appearing in the prognosis or contraindication clauses are set as the fourth probability value, and background probability values ​​are assigned to symptoms that do not appear. A score is obtained to indicate the correlation between diagnosis and symptoms. This score is then converted into a correction factor. When the score is higher than a preset baseline value, the corresponding conditional probability is increased. When the score is lower than the baseline value, the corresponding conditional probability is decreased.

[0009] In some implementations, the entropy reduction is bound to the input symptom features to form a consultation path, including: A consultation record is generated based on the symptom characteristics input each time. The consultation record includes symptom identifier, information entropy before update, information entropy after update, and entropy reduction. Multiple consultation records are linked sequentially to form a consultation path linked list.

[0010] In some implementations, calculating the expected information gain of the candidate symptom features includes: The probability of occurrence and non-occurrence of candidate symptom features are calculated based on a conditional probability knowledge base. The diagnostic space of occurrence and non-occurrence probabilities is updated by Bayes' formula, the posterior probability distribution of occurrence and non-occurrence probabilities is obtained, and the corresponding information entropy is calculated. The expected entropy is calculated by weighted averaging the two information entropies. By comparing the current information entropy with the expected entropy, the information gain value corresponding to the candidate symptom is obtained; The desired information gain is obtained by weighting and correcting the information gain by combining the priority field in the diagnostic mapping rule table.

[0011] In some implementations, the information entropy is a measure of the uncertainty of the diagnostic probability distribution, the entropy reduction is the difference between the information entropy before and after two consecutive inputs, and the information gain is the difference between the current information entropy and the expected information entropy corresponding to the candidate symptom.

[0012] In some implementations, the recommendation problem is a candidate problem that maximizes the expected information gain, determined by combining the priority weights in the diagnostic mapping table.

[0013] In some implementations, it also includes: The interactive interface displays the diagnostic probability distribution and outputs the diagnosis with the highest probability and its corresponding probability value, information entropy, and recommended input symptom features calculated based on information gain in a visual form. Upon receiving a diagnosis submission instruction, an evaluation report is generated, which includes the input symptom characteristics, a consultation path list, the most probable diagnosis, and a recommended optimal path.

[0014] Secondly, this invention discloses a clinical teaching evaluation system for classical Chinese medicine formulas based on information entropy, comprising: The knowledge base construction module constructs a conditional probability knowledge base, which is based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonym entries into standard symptom identifiers, and the diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts. The interactive assessment module acquires the input symptom characteristics, calls the corresponding conditional probabilities based on the conditional probability knowledge base, performs Bayesian updates, and calculates the probability of the diagnostic outcome. After each Bayesian update, the diagnostic space information entropy is calculated, and the entropy reduction is determined based on the initial information entropy. The entropy reduction is then bound to the input symptom features to form a consultation path. The diagnostic probability distribution and entropy reduction are recorded, and the diagnostic distribution and entropy value changes are output. The interactive display module iterates through the candidate symptom features that have not been input, calculates the expected information gain of the candidate symptom features, and outputs a recommendation question based on the expected information gain.

[0015] Thirdly, a computer storage medium is disclosed, on which a computer program is stored, which, when executed by a processor, implements the information entropy-based clinical teaching evaluation method for TCM classic prescriptions as described in any of the above.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs a conditional probability knowledge base and introduces information entropy as a measure of uncertainty in the diagnostic and feature spaces. After inputting symptom features, it dynamically adjusts the diagnostic probability distribution through Bayesian updates, binding entropy reduction to the input symptom features to form a traceable consultation path. This invention calculates the change in information entropy after each acquisition of input symptom features, achieving dynamic quantification of the diagnostic probability distribution as an objective process evaluation indicator. Simultaneously, by calculating the expected information gain for outputting recommendation questions for uninputted symptom features, it provides real-time and objective evaluation and guidance of students' reasoning paths in simulated consultations, achieving precise modeling and quantitative analysis of students' reasoning paths in the context of TCM formula teaching evaluation. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention. Figure 2 This is a flowchart illustrating step S1 in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention. Figure 3 This is a schematic diagram of the symptom dictionary database in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention; Figure 4 This is an intentional representation of the diagnostic mapping rules in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention; Figure 5 This is a flowchart illustrating step S13 in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention. Figure 6 This is a flowchart illustrating step S5 in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention. Figure 7 This is a schematic diagram of the interactive interface in the clinical teaching evaluation method for TCM classic prescriptions based on information entropy provided by the present invention. Figure 8 This is a schematic diagram of the evaluation report in the clinical teaching evaluation method of TCM classic prescriptions based on information entropy provided by the present invention; Figure 9 This is a schematic diagram of the evaluation report in the clinical teaching evaluation method of TCM classic prescriptions based on information entropy provided by the present invention; Figure 10 This is a schematic diagram of the evaluation report in the clinical teaching evaluation method of TCM classic prescriptions based on information entropy provided by the present invention; Figure 11 This is a schematic diagram of the evaluation report in the clinical teaching evaluation method of TCM classic prescriptions based on information entropy provided by the present invention. Detailed Implementation

[0018] To better understand and implement this invention, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] The terms “comprising” and “having” and any variations thereof in this invention are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products or devices.

