Lymphoma traditional Chinese medicine syndrome recognition model construction method and device and electronic equipment
By constructing a TCM syndrome identification model for lymphoma and utilizing a symptom coding mapping dictionary and machine learning model, the problem of inaccurate syndrome identification in TCM treatment of lymphoma was solved, achieving efficient and accurate TCM syndrome identification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THE FIRST AFFILIATED HOSPITAL OF ZHEJIANG CHINESE MEDICAL UNIVERSITY
- Filing Date
- 2026-01-19
- Publication Date
- 2026-05-05
AI Technical Summary
The lack of authoritative consensus and accurate syndrome identification methods in the treatment of lymphoma with traditional Chinese medicine leads to difficulties in clinical application.
By obtaining the TCM symptom tables of multiple lymphoma patients, generating a symptom coding table using a predefined symptom coding mapping dictionary, and training a machine learning model, a TCM syndrome recognition model for lymphoma was constructed.
It improves the efficiency and accuracy of TCM syndrome identification in lymphoma patients, and supports the automated identification and accurate prediction of TCM syndromes.
Smart Images
Figure CN121983281A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of TCM syndrome recognition technology, and in particular to a method, device, electronic device, storage medium and program product for constructing a TCM syndrome recognition model for lymphoma. Background Technology
[0002] Lymphoma is the most common hematologic malignancy. While current treatments such as small molecule drugs, monoclonal antibodies, antibody-drug conjugates, and cell therapy have significantly improved prognosis, unmet needs remain in its treatment, including: 1. Lack of effective interventions to prevent relapse in patients who have completed treatment; 2. Lack of specific interventions for lymphoma patients in the observation and waiting period; and 3. Lack of effective interventions for complications arising from Western medicine treatments. Traditional Chinese medicine (TCM) has inherent advantages in immunomodulation, and its interventions have shown definite efficacy in treating lymphoma-related immune system abnormalities, during the observation and waiting period, in maintenance therapy after treatment completion, and in patients with complications. However, there is a lack of authoritative consensus or guidelines for TCM treatment of this disease, and differing understandings among practitioners lead to a complex diagnostic system, hindering accurate identification of TCM syndromes and impeding the effective application of TCM in clinical practice. Summary of the Invention
[0003] This invention provides a method, device, electronic device, storage medium, and program product for constructing a TCM syndrome identification model for lymphoma, which can improve the efficiency and accuracy of TCM syndrome identification for lymphoma patients.
[0004] According to one aspect of the present invention, a method for constructing a TCM syndrome identification model for lymphoma is provided, the method comprising: Obtain a TCM (Traditional Chinese Medicine) symptom table for multiple lymphoma patients; wherein the TCM symptom table includes multiple symptom columns and the corresponding symptoms for each symptom column, and the symptoms include single symptoms and compound symptoms; Based on a predefined symptom coding mapping dictionary, determine the numerical codes corresponding to single symptoms and compound symptoms in the TCM waiting symptom table, and generate the waiting symptom coding table. The machine learning model was trained based on the aforementioned symptom coding table to obtain a TCM syndrome recognition model for lymphoma.
[0005] According to another aspect of the present invention, a device for constructing a TCM syndrome identification model for lymphoma is provided, the device comprising: The data acquisition module is used to acquire the TCM waiting symptom table of multiple lymphoma patients; wherein, the TCM waiting symptom table includes multiple symptom columns and the symptoms corresponding to each symptom column, and the symptoms include single symptoms and compound symptoms; The data encoding module is used to determine the numerical code corresponding to a single symptom and the numerical code string corresponding to a compound symptom in the TCM waiting symptom table according to a predefined symptom encoding mapping dictionary, and generate the waiting symptom encoding table. The model training module is used to train the machine learning model based on the waiting symptom coding table to obtain a TCM syndrome recognition model for lymphoma.
[0006] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the lymphoma TCM syndrome identification model construction method according to any embodiment of the present invention.
[0007] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the method for constructing a lymphoma TCM syndrome recognition model according to any embodiment of the present invention.
[0008] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method for constructing a lymphoma TCM syndrome recognition model as described in any of the embodiments of the present disclosure.
