Traditional Chinese medicine compound screening method based on machine learning

Abstract

The invention discloses a traditional Chinese medicine compound screening method based on machine learning, and relates to the technical field of biological medicine.According to the traditional Chinese medicine compound screening method, multi-source traditional Chinese medicine compound spleen tonifying related component data are collected and integrated, a data set containing information such as components, medicine effects and indications is constructed, and follow-up model training is supported; utilizing machine learning to train a compound screening model, identifying drug effect related components, predicting effective parts of traditional Chinese medicinal materials of the components, and revealing contribution of the components to the drug effect; structural characteristics of chemical components are extracted, potential pharmacological effects of the chemical components are analyzed, and drug effect and toxicity analysis is supported; and a pharmacodynamic index set is constructed to evaluate the effect, toxicity and timeliness of the components, and support is provided for pharmacodynamic optimization. Animal experiments are combined to verify the synergistic effect among the components, so that the model prediction accuracy is ensured. The proportion of the compound components is adjusted according to the experimental result to maximize the drug effect and reduce the side effect to meet the requirements of different patients.

Description

Technical Field

[0001] The present invention relates to the field of biomedicine technology, and in particular to a method for screening a traditional Chinese medicine compound based on machine learning. Background Art

[0002] Ulcerative colitis is a chronic, recurrent, inflammatory gastrointestinal disease that can cause inflammatory damage to the colon mucosa, destroy the gastrointestinal mucosal barrier, and increase the risk of colorectal cancer. At present, the treatment of ulcerative colitis mainly includes drug therapy and surgical treatment, but the effect is limited and there are certain safety risks. Therefore, the treatment of ulcerative colitis with Chinese herbal compound drugs is one of the effective treatment methods. In order to make the drugs for the treatment of ulcerative colitis suitable for the symptoms and with the best efficacy, it is necessary to screen the ingredients of various Chinese herbal compound drugs for the treatment of ulcerative colitis on the market in order to obtain effective drugs for the treatment of ulcerative colitis.

[0003] In summary, how to improve the efficacy of Chinese herbal compound ingredients in the treatment of ulcerative colitis is an urgent problem to be solved and optimized in the Chinese herbal compound screening method based on machine learning. Summary of the invention

[0004] The present invention provides a method for screening a Chinese medicine compound based on machine learning, which solves the technical problem of how to improve the efficacy of Chinese medicine compound ingredients in treating ulcerative colitis.

[0005] In order to solve the above technical problems, the present invention provides a method for screening Chinese medicine compound prescriptions based on machine learning, and the specific technical scheme is as follows:

[0006] A method for screening a Chinese medicine compound based on machine learning comprises the following steps:

[0007] Collect the data of spleen-tonifying ingredients in traditional Chinese medicine compound to obtain the traditional Chinese medicine compound dataset;

[0008] Integrate and encode each data item in the Chinese medicine compound data set into sequence data, train the sequence data to construct a compound screening model, and output the results of identifying the efficacy of the components and the effective parts of the components of the Chinese medicine compound;

[0009] Inputting at least one data item in the newly entered Chinese medicine compound data set into the compound screening model to output the corresponding Chinese medicine compound component efficacy and component effective part identification results after screening;

[0010] Based on the recognition results of the compound screening model, the chemical component structure characteristics of the principal component data in the traditional Chinese medicine compound data set are extracted; based on the chemical component structure characteristics, the principal component efficacy data, toxicity data and time-effect data are obtained to construct an efficacy index set; based on the chemical component structure characteristics and the efficacy index set, the auxiliary added value of the component efficacy is evaluated;

[0011] The component data in the traditional Chinese medicine compound data set is characterized and analyzed to obtain the size accuracy of the compound components; the size accuracy of the compound components, the structural characteristics of the components, and the set of efficacy indicators are used to establish an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound components; based on the experimental results, the proportion of the compound Chinese medicine ingredients is adjusted to maximize the efficacy of each component data.