[0020] This invention discloses a clinical teaching evaluation method for TCM classic formulas based on information entropy. Using a conditional probability knowledge base as a foundation, and combining Bayesian updating and information entropy calculation, it achieves an objective quantitative evaluation of students' reasoning pathways and dynamically optimizes and guides questioning strategies through information gain calculation. In this invention, entropy is transformed for the first time into a quantitative indicator of students' reasoning efficiency. Each symptom question corresponds to a change in entropy value, and the entropy reduction ΔH becomes a teaching signal used to evaluate whether students have grasped the key symptoms.

[0021] like Figure 1 As shown, this method includes: Step S1: Construct a conditional probability knowledge base. The conditional probability knowledge base is constructed based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonym entries into standard symptom identifiers. The diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts.

[0022] The conditional probability knowledge base is used to describe the conditional dependencies between diagnoses in the diagnosis space D and symptom questions in the feature space Q. The conditional probability P(q) j |d i ) indicates "when the diagnosis is determined to be d" i At that time, symptoms q j "Possibility of occurrence". For example, Figure 2 As shown, the construction method includes the following steps: Step S11: Obtain the diagnostic mapping rule table and the symptom dictionary. The symptom dictionary normalizes the synonyms or variant expressions in ancient books into unified symptom identifiers and assigns a unique index to each symptom identifier. The original ancient texts are preprocessed. Based on representative ancient Chinese medicine texts such as the *Treatise on Febrile and Miscellaneous Diseases*, the entire text is digitized, segmented, denoised, punctuated, and formatted to form a standardized corpus. These steps ensure the stability and consistency of subsequent retrieval and mapping operations.

[0023] A symptom dictionary database is established to unify the various expressions found in ancient texts. In this implementation, such as... Figure 3 As shown, the symptom dictionary uses symptom IDs as indexes, with each symptom corresponding to a standardized symptom name, and includes synonyms or variant expressions that may appear in ancient Chinese medical texts. For example, symptom Q001 is defined as "fever," and its synonyms include "spontaneous fever," "body heat," "evaporative fever," "mild fever," and "fever," etc.; symptom Q002 is defined as "chills / aversion to wind," and its synonyms include "chills and aversion to wind," "fear of cold," and "aversion to cold," etc. In this way, the complex and diverse natural language expressions in the text can be uniformly mapped to unique symptom identifiers, solving the problem of multiple expressions for the same diagnosis in ancient texts.

[0024] When certain symptoms play a crucial role in the diagnostic process, they are annotated in the symptom dictionary. For example, symptom Q003 "sweating" is marked as "key to differentiating stroke / typhoid fever," and symptom Q004 "anhidrosis" is marked as "core of Ma Huang Tang syndrome." These annotations with diagnostic hints allow subsequent probability assignments and diagnostic reasoning to incorporate the unique clinical logic of Traditional Chinese Medicine. The symptom dictionary forms a data table structured as [symptom ID, standard symptom name, synonym set, annotations], providing a unified input interface for subsequent diagnostic reasoning.

[0025] like Figure 4 As shown, the diagnostic mapping rule table is used to define diagnostic identifiers and their associated contextual information. Each diagnostic entry includes a diagnostic criterion name, core keywords, priority, necessary context, and exclusion context. The diagnostic mapping rule table is indexed by the diagnostic ID. The necessary context in the diagnostic mapping rule table includes a predefined set of diagnostic keywords. The set of diagnostic keywords includes at least one syndrome keyword and at least one symptom keyword. The exclusion context includes a set of exclusion keywords.

[0026] For example, D001 is defined as "Taiyang stroke (Guizhi Tang syndrome)," with the core keywords "Taiyang stroke" and "Guizhi Tang," a priority of 1, and the necessary context being "Taiyang." Exclusionary contexts include "postpartum," "spasm," "illness," and "Ephedra." In other words, when both "Taiyang" and the core keywords appear in the text, the probability of this diagnosis is high; however, if the keywords in the exclusionary context also appear, the probability of this diagnosis will be significantly reduced. For example, D002 is defined as "Taiyang cold damage (Ephedra Tang syndrome)," with the necessary contexts being "Taiyang," "no sweating," and "body pain," and the exclusionary contexts being "sweating" and "Guizhi." By using necessary and exclusionary contexts, "Taiyang stroke" and "Taiyang cold damage" can be effectively distinguished, avoiding confusion between different syndromes due to similar symptom descriptions.

[0027] The priority field is used to adjust the weight of different diagnoses when multiple diagnoses coexist. For example, the priority of classical formula syndromes in Taiyang disease (such as Guizhi Tang syndrome and Mahuang Tang syndrome) is set to 1, while the priority of certain concurrent syndromes and prognostic syndromes is set to 2 or 3. In this way, higher priority diagnoses are given greater initial weight in probability calculations, making the results more consistent with clinical logic.

[0028] The diagnostic mapping rule table ultimately forms a data table with the structure [Diagnosis ID, Diagnosis Name, Keyword, Priority, Necessary Context, Exclusion Context]. This data table, together with the symptom dictionary, forms the foundation of the conditional probability knowledge base, supporting subsequent Bayesian updates, entropy reduction calculations, and optimal question recommendation.