[0009] The technical solution of this invention involves obtaining a TCM (Traditional Chinese Medicine) symptom table for multiple lymphoma patients. This table includes multiple symptom columns and corresponding symptoms for each column, with symptoms including single and compound symptoms. Based on a predefined symptom coding mapping dictionary, numerical codes for single and compound symptoms are determined, generating a symptom coding table. A machine learning model is trained using this symptom coding table to obtain a TCM syndrome recognition model for lymphoma. This technical solution, by encoding single and compound symptoms in the TCM symptom table according to a symptom coding mapping dictionary, preserves the characteristic information of multiple coexisting symptoms in TCM, generates a symptom coding table, and trains a machine learning model to obtain a TCM syndrome recognition model for lymphoma. This improves the efficiency and accuracy of TCM syndrome recognition for lymphoma patients.
[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description
[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0012] Figure 1 This is a flowchart of a method for constructing a TCM syndrome recognition model for lymphoma according to Embodiment 1 of the present invention; Figure 2 This is a flowchart of a method for constructing a TCM syndrome recognition model for lymphoma according to Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of a device for constructing a TCM syndrome recognition model for lymphoma according to Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device that implements the method for constructing a TCM syndrome recognition model for lymphoma according to an embodiment of the present invention. Detailed Implementation
[0013] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.
[0014] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0015] Example 1 Figure 1 This invention provides a flowchart of a method for constructing a TCM syndrome recognition model for lymphoma, as described in Embodiment 1. This embodiment is applicable to situations where a model is constructed to identify TCM syndromes in lymphoma patients. This method can be executed by a lymphoma TCM syndrome recognition model construction device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes: S110. Obtain the TCM waiting symptom table of multiple lymphoma patients; the TCM waiting symptom table includes multiple symptom columns and the corresponding symptoms for each symptom column, and the symptoms include single symptoms and compound symptoms.
[0016] The Traditional Chinese Medicine (TCM) symptom chart is used in an Excel spreadsheet to objectively record the symptoms of lymphoma patients diagnosed by the four diagnostic instruments. It includes multiple symptom columns (such as fever, chills, tongue coating, pulse, etc.) and the corresponding symptoms for each column. Symptoms can be single or complex, with complex symptoms consisting of a combination of multiple single symptoms. For example, fever may correspond to single symptoms such as low-grade fever, tidal fever, or high fever, while tongue coating may correspond to complex symptoms such as white, yellow, thick, or greasy tongue coating.
[0017] In this embodiment of the invention, a TCM symptom table of multiple lymphoma patients can be obtained, which is an objective symptom record obtained by diagnosing multiple lymphoma patients through the four diagnostic instruments. This solves the problem of lack of TCM doctors and strong subjective diagnoses in Western hospitals and integrated TCM and Western medicine hospitals, and provides a reliable data foundation for the subsequent construction of a TCM syndrome recognition model for lymphoma.
[0018] S120. Based on the predefined symptom coding mapping dictionary, determine the numerical codes corresponding to single symptoms and compound symptoms in the TCM waiting symptom table, and generate the waiting symptom coding table.
[0019] The structure of the symptom coding mapping dictionary can be {symptom column: {symptom: code}}. For example, the predefined symptom coding mapping dictionary includes: {Fever: {None, Low-grade fever: 0, Hot flashes: 1, High fever, Persistent fever: 2}}, {Cold intolerance: {None: 0, Aversion to wind: 1, Cold extremities, Cold intolerance: 2}}, {General symptoms: {Palpitation, Shortness of breath, Fatigue: 0, Weak voice, Dizziness, Insomnia, Tinnitus: 1, Lower back and knee weakness, Hot palms and soles: 2, Irritability, Bitter taste in mouth, Distension in the hypochondrium, Sighing, Depression: 2}}, {Sweating: {None, Normal: 0, Sweating with slight movement, Spontaneous sweating: 1, Sweating upon waking, Night sweats: 2}}, {Urination: {Normal: 0, Frequent urination and / or Clear urine: 1, Oliguria and / or Dark urine: 2}}, {Appetite: {Normal: 0, Loss of appetite: 1, Fasting after eating: 2}}, {Reproductive system: {Normal: 0, Decreased libido: 1, Hypersexuality: 2}} Intent: 2, Nocturnal emission: 3, Scrotal dampness: 4, Yellow and foul-smelling vaginal discharge: 5, Lymph nodes: {Normal: 0, No pain or itching: 1, Mild pain: 2, Not palpable: 3, Soft texture: 4, Hard texture: 5, Tongue appearance: {Pale red: 1, Red: 2, Dark red: 3, Dark purple: 4, Red on the tip and sides of the tongue: 5, Tongue coating: {Thin white: 1, Yellow: 2, Thick and greasy: 3, Grayish black: 4, Patchy and peeled: 5, No tongue coating: 6}}, {tongue body: {normal: 0, teeth marks on the tongue side: 1, cracks: 2, swollen: 3, thin: 4, prickles on the tip of the tongue: 5, petechiae and ecchymosis: 6}}, {pulse: {normal or moderate pulse strength: 0, deep: 1, floating: 2, thin: 3, surging: 4, rapid: 5, slow: 6, even: 7, irregular or intermittent: 8, weak: 9, strong: 10, slippery: 11, hesitant, wiry, tight: 12}} etc.