[0012] As a further optimization scheme of the present invention, data on spleen-tonifying ingredients in a traditional Chinese medicine compound is collected to obtain a traditional Chinese medicine compound data set, including:

[0013] Collect multiple component data of a Chinese medicine compound for tonifying the spleen and resisting ulcerative colitis, and extract the chemical component data of the Chinese medicine compound to retain the effective Chinese medicine component data; the effective Chinese medicine component data includes Codonopsis pilosula, Atractylodes macrocephala, Poria cocos, Pulsatilla scutellaria, Taraxacum mongolicum, Sanguisorba officinalis, Bletilla striata and Herba Lycopodii;

[0014] Performing physical analysis on the effective Chinese medicine component data, and recording the smell, dosage and ratio of multiple components to obtain Chinese medicine compound component information data;

[0015] According to the Chinese medicine compound ingredient information data, combined with the input clinical trial data, to obtain the basic efficacy information data of the Chinese medicine compound;

[0016] Integrate the effective Chinese medicine ingredient data, Chinese medicine compound ingredient information data and Chinese medicine compound basic efficacy information data to obtain the Chinese medicine compound dataset.

[0017] As a further optimization scheme of the present invention, each data item in the Chinese medicine compound data set is integrated and encoded into sequence data, and the sequence data is trained to construct a compound screening model to output the results of identifying the efficacy of the components and the effective parts of the components of the Chinese medicine compound, including:

[0018] The data items in the Chinese medicine compound data set are supplemented or eliminated without duplication or missing items to obtain the pre-processed data items; the pre-processed data items are formatted in a unified format, and characteristic data with predictive power of the Chinese medicine compound are extracted, wherein the characteristic data with predictive power are Chinese medicine ingredients, proportions, medicinal properties and meridians;

[0019] The effective Chinese medicine component data except Bletilla striata is extracted and separated by a system solvent method to obtain effective component part information data of the Chinese medicine compound components;

[0020] Encoding the characteristic data of each Chinese medicine compound data item and the effective part information data of the components into a characteristic vector sequence data; to obtain a plurality of sequence data; performing normalization processing on the plurality of sequence data to obtain a sequence data set;

[0021] The sequence data set is input into the compound screening model; the compound screening model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and the intermediate representation data of multiple hidden layers are transmitted to the output layer, and the output layer outputs the identification results of the efficacy of the ingredients and the effective parts of the ingredients of the screened Chinese medicine compound.

[0022] As a further optimization scheme of the present invention, based on the recognition results of the compound screening model, the chemical component structure characteristics of the principal component data in the traditional Chinese medicine compound data set are extracted, including:

[0023] According to the recognition result of the compound screening model, principal component analysis and recognition are performed to obtain the principal component data of the Chinese medicine compound; the principal component data of the Chinese medicine compound is Bletilla striata;

[0024] Obtaining chemical structure characteristic data of the main components of the traditional Chinese medicine compound; the chemical structure characteristic data includes chemical molecular formula, molecular weight, functional group and stereo configuration;

[0025] According to the chemical structure characteristic data of the main components, the main components are extracted, separated and purified so that the components of the Chinese herbal medicine compound can play the role of both medicine and auxiliary.

[0026] As a further optimization scheme of the present invention, based on the structural characteristics of the chemical components, the main component efficacy data, toxicity data and time-effect data are obtained to construct a set of efficacy indicators, including:

[0027] According to the identification result of the compound screening model, the initial efficacy data of the main component is obtained; and based on the main component extraction, the actual proportion data of the main component in the traditional Chinese medicine compound is obtained to obtain the actual efficacy data;

[0028] Based on the actual efficacy data, the human body system and the main component dose analysis are performed to obtain the main component toxicity analysis results of the Chinese medicine compound, and the toxicity analysis results are the efficacy and side effect data; according to the analysis results, the main component side effect information data is obtained to ensure the safe use of the Chinese medicine compound;

[0029] Through pharmacological analysis, the efficacy analysis results of the Chinese medicine compound in N time periods after opening are obtained, and N={n 1 、n 2 ,…,n i The drug effects at multiple time points of} are recorded and integrated into a time-effect data set; where n i Represents the i-th time node n.

[0030] As a further optimization scheme of the present invention, according to the analysis results, the main component side effect information data is obtained to ensure the safe use of the Chinese medicine compound, including:

[0031] Obtaining the dosage values ​​of side effects of ingredients for the general population and special populations, wherein the special populations include the metabolism, absorption, and excretion capacity data of pregnant women, lactating women, children, the elderly, and patients with liver and kidney dysfunction; using network pharmacology to analyze the targets of the main ingredients in the traditional Chinese medicine compound according to the dosage values ​​of side effects of ingredients for the special populations, wherein the targets include enzymes, receptors, and signaling pathways;

[0032] Based on the target information of the principal component, ; to obtain the interaction between the main component dose and the human system; where E represents the effect caused by the drug; E max is the maximum effect that the drug can achieve; C is the concentration of the drug; is the drug concentration required to produce the h-potency effect; n is the Hill coefficient, describing the slope of the curve; the slope of the curve is related to the cooperativeness of the main component of the traditional Chinese medicine compound to bind to the receptor;