[0029] Step S12: Analyze the texts of ancient Chinese medicine books. When diagnostic keywords and symptom keywords co-occur in the same text or adjacent context, establish a correspondence between diagnostic markers and symptom markers. The ancient Chinese medicine books mentioned include at least the "Treatise on Febrile and Miscellaneous Diseases". Analyze the provisions in traditional Chinese medicine ancient books, perform context association and probability scoring. Traverse each diagnosis and its keywords in the diagnostic mapping rule table. In ancient books, diagnoses and keywords are usually accompanied by words such as "maintains", "ones", "disease", etc. Retrieve the core provisions containing these keywords in the corpus. In the retrieved core provisions and their adjacent contexts, detect whether the expressions of each symptom in the feature space Q appear. When the diagnostic keyword and the symptom keyword co-occur in the same provision or adjacent context, establish the corresponding relationship between the diagnostic identifier and the symptom identifier according to the association strength between the symptom and the diagnostic keyword.

[0030] Step S13. Determine the initial prior probability of the co-occurrence scenario; as Figure 5 shown, including: Step S131. Classify the provisions in traditional Chinese medicine ancient books. According to the importance of the provisions for diagnosis, classify the provisions into core definition provisions, formula-syndrome corresponding provisions, differential diagnosis or concurrent syndrome provisions, and prognosis or taboo provisions.

[0031] Classify the provisions in traditional Chinese medicine ancient books such as "Treatise on Febrile and Miscellaneous Diseases". According to the importance of the provisions for a certain diagnosis, classify the provisions into four levels: L1 Core definition provisions: Directly give the main characteristics of the disease syndrome, such as "The disease of Taiyang is characterized by floating pulse, stiffness and pain in the head and neck, and aversion to cold", belonging to a very strong association; L2 Formula-syndrome corresponding provisions: Clearly indicate the applicable syndromes of a certain formula, such as "For patients with Taiyang disease, headache, fever, sweating, and aversion to wind, Cinnamon Twig Decoction is the main treatment", belonging to a strong association; L3 Differential diagnosis or concurrent syndrome provisions: Used for syndrome type differentiation or concurrent syndrome description, such as "For patients with asthma, when using Cinnamon Twig Decoction, adding Magnolia officinalis and Armeniaca vulgaris is better", belonging to a medium association; L4 Prognosis or taboo provisions: Involve the outcome or taboo situation, such as "Those with dry throat should not induce sweating", belonging to a weak association.

[0032] Step S132. Set the symptoms that appear in the core definition provisions as the first probability value, the symptoms that appear in the formula-syndrome corresponding provisions as the second probability value, the symptoms that appear in the differential diagnosis or concurrent syndrome provisions as the third probability value, the symptoms that appear in the prognosis or taboo provisions as the fourth probability value, and assign the background probability value to the symptoms that do not appear; Exemplarily, the first probability value is 0.95; the second probability value is 0.85; the third probability value is 0.50; the fourth probability value is 0.20; the background probability value is 0.05. The specific probability values can be adjusted according to the actual situation and are not limited in this application.

[0033] Step S133: Obtain the score of the correlation between diagnosis and symptoms, and convert the score of the correlation between diagnosis and symptoms into a correction factor. When the score is higher than the preset benchmark value, increase the corresponding conditional probability; when it is lower than the benchmark value, decrease the corresponding conditional probability.

[0034] In some implementations, the score for the correlation between diagnosis and symptoms can be determined by multiple TCM clinical or classical formula research experts with extensive TCM knowledge. Each group of diagnosis-symptom correlations is scored independently, with a score range of 1-10. The scoring criteria are: 10 points for core symptoms, 7-9 points for important symptoms, 4-6 points for common or likely symptoms, and 1-3 points for minor or rare symptoms. To reduce the interference of individual extreme values, the highest and lowest scores are removed, and the average is taken as the final score for the correlation between diagnosis and symptoms.

[0035] The scores reflecting the correlation between diagnosis and symptoms are converted into correction factors to adjust the initial probability values ​​obtained from the categorization. The correction principle is as follows: when the score is higher than the baseline value, the conditional probability increases accordingly; when the score is lower than the baseline value, the conditional probability decreases accordingly. For example, a score of 9.0 increases the conditional probability by approximately 17%; a score of 2.0 decreases the conditional probability by approximately 17%.

[0036] The final conditional probability can be determined by combining the categorical probability with an expert correction factor. To avoid zero or absolute probability, the conditional probability is limited to [0.01, 0.99]. For example, for the symptom "sweating" under the diagnosis of "Taiyang stroke," the corresponding categorical provision is "Guizhi Tang is the main treatment," the categorical level is L2, the initial probability is 0.85, the score is 9.5, the conversion correction factor is 1.20, the corrected probability value is 1.02, and the final probability obtained after interval limitation is 0.99.

[0037] By combining the classification of provisions with the scoring used for correction, conditional probabilities are based on classic provisions while incorporating the experience and judgment of clinical experts, providing more reliable data support for subsequent diagnostic probability updates, information entropy calculations, and optimal question recommendations.