[0020] In this embodiment of the invention, a single symptom in the TCM (Traditional Chinese Medicine) waiting symptom table can be converted into a numerical code based on a predefined symptom coding mapping dictionary, and a compound symptom in the TCM waiting symptom table can be converted into a numerical code string to generate a waiting symptom coding table, making the data compatible with machine learning modeling. By converting single symptoms into numerical codes and compound symptoms into numerical code strings, the traditional single-value feature assumption can be broken, the characteristic information of "multiple symptoms coexisting" in TCM can be preserved, and the foundation for accurate modeling of high-dimensional sparse data can be laid.
[0021] Optionally, determining the numerical code corresponding to a single symptom and the numerical code string corresponding to a compound symptom in the TCM waiting symptom table according to a predefined symptom coding mapping dictionary includes: for the single symptom, matching the single symptom with preset symptoms in the symptom coding mapping dictionary, and using the numerical code corresponding to the preset symptom that successfully matches the single symptom as the numerical code of the single symptom; for the compound symptom, splitting the compound symptom into multiple single symptoms according to a first preset separator, determining the numerical code corresponding to each single symptom, and concatenating the numerical codes corresponding to the multiple single symptoms into a numerical code string using a second preset separator.
[0022] The first and second preset separators can be set by technicians according to the actual situation. For example, the first preset separator can be " The second preset separator can be "," etc., and the embodiments of the present invention do not limit it to this.
[0023] In this embodiment of the invention, based on a predefined symptom coding mapping dictionary, the numerical codes corresponding to single symptoms and compound symptoms in the TCM consultation symptom table are determined. Specifically: for single symptoms, mapping can be performed directly, matching the single symptom with preset symptoms in the symptom coding mapping dictionary, and using the numerical code corresponding to the preset symptom that successfully matches the single symptom as the numerical code of the single symptom. For compound symptoms, the numerical code can be determined first according to a first preset separator in the compound symptom (such as "..."). The method breaks down complex symptoms into multiple individual symptoms, maps each individual symptom item by item to determine its corresponding numerical code, and concatenates these numerical codes into a numerical code string using a second preset separator (such as ","). This three-step vectorization paradigm of "splitting-mapping-concatenation" achieves zero-information-entropy loss numerical encoding of complex symptoms. Furthermore, during the numerical encoding process, undefined symptoms in the Chinese symptom waiting table can be initially assigned a value of 0, and missing values in the table can be retained as None to support subsequent filling.
[0024] S130. The machine learning model is trained based on the waiting symptom coding table to obtain the TCM syndrome recognition model for lymphoma.
[0025] In this embodiment of the invention, before training the machine learning model based on the waiting symptom coding table, the symptom columns can be converted from Chinese to English abbreviations through header mapping, such as converting fever to "Fever," and non-feature columns such as "pathological type" can be removed from the waiting symptom coding table. Then, the machine learning model can be trained based on the processed waiting symptom coding table to obtain a lymphoma TCM syndrome recognition model. Optionally, experts can annotate the TCM syndromes corresponding to the waiting symptom coding table based on experience, and then train the machine learning model based on the annotated waiting symptom coding table to obtain a lymphoma TCM syndrome recognition model.