[0033] Based on the interaction between the main component dose and the human body system, the preset dose of the main component of the traditional Chinese medicine compound is subjected to allergen screening to obtain the allergic factor of the preset dose of the main component; according to the allergic factor data combined with the interaction between the main component dose and the human body system, ; To obtain the risk assessment index of the main component drug dosage information for special populations;

[0034] When RI>1, it means that the main ingredient dose poses a risk to special populations. In the formula, RI represents the main ingredient drug dose information risk assessment index; T represents the side effect parameter; Q represents the number of allergic factors; and S represents the safe dose limit.

[0035] As a further optimization scheme of the present invention, based on the interaction between the main component dosage and the human body system, the preset dosage of the main component of the traditional Chinese medicine compound is subjected to allergen screening processing to obtain the allergic factors of the preset dosage of the main component, including:

[0036] Obtain the dosage range of the main ingredients in the Chinese medicine compound, including the effective dose or the dose that causes adverse reactions; organize the known dose-response relationship data based on the side effect dose of the main ingredients in the Chinese medicine compound, including the activity threshold and allergic threshold of the ingredients;

[0037] Based on the dose-response relationship data, the main components of the Chinese herbal compound were used to treat the immune cell model of allergic constitution at different dose concentrations to verify the allergic reaction threshold and mechanism of action;

[0038] Record the incidence of allergic reactions in small-scale clinical trials and validate the predicted allergenic factors to obtain validation results;

[0039] According to the validation results, the dose threshold and the category of allergic factors are revised and updated; the key molecular features that cause allergic factors are extracted, and the key molecular features include molecular structure fragments and specific metabolites; they are classified and identified as direct antigens or indirect antigens;

[0040] Based on the experimental and calculation results, the preset dosage range of the main ingredients is adjusted to reduce the risk of allergic reactions.

[0041] As a further optimization scheme of the present invention, the auxiliary added value of the component efficacy is evaluated according to the structural characteristics of the chemical components and the efficacy index set, including:

[0042] Based on the chemical composition structure characteristics and efficacy index set of the main components of the traditional Chinese medicine compound, the quantitative relationship between the molecular structure characteristics of the chemical components and the efficacy index is analyzed by establishing the efficacy-structure association mechanism;

[0043] pass , to obtain the enhancement effect of the main component and other components in the traditional Chinese medicine compound on the efficacy; where, is the enhancement effect index, It is the efficacy of the compound combination. It is the sum of the effects of the individual ingredients. >1, indicating that the combined effect is greater than the total effect of a single component, indicating an enhanced effect;

[0044] According to the enhancing effect of the main component and other components on the efficacy, it is obtained that the main component has an auxiliary added value of efficacy, and the added value is ≥1.

[0045] As a further optimization scheme of the present invention, characterization analysis is performed on the component data in the traditional Chinese medicine compound data set to obtain the size accuracy of the compound components, including:

[0046] Based on the efficacy index set, ; To obtain the nonlinear relationship between the data items of the components;

[0047] In the formula, D represents the current molecular size of each data item of the Chinese medicine compound; D 0 Indicates the size value with the highest efficacy; K 1 , K 2 , K 3 , K 4 represents the model parameters, so that each data item is fitted and optimized; A represents the binding affinity parameter of each data item; S represents the efficacy enhancement value of the Chinese medicine compound; the size value with the highest efficacy is obtained by Get; where, Represents the shortest distance between the i-th atom and the j-th atom in a molecule;

[0048] Based on the nonlinear relationship between the data items of the components, ; To obtain the molecular size accuracy of each component when the Chinese medicine compound has the maximum efficacy; where, It represents the molecular size accuracy function of each component at the time of maximum efficacy.

[0049] As a further optimization scheme of the present invention, the size accuracy, component structure characteristics and efficacy index set of the compound components are established by establishing an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound components, including:

[0050] Based on the size accuracy, structural characteristics and efficacy index set of compound ingredients, an animal pharmacology experimental model is established; the Chinese medicine compound is screened according to the size accuracy, structural characteristics and efficacy index set of compound ingredients to obtain the screened Chinese medicine compound;

[0051] The screened Chinese herbal compound is tested through animal pharmacology experimental models, and according to the obtained Chinese herbal compound information data, ; to obtain the synergistic effect between the components of the Chinese medicine compound; where I represents the overall drug interaction index; CI is the synergistic index of the single drug combination, which is obtained by comparing the ratio of the combined efficacy to the efficacy of each drug alone; m is the number of drugs in the combination;

[0052] Based on the synergistic effects among the ingredients of traditional Chinese medicine compound, the dosage combination of different ingredients and their contribution to the efficacy are evaluated, and the dosage of each ingredient is adjusted in real time. At the same time, the toxic and side effects of each ingredient are taken into consideration to optimize the formula to maximize the efficacy and minimize the side effects.