[0038] Furthermore, to improve the accuracy of the initial prior probability, it can be modified by combining the necessary and exclusion contexts in the diagnostic mapping rule table. For example, the necessary contexts for diagnosing "Taiyang typhoid fever" are "Taiyang," "no sweating," and "body pain," while the exclusion contexts are "sweating" and "cinnamon twig." When the input symptom co-occurs with the necessary context keywords, the conditional probability increases; when the input symptom co-occurs with the exclusion context keywords, the conditional probability decreases to the background level.

[0039] Step S14: Store the diagnostic identifier, symptom identifier and corresponding probability value in the form of triples to form the conditional probability knowledge base.

[0040] All calculated conditional probability results are presented in triplet form [diagnostic identifier d] i Symptom marker q j The conditional probability values ​​are stored to form a Conditional Probability Table (CPT). This knowledge base supports subsequent updates of diagnostic probabilities, calculation of information entropy, and analysis of information gain.

[0041] Step S2: Initialize the diagnostic space and feature space, and determine the initial prior probability for each diagnostic feature in the diagnostic space and each symptom question in the feature space based on the conditional probability knowledge base, and calculate the initial information entropy; Since this application is applied in the teaching scenario of TCM classic formulas, it uses ancient books such as "Treatise on Febrile and Miscellaneous Diseases" as knowledge sources to construct a diagnostic space D and a feature space Q, providing a basis for the quantitative assessment of students' inquiry-based learning and critical thinking process.

[0042] Based on the diagnostic mapping rule table in the conditional probability knowledge base, a diagnostic space D is constructed. The diagnostic space D = {d1, d2, ..., dn}, where each diagnostic d1, d2, ..., dn is represented by a rule table. i All are derived from the formulas and disease patterns in the *Treatise on Febrile and Miscellaneous Diseases*, and stored in a standardized format. For example, d1 = "Taiyang Wind-Stroke (Guizhi Tang Syndrome)", d2 = "Taiyang Cold Damage (Mahuang Tang Syndrome)", d3 = "Shaoyang Disease (Xiao Chaihu Tang Syndrome)", d4 = "Taiyin Disease (Lizhong Tang Syndrome)", d5 = "Yangming Bowel Stagnation Syndrome (Chengqi Tang Syndrome)". This method ensures that the diagnostic space fully covers the main classical formula patterns that need to be identified in teaching.

[0043] Based on the symptom dictionary in the conditional probability knowledge base, a feature space Q is constructed. The feature space Q = {q1, q2, ..., qm}, where each feature q... j The atomized issues originate from the symptoms, signs, or pulse characteristics in the original text and have undergone unified normalization. For example, q1=Q001: "Is there a fever?", q2=Q002: "Is there aversion to cold / wind?", q3=Q003: "Is there sweating?", q4=Q004: "Is there no sweating?", q5=Q005: "Is the pulse floating?". Synonymous or variant expressions are uniformly mapped to the same standard identifier in the symptom dictionary; for example, "sweating", "spontaneous sweating", and "night sweats" are all mapped to q3.

[0044] After establishing the diagnostic space D and the feature space Q, the [diagnostic identifier d] in the conditional probability knowledge base is called. i Symptom marker q j [Conditional probability value] triple, for each diagnosis d i With each symptom question q jAssign initial prior probabilities. Under no additional teaching conditions, prior probabilities can be set equally; however, if combined with the priority field in the diagnostic mapping rule table, diagnoses with higher priority receive higher initial probability weights. Calculate the initial information entropy H0(D) of the diagnostic space based on the prior probabilities. This value serves as the baseline uncertainty for the teaching diagnostic reasoning process, measuring the impact of subsequent inputs on the convergence of the diagnostic space. Through these steps, the initialization of the diagnostic space and feature space is completed, and it is linked with the conditional probability knowledge base in step S1, providing a calculable and traceable reasoning environment for students' critical thinking process in the teaching scenario.

[0045] Step S3: Obtain the input symptom features, call the corresponding conditional probabilities according to the conditional probability knowledge base, perform Bayesian update, and calculate the diagnostic probability distribution.

[0046] When students enter the teaching simulation consultation session, a dynamic update process is executed based on the conditional probability knowledge base and the initialized diagnostic space constructed in steps S1 and S2, so as to realize the real-time adjustment of the diagnostic probability distribution.

[0047] First, at the start of a new medical case, load the complete diagnostic space D={d1,d2,…,d n Before any symptom information is entered, it is assumed that all diagnostic conclusions have equal prior probabilities, i.e., for each diagnosis d... i Assign an initial prior probability P0(d) i )=1 / n. At this point, the uncertainty of the diagnostic space is at its maximum, and the information entropy H0(D)=log2(n) is used as the initial benchmark for the reasoning process.

[0048] Subsequently, the student selects a symptom question q from the feature space Q. ⱼ Ask a question and return a "yes" or "no" answer based on preset medical case data. Call the corresponding conditional probability P(q) from the conditional probability knowledge base. j |d i And perform a Bayesian update based on the response results: If the medical record data returns a "yes" answer, then the conditional probability P(q) is directly called. j )|d i ); If the medical record data returns a "no" answer, then the supplementary probability 1-P(q) is called. j )|d i ).

[0049] During the Bayesian update process, each diagnosis d i The conditional probability Pold(d) i The result will be corrected to the posterior probability Pnew(d). i The calculation formula is as follows: Pnew(d i )=[P(q j )|d i )×Pold(d i )] / P(q j )) Where P(q) j ) is a normalization constant used to ensure that the sum of the probabilities of all diagnoses is 1.