[0026] The technical solution of this invention involves obtaining a TCM (Traditional Chinese Medicine) symptom table for multiple lymphoma patients. This table includes various symptom columns and corresponding symptoms for each column, with symptoms including single and compound symptoms. Based on a predefined symptom coding mapping dictionary, numerical codes for single and compound symptoms are determined, generating a symptom coding table. A machine learning model is trained using this coding table to obtain a TCM syndrome recognition model for lymphoma. This technical solution, by encoding single and compound symptoms in the TCM symptom table according to a symptom coding mapping dictionary, preserves the characteristic information of multiple coexisting symptoms in TCM, generates a symptom coding table, and trains a machine learning model to obtain a TCM syndrome recognition model for lymphoma. This improves the efficiency and accuracy of TCM syndrome recognition for lymphoma patients.
[0027] Example 2 Figure 2 This is a flowchart of a method for constructing a TCM syndrome recognition model for lymphoma according to Embodiment 2 of the present invention. The embodiments of the present invention are optimized based on the above embodiments. Solutions not described in detail in the embodiments of the present invention are described in the above embodiments. Figure 2 As shown, the method includes: S210. Obtain the TCM waiting symptom tables of multiple lymphoma patients.
[0028] S220. Based on the predefined symptom coding mapping dictionary, determine the numerical codes corresponding to single symptoms and compound symptoms in the TCM waiting symptom table, and generate the waiting symptom coding table.
[0029] S230. Determine the first symptom feature of a single symptom and the second symptom feature of a compound symptom in the waiting symptom coding table through unique hot coding, and generate a waiting symptom feature table corresponding to the waiting symptom coding table.
[0030] One-hot encoding is a commonly used encoding technique in machine learning and data preprocessing. It is mainly used to convert categorical variables (discrete features) into numerical forms that can be processed by machine learning algorithms. Its core idea is to map each category to a binary vector, in which only one position is 1 and the rest are 0, thereby avoiding the misinterpretation of the numerical relationship between categories.
[0031] In this embodiment of the invention, one-hot encoding can be used to determine the first symptom feature of a single symptom and the second symptom feature of a compound symptom based on the numerical code of a single symptom and the numerical code string of a compound symptom in the waiting symptom encoding table. A waiting symptom feature table corresponding to the waiting symptom encoding table can be generated based on the first symptom feature and the second symptom feature, providing structured input to the model and supporting feature importance analysis.
[0032] Optionally, determining the first symptom feature of a single symptom and the second symptom feature of a compound symptom in the waiting symptom coding table through one-hot encoding includes: for each symptom column in the waiting symptom coding table, determining the symptom type included in the symptom column; determining the binary dummy variable of the single symptom as the first symptom feature by one-hot encoding based on the symptom type and the numerical code corresponding to the single symptom in the symptom column; and determining the symptom existence binary matrix of the compound symptom as the second symptom feature by one-hot encoding based on the symptom type and the numerical code string corresponding to the compound symptom in the symptom column.
[0033] In this embodiment of the invention, the first symptom feature of a single symptom and the second symptom feature of a compound symptom in the symptom coding table are determined by unique-hot encoding. Specifically, firstly, the symptom types included in each symptom column of the symptom coding table are determined, such as tongue coating including thin white, yellow, thick and greasy, gray-black, and mottled. Secondly, through unique-hot encoding, based on the symptom type corresponding to the symptom column and the numerical code corresponding to the single symptom in the symptom column, a binary dummy variable for the single symptom is determined as the first symptom feature. For example, if the numerical code corresponding to the single symptom in the tongue coating is 1, the single symptom can be determined to be thin white, and the binary dummy variable determined by unique-hot encoding can be [1,0,0,0,0,0]. Finally, through unique-hot encoding, based on the symptom type corresponding to the symptom column and the numerical code string corresponding to the compound symptom in the symptom column, the numerical code string is parsed according to a second preset separator to determine the binary dummy variables corresponding to each numerical code in the numerical code string, and a symptom existence binary matrix of the compound symptom is generated as the second symptom feature. For example, if the numerical encoding string corresponding to a complex symptom in the tongue coating is "1,3", it can be determined that the complex symptom is a combination of thin white and thick greasy coatings. The binary matrix of symptom presence determined by unique thermal encoding can be... .