[0053] The present invention has at least the following beneficial effects: the present invention first needs to collect and integrate data on spleen-tonifying ingredients of traditional Chinese medicine compound from multiple sources to construct a data set containing multiple information such as ingredients, efficacy, indications, etc.; by collecting a large amount of traditional Chinese medicine compound ingredient data, a relatively complete data set can be established to ensure that subsequent model training has sufficient sample data support; the obtained data set should include detailed information such as efficacy, pharmacological mechanism, chemical composition, clinical application, etc., which is helpful to fully understand the efficacy of compound traditional Chinese medicine; data collection is not limited to traditional literature, but can also be integrated in combination with modern databases (such as traditional Chinese medicine ingredient databases, clinical data, etc.) to improve data quality.

[0054] The various ingredients and efficacy information in the data are converted into a serialized format, and machine learning methods are used to train the compound screening model. The task of the model is to identify which ingredients have a positive effect on the efficacy and predict their effective parts; by integrating various types of data, the model can extract the most influential features on the efficacy from the compound ingredients, such as molecular structure, chemical properties, target information, etc.; by analyzing the trained model, the contribution of different ingredients and different parts to the efficacy can be revealed, helping us understand the mechanism of action of traditional Chinese medicine compound; the compound screening model can efficiently screen out those traditional Chinese medicine ingredients with higher efficacy, providing a theoretical basis for subsequent optimization.

[0055] New Chinese medicine compound data are input into the trained compound screening model, which outputs the efficacy value of each ingredient and its possible effective part. The newly input data can update the model in real time, helping researchers to quickly identify new ingredients and evaluate their efficacy. The screening model can accurately predict the efficacy of each ingredient and provide direction for subsequent pharmacological research through the identification of effective parts. It can process large amounts of compound data and reduce the time and cost of manual screening.

[0056] Based on the active ingredients identified by the model, the chemical structure characteristics of these ingredients are extracted, such as molecular formula, molecular weight, functional group, chemical reaction activity, etc.; by extracting the structural characteristics of chemical components, the potential pharmacological effects of each active ingredient can be accurately understood at the molecular level; structural feature support is provided for subsequent efficacy and toxicity data analysis, forming a more complete component database; through the analysis of structural characteristics, it is helpful to deeply understand the relationship between molecular structure and efficacy, and provide a basis for drug design and optimization.

[0057] Based on the structural characteristics of chemical components, combined with existing efficacy data, toxicity data, etc., a set of efficacy indicators is constructed; the efficacy indicator set can comprehensively evaluate the effects of the ingredients, including efficacy, toxicity and timeliness, to avoid misjudgment caused by a single indicator; it provides data support for efficacy optimization and facilitates the formulation of more scientific Chinese medicine compound ingredient ratio schemes; through toxicity data, it can provide a scientific basis for drug safety assessment and reduce drug side effects.

[0058] Based on the constructed set of efficacy indicators and the structural characteristics of the ingredients, the auxiliary efficacy added value of each ingredient in the compound is evaluated; the role of each ingredient in the compound is evaluated, not only considering the efficacy of a single ingredient, but also revealing the synergistic effect between the ingredients in the compound; it helps to formulate a ratio between ingredients to maximize efficacy and reduce side effects, and extracts Bletilla striata, the main ingredient of the Chinese medicine compound, as a good matrix for the colon drug release system, playing a "dual-purpose drug and auxiliary" role.

[0059] By characterizing the ingredients of the compound, calculating the size accuracy of each ingredient, and quantifying the physicochemical properties of the compound ingredients, a more detailed reference can be provided for the optimization of drug efficacy. The size accuracy information of the ingredients can help discover which ingredients may be more easily absorbed in the body or more easily bind to the target, and by extracting and compressing the volume of the Chinese medicine compound ingredients while maintaining the drug efficacy, it can meet the requirements of modern preparations.