[0050] The posterior probability of a diagnosis that matches the question-and-answer results will increase, while the posterior probability of a diagnosis that does not match will decrease, and the probability distribution in the diagnostic space will gradually converge towards the optimal diagnostic result. Simultaneously, the updated diagnostic probability distribution is recorded. In this way, this embodiment achieves dynamic Bayesian updating based on input symptom features in a teaching simulation consultation scenario, making the diagnostic reasoning process computable and traceable.

[0051] Step S4: After each Bayesian update, calculate the diagnostic space information entropy, determine the entropy reduction based on the initial information entropy, bind the entropy reduction with the input symptom features to form a consultation path; record the diagnostic probability distribution and the entropy reduction, and output the diagnostic distribution and entropy value changes.

[0052] After each Bayesian update, information entropy calculation is introduced to dynamically evaluate the effectiveness of the consultation process. Based on the posterior probability Pnew(dᵢ), the updated diagnostic space information entropy Hnew(D) is calculated. Information entropy, as a measure of the uncertainty of the diagnostic space, reflects the degree of ambiguity of the student's diagnostic conclusion under the current question. It is calculated by weighted summation of all diagnostic probabilities in the diagnostic space and then taking the logarithm.

[0053] By comparing the updated entropy value with the previous entropy value, we obtain the entropy reduction ΔH = Hold(D) – Hnew(D). This value quantifies the degree of uncertainty reduction brought about by the current question. If ΔH is large, it indicates that the question reduces the uncertainty of the diagnostic space, making it an efficient question; if ΔH is small, it means that the question provides limited information and has a weak effect on the convergence of the diagnostic space. In this way, the value of the student's critical thinking process is objectively characterized through a numerical indicator.

[0054] To facilitate subsequent analysis and review, each input is linked to its corresponding entropy reduction, generating a consultation record. Each consultation record contains at least the following fields: symptom identifier, information entropy before update, information entropy after update, and entropy reduction ΔH. Multiple consultation records are linked sequentially to form a complete consultation path linked list, preserving the dynamic trajectory of the diagnostic probability evolution and recording the information acquisition efficiency of students at each step of the reasoning process during teaching.

[0055] Traditional Chinese medicine (TCM) diagnosis differs from modern classification logic, exhibiting complex relationships such as different symptoms for the same disease and the same symptoms for different diseases. Traditional statistical or classification methods struggle to handle this complex interplay. This invention utilizes information entropy to accurately represent the uncertainty between symptoms. For example, the same symptom of "fever" could indicate either Taiyang stroke or Taiyang typhoid fever; the entropy value reflects the undecided diagnostic state. Through continuous questioning, the entropy value decreases, corresponding to the gradual clarification of the diagnosis. This unique explanatory power in TCM teaching is irreplaceable by existing information entropy applications.

[0056] Unlike traditional techniques that only output diagnostic results or provide single probability updates, this embodiment introduces information entropy, a global metric, into the reasoning process. Through the binding mechanism of entropy reduction and symptom input, it transforms abstract diagnostic uncertainty into a traceable path structure. This provides teachers or the system with a basis for review, indicating which steps students ask questions efficiently and which steps are redundant. It also provides input for subsequent optimal question recommendations, making the teaching process quantitative, transparent, and interpretable.

[0057] Step S5: Traverse the candidate symptom features that have not been input, calculate the expected information gain of the candidate symptom features, and output the recommendation question based on the expected information gain.

[0058] After completing each diagnostic probability update and information entropy calculation, this embodiment further executes optimal path guidance to provide students with intelligent question suggestions. It iterates through all candidate symptom questions qk that have not yet been asked in the feature space Q. For each candidate question, its expected information gain IG(qk) is calculated to measure the discriminative value of the question in the current diagnostic state. Using the entropy value Hcurrent(D) of the current diagnostic space as a benchmark, the two possible answers (present or absent) to candidate question qk are deduced separately. Specifically, as follows... Figure 6 As shown, it includes the following steps: Step S51: Calculate the probability of occurrence and the probability of non-occurrence of candidate symptom features based on the conditional probability knowledge base; Step S52: Update the diagnostic space of occurrence probability and non-occurrence probability using Bayes' theorem, obtain the posterior probability distribution of occurrence probability and non-occurrence probability, and calculate the corresponding information entropy. Step S53: Calculate the expected entropy by weighted averaging of the two information entropies; Step S54: By comparing the current information entropy with the expected entropy, the information gain value corresponding to the candidate symptom is obtained; Step S55: Combine the priority field in the diagnostic mapping rule table to perform weighted correction and obtain the expected information gain.

[0059] Based on the conditional probability knowledge base, the probability of candidate symptom feature qk appearing and not appearing is first calculated. These probability values ​​are derived from the conditional probability table corresponding to diagnosis and symptom, and weighted in conjunction with the current diagnosis probability distribution. For the two possible responses to candidate symptom feature qk, a Bayesian update of the diagnosis space is performed respectively: In the case of "symptom qk appears", the diagnostic space is modified using the conditional probability knowledge base to obtain a new posterior probability distribution, and the corresponding information entropy H(D|qk=appears) is calculated. In the case where "symptom qk does not appear", a Bayesian update is performed to obtain another set of posterior probability distributions, and the corresponding information entropy H(D|qk=not appear) is calculated.