[0034] S240. Perform recursive clustering on the waiting symptom feature table and determine the TCM syndrome corresponding to each category of waiting symptom feature table as the cluster label.
[0035] Recursive clustering is a method for grouping data through iterative or hierarchical strategies. Its core idea is to gradually decompose the data into smaller subsets (subclusters) until a specific termination condition is met.
[0036] In this embodiment of the invention, after generating a waiting symptom feature table corresponding to the waiting symptom coding table, the waiting symptom feature table can be recursively clustered. Specifically, recursive clustering with recursive binary division maximum clustering can be used to significantly reduce the clustering variance under high-dimensional sparse data and improve clustering stability and interpretability. Then, the TCM syndrome corresponding to each type of waiting symptom feature table is determined as the clustering label.
[0037] S250. The waiting symptom feature table is labeled according to the cluster labels, and the random forest model is trained according to the labeled waiting symptom feature table to obtain the lymphoma TCM syndrome recognition model.
[0038] In this embodiment of the invention, the symptom feature table for outpatient visits can be labeled according to clustering labels, and a random forest model can be trained based on the labeled symptom feature table to obtain a TCM syndrome recognition model for lymphoma. By using clustering labels as a standard to drive the random forest model for high-precision supervised learning, a knowledge distillation mechanism from unlabeled to labeled can be formed, realizing automated syndrome type discovery and accurate prediction, balancing exploration and accuracy. Furthermore, the technical solution in this invention, from obtaining the TCM symptom table to training the TCM syndrome recognition model for lymphoma, is purely file-driven, requires no database deployment, has a low installation threshold, and has the potential for large-scale promotion.
[0039] Optionally, the step of labeling the waiting symptom feature table according to the cluster labels and training the random forest model based on the labeled waiting symptom feature table to obtain the lymphoma TCM syndrome recognition model includes: inputting the labeled waiting symptom feature table into the random forest model, using the labeled cluster labels as classification targets, determining the optimal parameter combination among multiple parameter combinations through grid search with cross-validation; and determining the random forest model with the optimal parameter combination as the model parameters as the lymphoma TCM syndrome recognition model.
[0040] In this embodiment of the invention, a patient symptom feature table is labeled according to cluster labels, and a random forest model is trained based on the labeled patient symptom feature table to obtain a lymphoma TCM syndrome recognition model. Specifically, the labeled patient symptom feature table is input into the random forest model, the labeled cluster labels are used as the classification target, and the optimal parameter combination is determined from multiple parameter combinations through a grid search with cross-validation. The random forest model with the optimal parameter combination as the model parameters is then determined as the lymphoma TCM syndrome recognition model. For example, the labeled cluster labels can be used as the classification target, and the optimal parameter combination can be determined from parameter combinations of tree number [100, 200] and tree depth [unlimited, 10, 20] through a grid search with 3-fold cross-validation. The random forest model with the optimal parameter combination as the model parameters is then determined as the lymphoma TCM syndrome recognition model.
[0041] Optionally, the step of labeling the waiting symptom coding table according to the clustering labels and training the random forest model based on the labeled waiting symptom coding table to obtain the lymphoma TCM syndrome recognition model further includes: verifying the performance of the lymphoma TCM syndrome recognition model through a pre-set test set, and calculating the contribution value of each symptom feature in the waiting symptom feature table in the test set when participating in model prediction; generating a global feature importance ranking based on the contribution value, and generating a force-directed graph and a summary bee colony graph corresponding to each waiting symptom feature table.
[0042] Force-directed graphs are a visualization technique that uses simulated physical forces (such as attraction and repulsion) to demonstrate the relationship between symptom characteristics and syndrome categories. Nodes represent symptoms or syndromes, and edges represent the strength of the association. Beehive graphs, on the other hand, are a data visualization method that uses clusters of points to show the distribution of symptom characteristics across different syndrome categories, helping to intuitively understand the association patterns between symptoms and syndromes.