[0060] Based on the above data, animal experimental models are used to verify the synergistic effects and efficacy contributions of the ingredients. The experiment can detect the effects of different ratios of the compound on the pharmacological indicators in animals; the efficacy of the compound is verified through animal experiments to ensure the reliability of the model prediction results; experimental data can reveal the synergistic effects between the ingredients in the compound, which helps to optimize the formula.

[0061] According to the experimental results, the ratio of the ingredients in the compound is adjusted to maximize the efficacy of the compound; by finely adjusting the ratio of the ingredients in the compound, the overall efficacy of the compound can be improved and the synergistic effect between the ingredients can be ensured; according to the needs of different patients or different indications, the ratio of the ingredients is adjusted to achieve the best therapeutic effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a flow chart of a method for screening a compound Chinese medicine based on machine learning provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0064] This embodiment provides a method for screening a Chinese medicine compound based on machine learning, and the specific implementation method is as follows:

[0065] like Figure 1 As shown, a method for screening a Chinese medicine compound based on machine learning comprises the following steps:

[0066] Step 1, collecting the spleen-tonifying component data of the traditional Chinese medicine compound to obtain the traditional Chinese medicine compound data set;

[0067] Step 2, integrating and encoding each data item in the Chinese medicine compound data set into sequence data, training the sequence data to construct a compound screening model, so as to output the identification results of the efficacy of the components and the effective parts of the components of the Chinese medicine compound;

[0068] Step 3, inputting at least one data item in the newly entered Chinese medicine compound data set into the compound screening model to output the efficacy of the corresponding Chinese medicine compound components and the identification results of the effective parts of the components after screening;

[0069] Step 4, based on the recognition results of the compound screening model, extract the chemical component structure characteristics of the principal component data in the traditional Chinese medicine compound data set; based on the chemical component structure characteristics, obtain the principal component efficacy data, toxicity data and time-effect data to construct an efficacy index set; according to the chemical component structure characteristics and the efficacy index set, evaluate the auxiliary added value of the component efficacy;

[0070] Step 5, characterize and analyze the component data in the traditional Chinese medicine compound data set to obtain the size accuracy of the compound components; the size accuracy of the compound components, the component structure characteristics and the efficacy index set are used to establish an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound components; based on the experimental results, adjust the ratio of the compound Chinese medicine ingredients to maximize the efficacy of each component data.

[0071] In this embodiment, in step 1, data of Chinese herbal compound prescriptions for invigorating the spleen are collected, including ingredients, proportions, pharmacological effects and related experimental results; the data are derived from literature and databases (such as Chinese medicine databases), including but not limited to: using modern key technologies of Chinese medicine, extracting and separating the effective parts and main effective ingredients of the compound Chinese medicine for invigorating the spleen (Codonopsis pilosula, Atractylodes macrocephala, Poria cocos, Pulsatilla scutellaria, Taraxacum mongolicum, Sanguisorba officinalis, Bletilla striata, Herba Lycopodii) to compress the volume of Chinese medicine; collating the data items of the obtained Chinese herbal compound prescription ingredients, and ensuring the integrity and consistency of the data; constructing a systematic Chinese herbal compound prescription data set, which provides a basis for subsequent analysis; through data accumulation and integration, the common ingredients and rules of Chinese medicine for invigorating the spleen can be clarified, providing comprehensive information support for model establishment.

[0072] Step 2 converts the TCM compound data set obtained in step 1 into sequence data that can be understood by machine learning algorithms (such as one-hot encoding, molecular structure feature encoding) through encoding; trains machine learning or deep learning models (such as neural networks, random forests) to predict the efficacy and effective parts of the compound; the model learns the complex associations between data through input and output relationships; quickly screens effective TCM compounds, reducing the time and cost of experimental screening; by encoding data, the potential relationships between TCM ingredients can be captured more systematically, providing a new perspective for efficacy research.

[0073] In step 3, the new TCM compound data is input into the compound screening model trained in step 2; the model outputs prediction results, including the main efficacy of the compound and the site of action of the active ingredient; the compound design can be adjusted according to the results or the accuracy of the model prediction can be further verified; a tool for rapid evaluation of new TCM compounds is provided; time is saved for drug development, and accurate ingredient identification results are provided to provide guidance for experimental verification.