[0060] Based on the probability of occurrence and non-occurrence of candidate symptoms, the information entropy under the above two scenarios is weighted and averaged to obtain the expected entropy E[H(D|qk)] of candidate question qk. By comparing the information entropy H_current(D) of the current diagnostic space with the expected entropy E[H(D|qk)], the original information gain value of the candidate question is obtained, reflecting the ability of candidate question qk to reduce the uncertainty of the diagnostic space in the current state. Combining the priority field in the diagnostic mapping rule table, the original information gain value is weighted and corrected to obtain the final expected information gain. Diagnoses with higher priority will receive greater weight in the correction process, thus making the recommended question more in line with the logic of traditional Chinese medicine diagnosis. After traversing all candidate questions, the IG values ​​of each question are compared, and the symptom question with the largest information gain is selected as the recommended question, which is output as a prompt in the teaching interface. For example, when the information gain of judging "whether there is sweating" is the largest, the prompt box will give "Recommended next question: Is there sweating?".

[0061] This embodiment calculates the expected information gain of unasked symptoms in real time and dynamically outputs the optimal questioning suggestions. This makes the teaching process not only guiding but also supports the recommendation results with objective information gain numerical indicators. It helps students learn how to grasp key information points in dialectical thinking and optimizes and interprets the diagnostic path, thereby achieving intelligent guidance for the reasoning path.

[0062] After guiding students to the optimal path and outputting recommended questions, this embodiment also includes an explanation and fault-tolerance mechanism to enhance the guidance and flexibility of teaching. After the recommended questions are generated, the interface will prompt students with the reasons for the recommended questions. Based on information gain calculation logic, the expected entropy reduction corresponding to the question and its role in the current diagnostic space will be displayed. For example, if the question "Are you sweating?" has the highest information gain value, it will prompt: "Asking 'Are you sweating?' can significantly narrow the diagnostic scope, making it clear that 'solar stroke' is different from 'solar typhoid fever'." Through this explanation, students see the recommended results, understand the underlying reasoning, and improve the transparency and persuasiveness of their learning.

[0063] Considering that students might reject recommended questions and insist on asking their own in a teaching setting, this embodiment employs a fault-tolerant strategy. When a student selects a non-recommended question, a Bayesian update and information entropy calculation are performed based on the input, and a new diagnostic probability distribution is generated. In subsequent recommendation steps, new optimal questions are output, but the student's failure to adopt the recommendation is marked in the evaluation log. This approach ensures flexibility in teaching, allowing students to explore different questioning paths.

[0064] Existing medical teaching assessments often only judge whether a student's final diagnosis is correct, ignoring the rationality of the thought process. This invention utilizes information entropy to transform teaching assessment from a single result of correctness / incorrectness to a comprehensive evaluation of the entire process. Even if a student's final diagnosis is correct, if many invalid questions with near-zero entropy reduction are asked in the process, the entropy trajectory will still point out the problem, and the score will decrease accordingly.

[0065] More importantly, the final evaluation report will review the student's thought process. If a student frequently chooses non-recommended questions, the overall entropy reduction in the diagnostic path will be low, potentially leading to an incorrect diagnosis and a lower score. Through this feedback mechanism, students can clearly see the difference between the optimal questioning path and their actual questioning path, thus understanding their shortcomings in the critical thinking process.

[0066] For example, in one teaching case, students did not ask the recommended questions but instead selected multiple symptoms with low information gain, resulting in a consistently dispersed diagnostic probability distribution and a final diagnosis that deviated from the optimal conclusion. This approach was marked as an inefficient questioning path in the evaluation report, and its corresponding score was lower than that of following the recommendations.

[0067] In specific teaching applications, it also includes: The interactive interface displays the diagnostic probability distribution and outputs the diagnosis with the highest probability and its corresponding probability value, information entropy, and recommended input symptom features calculated based on information gain in a visual form. Upon receiving a diagnosis submission instruction, an evaluation report is generated, which includes the input symptom characteristics, a consultation path list, the most probable diagnosis, and a recommended optimal path.

[0068] like Figure 7 , 8 As shown in Figures 9, 10, and 11, during the teaching process, the interactive interface can display the diagnostic probability distribution, information entropy, and recommended questions in real time. The left side of the interface displays the chief complaint of the medical record and records the student's consultation history in chronological order, forming a continuous question-and-answer list to help students clearly see the questions they have asked and the answers they have received. The right side of the interface dynamically presents the analysis results, including curves or bar graphs showing the changes in diagnostic probabilities, displaying the highest-probability diagnoses and their values ​​in real time. As students ask questions, the probabilities of some diagnoses increase while others decrease, allowing students to intuitively perceive the diagnostic space gradually converging.

[0069] Furthermore, the degree of uncertainty is displayed through dashboard-style visualizations. Each time a student completes a question and receives an answer, the information entropy value is updated, and the pointer on the dashboard moves accordingly. This makes the originally abstract change in entropy value visually corresponding, allowing students to more clearly understand the role of questioning in reducing uncertainty.