[0043] In this embodiment of the invention, the performance of the lymphoma TCM syndrome recognition model can be verified using a pre-set test set, and a full-link SHAP contribution value interpretability framework can be constructed. Specifically, the SHAP contribution value of each symptom feature in the waiting symptom feature table in the test set when participating in model prediction can be calculated. A global feature importance ranking is generated based on the SHAP contribution value, and a force-directed graph and a summary beehive graph corresponding to each waiting symptom feature table are generated, forming a transparent interpretation path from model output to doctor's judgment, significantly improving the clinical credibility and auditability of AI-assisted diagnosis. Optionally, the symptom columns in the output results can be converted from English abbreviations to clinical Chinese expressions, such as converting "fever" to "excitement".
[0044] S260. Obtain the target TCM waiting symptom table for the target lymphoma patient, and generate the target waiting symptom feature table based on the target TCM waiting symptom table.
[0045] In this embodiment of the invention, a target TCM (Traditional Chinese Medicine) symptom table for target lymphoma patients can be obtained, and a data processing procedure consistent with that used in the model training phase can be performed on the target TCM symptom table to generate a target symptom feature table corresponding to the target TCM symptom table. This unified data processing procedure can prevent feature dimension misalignment or model collapse caused by missing symptoms in new samples, ensuring the model has industrial-grade deployment stability.
[0046] S270. Input the target symptom feature table into the lymphoma TCM syndrome recognition model to obtain the lymphoma TCM syndrome recognition results corresponding to the target TCM symptom table, and generate the force-directed graph and summary beehive graph corresponding to each target symptom feature table.
[0047] In this embodiment of the invention, after generating the target symptom feature table, the target symptom feature table can be input into the lymphoma TCM syndrome recognition model to obtain the lymphoma TCM syndrome recognition result corresponding to the target TCM symptom table, and generate a force-directed graph and a summary beehive graph corresponding to each target symptom feature table to support clinicians in tracing decision-making basis.
[0048] The technical solution of this invention involves: acquiring a TCM (Traditional Chinese Medicine) symptom table for multiple lymphoma patients; determining the numerical codes corresponding to single symptoms and compound symptoms in the TCM symptom table based on a predefined symptom coding mapping dictionary, generating a symptom coding table; determining the first symptom feature of a single symptom and the second symptom feature of a compound symptom in the symptom coding table through one-hot coding, generating a symptom feature table corresponding to the symptom coding table; recursively clustering the symptom feature table, determining the TCM syndrome corresponding to each category of symptom feature table as a cluster label; labeling the symptom feature table according to the cluster label, and training a random forest model based on the labeled symptom feature table to obtain a lymphoma TCM syndrome recognition model; acquiring a target TCM symptom table for a target lymphoma patient, and generating a target symptom feature table based on the target TCM symptom table; inputting the target symptom feature table into the lymphoma TCM syndrome recognition model to obtain the lymphoma TCM syndrome recognition result corresponding to the target TCM symptom table, and generating a force-directed graph and a summary beehive graph corresponding to each target symptom feature table. The technical solution of this invention firstly determines the numerical codes for single symptoms and compound symptoms in the TCM (Traditional Chinese Medicine) symptom identification table by using a symptom coding mapping dictionary, preserving the characteristic information of multiple coexisting symptoms in TCM, and generating a symptom identification table. Secondly, a learning method from unsupervised clustering to supervised classification is adopted to train a machine learning model using the symptom identification table, resulting in a TCM syndrome identification model for lymphoma. This achieves automated syndrome type discovery and accurate prediction, balancing exploratory nature and accuracy. Finally, the TCM syndrome identification model for lymphoma is used to identify the TCM syndromes in the target TCM symptom identification table, improving the efficiency and accuracy of TCM syndrome identification for lymphoma patients.
[0049] Example 3 Figure 3 This is a schematic diagram of a device for constructing a TCM syndrome recognition model for lymphoma, provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes: The data acquisition module 310 is used to acquire a TCM symptom table for multiple lymphoma patients; wherein, the TCM symptom table includes multiple symptom columns and the symptoms corresponding to each symptom column, and the symptoms include single symptoms and compound symptoms; Data encoding module 320 is used to determine the numerical code corresponding to a single symptom and the numerical code string corresponding to a compound symptom in the TCM waiting symptom table according to a predefined symptom encoding mapping dictionary, and generate a waiting symptom encoding table. The model training module 330 is used to train the machine learning model based on the waiting symptom coding table to obtain a lymphoma TCM syndrome recognition model.