[0074] Step 4 extracts the chemical structure characteristics (such as molecular weight, functional group distribution, and 3D spatial structure) of the main components based on the compound screening model identification results obtained in step 3; establishes a set of efficacy indicators to quantify the efficacy, toxicity, time-effect and other information of the components; combines chemical characteristics with efficacy data to analyze and evaluate the efficacy contribution value and potential added value of each component; systematically analyzes the role of the main components of the traditional Chinese medicine compound and clarifies its efficacy contribution; establishes a quantitative evaluation system to help screen and optimize high-efficiency and low-toxic traditional Chinese medicine compounds; and improves the scientificity and rationality of compound drug development.

[0075] Step 5: Based on step 4, characterization techniques (such as mass spectrometry, nuclear magnetic resonance, and X-ray crystallography) are used to analyze the physical and chemical properties of the compound ingredients; an animal pharmacology experimental model is established to verify the synergistic effect of the compound ingredients and the contribution of each ingredient to the overall efficacy; based on the experimental data, the ingredient ratio is optimized to maximize the efficacy; the synergistic mechanism between the ingredients in the compound is revealed and the compatibility ratio is optimized; the experimental data provides reliable support for the modernization of traditional Chinese medicine research and promotes the international development of compound drugs; through synergistic effect research, the potential role of traditional Chinese medicine compound can be further explored, providing a theoretical basis for clinical use.

[0076] Through data-driven and model prediction, the screening and optimization cycle of traditional Chinese medicine compound prescriptions can be greatly shortened; machine learning technology can be used to make traditional Chinese medicine research more scientific and quantitative; and based on toxicity assessment and efficacy optimization, the risk of adverse reactions can be reduced and the efficacy can be improved; the mechanism of traditional Chinese medicine compound prescriptions can be revealed from the molecular and structural levels, providing new ideas for screening traditional Chinese medicine compound prescription data for the modernization of traditional Chinese medicine and the development of maximum efficacy;

[0077] In a preferred embodiment of the present invention, step 1 comprises the following steps:

[0078] Step 11, collecting multiple component data of the Chinese medicine compound for tonifying the spleen and resisting ulcerative colitis, and extracting the chemical component data of the Chinese medicine compound to retain the effective Chinese medicine component data; the effective Chinese medicine component data includes Codonopsis pilosula, Atractylodes macrocephala, Poria cocos, Pulsatilla scutellaria, Taraxacum mongolicum, Sanguisorba officinalis, Bletilla striata and Herba Lycopodii;

[0079] Step 12, physically analyzing the effective Chinese medicine component data, and recording the smell, dosage and ratio of multiple components to obtain Chinese medicine compound component information data;

[0080] Step 13, obtaining basic efficacy information data of the Chinese medicine compound according to the Chinese medicine compound ingredient information data and in combination with the entered clinical trial data;

[0081] Step 14, integrating the effective Chinese medicine component data, the Chinese medicine compound ingredient information data and the Chinese medicine compound basic efficacy information data to obtain the Chinese medicine compound data set.

[0082] In the embodiment of the present invention, in step 11, the ingredients of the Chinese herbal compound related to tonifying the spleen and resisting ulcerative colitis are collected, with a focus on the Chinese herbal ingredients Codonopsis pilosula, Atractylodes macrocephala, Poria cocos, Pulsatilla scabra, Taraxacum mongolicum, Sanguisorba officinalis, Bletilla striata and Herba Lycopodiellae, which are believed to have a certain effect on the treatment of colitis; the compound formula containing these key ingredients is screened out from a large number of literature or databases to ensure that the selected ingredients are effective and relevant to the treatment of ulcerative colitis; the chemical characteristics of each active ingredient are extracted and confirmed through chemical analysis (such as gas chromatography-mass spectrometry technology, etc.) to ensure that the effective drug ingredients are retained in the compound; through comprehensive literature research and database query, it is ensured that the selected ingredients have empirical support in the efficacy of anti-ulcerative colitis; the key active ingredients in the compound are clearly extracted, which is conducive to in-depth analysis of the Chinese herbal ingredients, thereby effectively evaluating their efficacy.

[0083] Step 12: Analyze the physical properties of the effective Chinese medicine component data obtained in step 11, such as smell, color, taste, solubility, etc., which are traditional methods for evaluating the quality and effectiveness of Chinese medicine; record the basic information of each effective ingredient, such as the standard dosage and ratio requirements of each medicinal material. This step can be carried out through laboratory analysis, combined with literature standards and pharmacopoeia regulations; study the proportion and function of each ingredient in the compound, explore their compatibility rules and contribution to the overall efficacy; record the physical properties of Chinese medicine to provide a basis for subsequent compound ratio optimization and dosage form research and development; provide support for maximizing the efficacy of the compound and help determine the optimal ratio of each ingredient in the compound; combine traditional physical analysis methods with modern science and technology to provide data support for the modernization of Chinese medicine.