[0070] During the consultation, a prompt box can recommend the optimal question for the current situation. The recommended question is generated based on information gain calculations of unasked symptoms. The prompt box will display the specific wording of the question along with its corresponding informational value. For example, it might suggest: "We recommend asking 'Are you sweating?' This question helps to best differentiate between solar stroke and solar typhoid fever." This recommendation is not mandatory but serves as guidance to help students understand how to choose questions with greater information content.

[0071] Once students complete their consultation and submit their final diagnosis, a comprehensive evaluation report is automatically generated. The report compares the student's diagnosis with the highest probability result, i.e., the optimal diagnosis, demonstrating their consistency. It also presents a complete consultation path list, mapping each question to its corresponding entropy reduction ΔH, visually illustrating each question's contribution to uncertainty reduction. The evaluation report further includes the calculated optimal consultation path and, by comparing the student's path with the optimal path, identifies where the student deviated from the most efficient reasoning logic. In this way, students receive real-time feedback during the consultation process and can review the report afterward to understand the rationality and shortcomings of their questioning during reasoning.

[0072] Based on the same inventive concept, this application also provides a clinical teaching evaluation system for TCM classic formulas based on information entropy, including: The knowledge base construction module constructs a conditional probability knowledge base based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonyms into standard symptom identifiers, and the diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts. The module initializes the diagnostic space and feature space, and assigns initial prior probabilities to each diagnostic feature in the diagnostic space and each symptom question in the feature space based on the conditional probability knowledge base, calculating the initial information entropy. The interactive assessment module acquires the input symptom characteristics, calls the corresponding conditional probabilities based on the conditional probability knowledge base, performs Bayesian updates, and calculates the posterior probability of diagnosis. After each Bayesian update, the diagnostic space information entropy is calculated, and the entropy reduction is determined based on the initial information entropy. The entropy reduction is then bound to the input symptom features to form a consultation path. The diagnostic probability distribution and entropy reduction are recorded, and the diagnostic distribution and entropy value changes are output. The interactive display module iterates through the candidate symptom features that have not been input, calculates the expected information gain of the candidate symptom features, and outputs a recommendation question based on the expected information gain.

[0073] This invention constructs a conditional probability knowledge base and introduces information entropy as a measure of uncertainty in the diagnostic and feature spaces. After inputting symptom features, it dynamically adjusts the diagnostic probability distribution through Bayesian updates, binding entropy reduction to the input symptom features to form a traceable consultation path. This invention calculates the change in information entropy after each acquisition of input symptom features, achieving dynamic quantification of the diagnostic probability distribution as an objective process evaluation indicator. Simultaneously, by calculating the expected information gain for outputting recommendation questions for uninputted symptom features, it provides real-time and objective evaluation and guidance of students' reasoning paths in simulated consultations, achieving precise modeling and quantitative analysis of students' reasoning paths in the context of TCM formula teaching evaluation.

[0074] This invention introduces information entropy into the teaching and assessment of TCM classic formulas for the first time. Unlike traditional information theory applications where it is merely used as an abstract measure, this invention binds entropy reduction to symptom input, forming a linked list of consultation paths. This ensures that each question corresponds to a quantitative record of uncertainty reduction. Through this design, information entropy is no longer an internal parameter of the algorithm, but a relevant indicator for measuring students' diagnostic efficiency. This achieves a shift from outcome-oriented to process-oriented approaches, solving the problem of quantification difficulties in the many-to-many relationships of TCM diagnosis, and providing an interpretable and reviewable feedback mechanism for teaching.

[0075] Based on the same inventive concept, the present invention also provides a computer device, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and executed the steps of the above-described clinical teaching and assessment method for TCM classic formulas based on information entropy.

[0076] The processing methods for computer devices can be referred to the description of the methods above, and will not be repeated here.

[0077] This application also provides a non-transitory machine-readable storage medium storing an executable program, which, when run by a microprocessor, causes the processor to execute the method provided in the above embodiments.

[0078] This invention discloses a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform the described methods.

[0079] This invention discloses a computer program product including a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform the described method.

[0080] The embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0081] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0082] Finally, it should be noted that the embodiments disclosed in this invention are merely preferred embodiments of this invention and are only used to illustrate the technical solutions of this invention, not to limit it. Although this invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this invention.

Claims

1. A clinical teaching evaluation method for TCM classic formulas based on information entropy, characterized in that, The method includes: A conditional probability knowledge base is constructed based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonym entries into standard symptom identifiers, and the diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts. Initialize the diagnostic space and feature space, and determine the initial prior probability for each diagnostic feature in the diagnostic space and each symptom question in the feature space based on the conditional probability knowledge base, and calculate the initial information entropy; The system acquires the input symptom features, calls the corresponding conditional probabilities from the conditional probability knowledge base, performs a Bayesian update, and calculates the posterior probability of diagnosis. After each Bayesian update, the diagnostic space information entropy is calculated, and the entropy reduction is determined based on the initial information entropy. The entropy reduction is then bound to the input symptom features to form a consultation path. The diagnostic probability distribution and entropy reduction are recorded, and the diagnostic distribution and entropy value changes are output. Iterate through the candidate symptom features that have not been input, calculate the expected information gain of the candidate symptom features, and output the recommendation question based on the expected information gain.