[0050] Optionally, the data encoding module 320 includes: A single symptom encoding unit is used to match the single symptom with preset symptoms in the symptom encoding mapping dictionary, and use the numerical code corresponding to the preset symptom that successfully matches the single symptom as the numerical code of the single symptom. The composite symptom encoding unit is used to split the composite symptom into multiple individual symptoms according to a first preset separator, determine the numerical code corresponding to each individual symptom, and concatenate the numerical codes corresponding to the multiple individual symptoms into a numerical code string using a second preset separator.
[0051] Optional, model training module 330 includes: The symptom feature determination unit is used to determine the first symptom feature of a single symptom and the second symptom feature of a compound symptom in the waiting symptom coding table through unique hot coding, and generate a waiting symptom feature table corresponding to the waiting symptom coding table; The clustering label determination unit is used to recursively cluster the waiting symptom feature table and determine the TCM syndrome corresponding to each category of waiting symptom feature table as the clustering label. The identification model training unit is used to label the waiting symptom feature table according to the cluster labels, and to train the random forest model according to the labeled waiting symptom feature table to obtain the lymphoma TCM syndrome identification model.
[0052] Optionally, the symptom feature determination unit is specifically used for: For each symptom column in the waiting symptom coding table, determine the symptom type included in the symptom column; By using unique heat encoding, based on the symptom type and the numerical code corresponding to a single symptom in the symptom list, a binary dummy variable of the single symptom is determined as the first symptom feature. By using one-hot encoding, based on the symptom type and the numerical encoding string corresponding to the compound symptom in the symptom column, a binary matrix of the symptom existence of the compound symptom is determined as the second symptom feature. Optionally, the recognition model training unit is specifically used for: The labeled waiting symptom feature table is input into the random forest model, and the labeled clustering labels are used as the classification target. The optimal parameter combination is determined from multiple parameter combinations through grid search with cross-validation. The random forest model with the optimal parameter combination as model parameters was determined and used as a TCM syndrome identification model for lymphoma.
[0053] Optionally, the recognition model training unit is further specifically used for: The performance of the lymphoma TCM syndrome recognition model was verified by using a pre-set test set, and the contribution value of each symptom feature in the waiting symptom feature table in the test set when participating in model prediction was calculated. Based on the contribution values, a global feature importance ranking is generated, and a force-directed graph and a summary bee colony graph are generated for each waiting symptom feature table.
[0054] Optionally, the device further includes: The target symptom feature table generation module is used to obtain the target TCM waiting symptom table of the target lymphoma patient and generate the target waiting symptom feature table based on the target TCM waiting symptom table. The lymphoma TCM syndrome identification module is used to input the target syndrome feature table into the lymphoma TCM syndrome identification model, obtain the lymphoma TCM syndrome identification result corresponding to the target syndrome feature table, and generate a force-directed graph and a summary beehive graph corresponding to each target syndrome feature table. The lymphoma TCM syndrome identification model construction device provided in the embodiments of the present invention can execute the lymphoma TCM syndrome identification model construction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0055] Example 4 Figure 4 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.
[0056] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0057] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0058] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the method for constructing a TCM syndrome recognition model for lymphoma.
[0059] In some embodiments, the method for constructing a TCM syndrome recognition model for lymphoma can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the method for constructing a TCM syndrome recognition model for lymphoma described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to execute the method for constructing a TCM syndrome recognition model for lymphoma by any other suitable means (e.g., by means of firmware).
[0060] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0061] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0062] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0063] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0064] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0065] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0066] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.
[0067] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.