[0084] Step 13 will collect and integrate the efficacy data of the Chinese herbal compound in clinical trials in step 12, including clinical trial results such as efficacy observation, pathological research, and patient feedback; extract efficacy information related to the Chinese herbal compound through analysis of clinical trial data, including the effect of relieving colitis symptoms and the therapeutic effect on ulcerative colitis; quantify the efficacy data, such as using efficacy scores, treatment response rates, symptom improvement and other indicators, so as to provide a quantifiable efficacy basis for subsequent analysis; verify the efficacy through clinical trials, ensure the clinical application basis of the Chinese herbal compound, and enhance the credibility of efficacy evaluation; combine clinical efficacy data to further optimize the use and formulation of the compound, and improve its clinical application effect and safety; provide a scientific basis for the efficacy of the Chinese herbal compound, and contribute to its wider application in modern medicine.

[0085] Step 14 integrates the effective Chinese medicine ingredient data (such as ingredient name, chemical structure, etc.), ingredient information data (such as smell, ratio, etc.), and efficacy information data (such as clinical efficacy data) obtained in steps 11, 12, and 13 to form a complete Chinese medicine compound data set; the collected data is structured and a database or data model is constructed to ensure convenient access and analysis of the data. For example, a relational database or data warehouse can be used to store this information; based on the integrated data set, further analysis is performed, such as efficacy prediction, ingredient synergy analysis, etc., to help optimize the ingredient ratio and therapeutic effect of the compound; a systematic and comprehensive data foundation is provided for the research and optimization of Chinese medicine compound. By integrating information from multiple dimensions, researchers can have a more comprehensive understanding of the composition and efficacy of the compound; a structured data set is helpful for subsequent data mining work, which can reveal the interaction between ingredients, efficacy mechanism, etc., and further optimize the compound; it provides rich empirical data for the scientific research and clinical application of Chinese medicine compound and promotes the process of modernization of Chinese medicine.

[0086] Through this series of steps, we can not only provide scientific support for the traditional Chinese medicine compound for ulcerative colitis, but also provide a data basis for the optimization, standardization and modernization of the traditional Chinese medicine compound.

[0087] In a preferred embodiment of the present invention, the above step 2 includes the following steps:

[0088] Step 21, completing or eliminating the data items in the Chinese medicine compound data set without duplication or missing items to obtain the pre-processed data items; formatting the pre-processed data items into a unified format, and extracting the characteristic data with predictive power of the Chinese medicine compound, wherein the characteristic data with predictive power is the Chinese medicine ingredients, ratio, medicinal properties and meridians;

[0089] Step 22, using the system solvent method to extract and separate different material parts of the Chinese medicine compound components, and obtain the effective part information data of the components;

[0090] Step 23, encoding the characteristic data of each Chinese medicine compound data item and the effective part information data of the ingredients into a characteristic vector sequence data; to obtain a plurality of sequence data; and normalizing the plurality of sequence data to obtain a sequence data set;

[0091] Step 24, input the sequence data set into the compound screening model; the compound screening model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and transmits the intermediate representation data of multiple hidden layers to the output layer, and the output layer outputs the identification results of the efficacy of the ingredients and the effective parts of the ingredients of the screened Chinese medicine compound.

[0092] In an embodiment of the present invention, firstly, each data item in the Chinese medicine compound data set is checked through step 21, duplicate data is removed and missing items are processed. Missing items can be supplemented by interpolation technology, or data containing missing values ​​can be selected to ensure the quality and integrity of the data; in order to unify the data format, all data items will be converted to the same standard format, which is helpful for subsequent analysis and model processing; feature data with predictive power are extracted from the Chinese medicine compound. These feature data include the ingredients of Chinese medicine, the ratio of ingredients, medicinal properties (such as cold and hot, warm and cool, etc.) and meridians (i.e., the target organs or systems of drug action). These features can provide an in-depth understanding of drug efficacy, which is helpful for subsequent analysis and model training; improve the quality and consistency of data, and lay the foundation for subsequent analysis and modeling; through feature extraction, focus on the key data that affects drug efficacy, so that the model can better predict and screen.