2. The clinical teaching evaluation method for TCM classic formulas based on information entropy according to claim 1, characterized in that, Construct a conditional probability knowledge base, including: Obtain the diagnostic mapping rule table and the symptom dictionary, wherein the symptom dictionary normalizes the synonyms or variant expressions in ancient books into unified symptom identifiers and assigns a unique index to each symptom identifier; The texts of ancient Chinese medicine books are analyzed, and when diagnostic keywords and symptom keywords co-occur in the same text or in adjacent contexts, a correspondence between diagnostic markers and symptom markers is established; the ancient Chinese medicine books mentioned include at least the "Treatise on Febrile and Miscellaneous Diseases"; Determine the initial prior probability of the co-occurrence scenario; The diagnostic identifier, symptom identifier, and corresponding probability value are stored in the form of triples to form the conditional probability knowledge base.

3. The clinical teaching evaluation method for TCM classic prescriptions based on information entropy according to claim 2, characterized in that, Determine the initial prior probability of the co-occurrence scenario, including: The passages in ancient Chinese medicine books are classified according to their importance to diagnosis, into core definition passages, formula-symptom correspondence passages, differential diagnosis or concurrent symptoms passages, and prognosis or contraindication passages. Symptoms appearing in the core definition clauses are set as the first probability value, symptoms appearing in the corresponding clauses of the prescription are set as the second probability value, symptoms appearing in the identification or concurrent syndrome clauses are set as the third probability value, symptoms appearing in the prognosis or contraindication clauses are set as the fourth probability value, and background probability values ​​are assigned to symptoms that do not appear. A score is obtained to indicate the correlation between diagnosis and symptoms. This score is then converted into a correction factor. When the score is higher than a preset baseline value, the corresponding conditional probability is increased. When the score is lower than the baseline value, the corresponding conditional probability is decreased.

4. The clinical teaching evaluation method for TCM classic prescriptions based on information entropy according to claim 1, characterized in that, The entropy reduction is bound to the input symptom features to form a consultation path, including: A consultation record is generated based on the symptom characteristics input each time. The consultation record includes symptom identifier, information entropy before update, information entropy after update, and entropy reduction. Multiple consultation records are linked sequentially to form a consultation path linked list.

5. The clinical teaching evaluation method for TCM classic prescriptions based on information entropy according to claim 1, characterized in that, Calculating the expected information gain of the candidate symptom features includes: The probability of occurrence and non-occurrence of candidate symptom features are calculated based on a conditional probability knowledge base. The diagnostic space of occurrence and non-occurrence probabilities is updated by Bayes' theorem, the posterior probability distribution of occurrence and non-occurrence probabilities is obtained, and the corresponding information entropy is calculated. The expected entropy is calculated by taking a weighted average of the two information entropies. By comparing the current information entropy with the expected entropy, the information gain value corresponding to the candidate symptom is obtained; The desired information gain is obtained by weighting and correcting the information gain by combining the priority field in the diagnostic mapping rule table.

6. The clinical teaching evaluation method for TCM classic formulas based on information entropy according to claim 5, characterized in that, The information entropy is a measure of the uncertainty of the diagnostic probability distribution, the entropy reduction is the difference between the information entropy before and after two consecutive inputs, and the information gain is the difference between the current information entropy and the expected information entropy corresponding to the candidate symptom.

7. The clinical teaching evaluation method for TCM classic prescriptions based on information entropy according to claim 6, characterized in that, The recommendation problem is the candidate problem with the largest expected information gain, determined by combining the priority weights in the diagnostic mapping table.

8. The clinical teaching evaluation method for TCM classic formulas based on information entropy according to claim 1, characterized in that, Also includes: The interactive interface displays the diagnostic probability distribution and outputs the diagnosis with the highest probability and its corresponding probability value, information entropy, and recommended input symptom features calculated based on information gain in a visual form. Upon receiving a diagnosis submission instruction, an evaluation report is generated, which includes the input symptom characteristics, a consultation path list, the most probable diagnosis, and a recommended optimal path.

9. A clinical teaching evaluation system for classical Chinese medicine formulas based on information entropy, characterized in that, include: The knowledge base construction module constructs a conditional probability knowledge base, which is based on a symptom dictionary and a diagnostic mapping rule table. The symptom dictionary is used to normalize synonym entries into standard symptom identifiers, and the diagnostic mapping rule table is used to define diagnostic identifiers, corresponding keywords, necessary contexts, and exclusion contexts. The interactive assessment module acquires the input symptom characteristics, calls the corresponding conditional probabilities based on the conditional probability knowledge base, performs Bayesian updates, and calculates the posterior probability of diagnosis. After each Bayesian update, the diagnostic space information entropy is calculated, and the entropy reduction is determined based on the initial information entropy. The entropy reduction is then bound to the input symptom features to form a consultation path. The diagnostic probability distribution and entropy reduction are recorded, and the diagnostic distribution and entropy value changes are output. The interactive display module iterates through the candidate symptom features that have not been input, calculates the expected information gain of the candidate symptom features, and outputs a recommendation question based on the expected information gain.

10. A computer storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the steps of the information entropy-based clinical teaching evaluation method for TCM classic prescriptions as described in any one of claims 1-8.

Images