[0068] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A method for constructing a TCM syndrome recognition model for lymphoma, characterized in that, The method includes: Obtain a TCM (Traditional Chinese Medicine) symptom table for multiple lymphoma patients; wherein the TCM symptom table includes multiple symptom columns and the corresponding symptoms for each symptom column, and the symptoms include single symptoms and compound symptoms; Based on a predefined symptom coding mapping dictionary, determine the numerical codes corresponding to single symptoms and compound symptoms in the TCM waiting symptom table, and generate the waiting symptom coding table. The machine learning model was trained based on the aforementioned symptom coding table to obtain a TCM syndrome recognition model for lymphoma.
2. The method according to claim 1, characterized in that, The step of determining the numerical code for a single symptom and the numerical code string for a compound symptom in the TCM consultation symptom table according to a predefined symptom coding mapping dictionary includes: For the single symptom, the single symptom is matched with preset symptoms in the symptom coding mapping dictionary, and the numerical code corresponding to the preset symptom that successfully matches the single symptom is used as the numerical code of the single symptom. For the compound symptoms, the compound symptoms are split into multiple individual symptoms according to the first preset separator, the numerical code corresponding to each individual symptom is determined, and the numerical codes corresponding to the multiple individual symptoms are concatenated into a numerical code string by the second preset separator.
3. The method according to claim 1, characterized in that, The step of training a machine learning model based on the waiting symptom coding table to obtain a lymphoma TCM syndrome recognition model includes: The first symptom feature of a single symptom and the second symptom feature of a compound symptom in the waiting symptom coding table are determined by unique hot coding, and a waiting symptom feature table corresponding to the waiting symptom coding table is generated. Recursively cluster the waiting symptom feature table and determine the TCM syndrome corresponding to each category of waiting symptom feature table as the cluster label; The waiting symptom feature table is labeled according to the clustering labels, and the random forest model is trained according to the labeled waiting symptom feature table to obtain the lymphoma TCM syndrome recognition model.
4. The method according to claim 3, characterized in that, The process of determining the first symptom feature of a single symptom and the second symptom feature of a complex symptom in the waiting symptom coding table through unique thermal coding includes: For each symptom column in the waiting symptom coding table, determine the symptom type included in the symptom column; By using unique heat encoding, based on the symptom type and the numerical code corresponding to a single symptom in the symptom list, a binary dummy variable of the single symptom is determined as the first symptom feature. By using one-hot encoding, based on the symptom type and the numerical encoding string corresponding to the compound symptom in the symptom list, a binary matrix of the symptom existence of the compound symptom is determined as the second symptom feature.
5. The method according to claim 3, characterized in that, The step of labeling the waiting symptom feature table according to the cluster labels, and training the random forest model based on the labeled waiting symptom feature table to obtain the lymphoma TCM syndrome recognition model includes: The labeled waiting symptom feature table is input into the random forest model, and the labeled clustering labels are used as the classification target. The optimal parameter combination is determined from multiple parameter combinations through grid search with cross-validation. The random forest model with the optimal parameter combination as model parameters was determined and used as a TCM syndrome identification model for lymphoma.
6. The method according to claim 3, characterized in that, The step of labeling the waiting symptom coding table according to the cluster labels and training the random forest model based on the labeled waiting symptom coding table to obtain the lymphoma TCM syndrome recognition model further includes: The performance of the lymphoma TCM syndrome recognition model was verified by using a pre-set test set, and the contribution value of each symptom feature in the waiting symptom feature table in the test set when participating in model prediction was calculated. Based on the contribution values, a global feature importance ranking is generated, and a force-directed graph and a summary bee colony graph are generated for each waiting symptom feature table.
7. The method according to claim 1, characterized in that, After training the machine learning model based on the waiting symptom coding table to obtain the lymphoma TCM syndrome recognition model, the method further includes: Obtain the target TCM waiting symptom table of the target lymphoma patient, and generate the target waiting symptom feature table based on the target TCM waiting symptom table; The target symptom feature table is input into the lymphoma TCM syndrome recognition model to obtain the lymphoma TCM syndrome recognition result corresponding to the target TCM symptom feature table, and a force-directed graph and a summary beehive graph corresponding to each target symptom feature table are generated.
8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the method for constructing a lymphoma TCM syndrome recognition model according to any one of claims 1-7.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the method for constructing a lymphoma TCM syndrome recognition model as described in any one of claims 1-7.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for constructing a lymphoma TCM syndrome recognition model as described in any one of claims 1-7.