[0093] Step 22 uses a system solvent method to extract the components of the Chinese medicine compound in step 21 and separate them into effective parts of different Chinese medicine components, with the purpose of separating the effective components of Chinese medicine by using different solvents. This method can effectively extract different parts of the substance, making the analysis of drug efficacy more accurate; the components of different parts of the substance extracted will be further analyzed for their role in drug efficacy. The effective information of these parts plays an important role in predicting drug efficacy and identifying the biological activity of the components; the system solvent method can extract multiple components from the compound, providing support for a deeper understanding of its pharmacological effects; determining the effective parts of the components helps to clarify which components have stronger drug efficacy for the treatment of specific diseases.

[0094] Step 23 encodes the characteristic data of each data item and effective part of the Chinese medicine compound to form a feature vector sequence. Each compound data will be converted into a vector, containing information such as drug ingredients, ratios, medicinal properties, meridians, etc.; through normalization processing, all feature vectors are converted to a unified scale so that the differences between the features will not affect the training effect of the model. Normalization helps to improve the training efficiency and stability of the model; feature vector encoding simplifies the data format, making it suitable for input into the neural network model; normalization helps to improve the convergence speed and stability of the model and reduce the bias of certain features on the results.

[0095] Step 24 inputs the data obtained in step 23 into the compound screening model and performs efficacy prediction; the model consists of multiple layers, including an input layer, multiple hidden layers, and an output layer. The input layer receives the normalized feature vector data and passes it to the hidden layer. The hidden layer is responsible for abstracting and extracting the potential rules in the data, and finally the output layer outputs the identification results of the compound's component efficacy and the effective part of the component; the intermediate representation data of each hidden layer will be passed to the next layer, and the key information of the compound's efficacy will be gradually extracted and refined; the final efficacy and effective part identification results are obtained through the output layer, providing a basis for the optimization and further application of the Chinese medicine compound; the neural network model can process complex feature data, extract implicit rules in the data, and help accurately predict the efficacy; multi-level hidden layer processing helps to capture more detailed feature information and improve the model's prediction ability; the output efficacy and effective part results can provide a scientific basis for the precise application of the Chinese medicine compound and optimize the use effect of the compound.

[0096] An efficient model screening process was established through data preprocessing, feature extraction, component separation and coding. This process not only provides an in-depth understanding of the ingredients and efficacy of Chinese medicine compound prescriptions, but also helps to accurately identify the key active ingredients and their effective parts in the compound prescriptions, providing theoretical support and technical guarantee for the innovation and application of Chinese medicine.

[0097] In a preferred embodiment of the present invention, in the above step 4, based on the recognition result of the compound screening model, extracting the chemical component structure characteristics of the principal component data in the traditional Chinese medicine compound data set comprises the following steps:

[0098] Step 41, according to the recognition result of the compound screening model, principal component analysis and recognition are performed to obtain principal component data of the Chinese medicine compound; the principal component data of the Chinese medicine compound is Bletilla striata;

[0099] Step 42, obtaining chemical structure characteristic data of the main component of the traditional Chinese medicine compound; the chemical structure characteristic data includes chemical molecular formula, molecular weight, functional group and stereo configuration;

[0100] Step 43, extracting, separating and purifying the main component according to the chemical structure characteristic data of the main component, so that the Chinese medicine compound component can play the dual role of medicine and auxiliary.

[0101] In an embodiment of the present invention, step 41 reduces the dimension and analyzes the recognition results of the compound screening model by the principal component analysis (PCA) method; PCA is a statistical technique that can help identify the main components in the data set and convert the data from high dimensions to lower dimensions. It reduces the redundant information in the data and finds the principal components that best represent the data characteristics; through PCA analysis, the system can identify the principal components in the compound, that is, the components that can most effectively represent the efficacy of the compound. In this step, the key ingredient Bletilla striata in the traditional Chinese medicine compound is mainly identified; PCA helps reduce the dimension and extract key features, so that the model can focus on the most representative ingredients, improving the efficiency and accuracy of the analysis; identifying the principal components in the compound (such as Bletilla striata) helps to clarify the key active ingredients of the compound and provide a basis for efficacy analysis and ingredient optimization.

[0102] Step 42 extracts relevant chemical structure characteristic data based on the chemical structure of the main component of the Chinese medicine compound obtained in step 41; the chemical structure characteristics include the following aspects: describing the type and quantity of each element in the molecule, which is the basic information of the chemical molecule; reflecting the relative mass of the molecule, which is usually related to its pharmacological activity and stability; the specific atomic groups involved in the reaction in the chemical molecule, which usually determine the chemical reactivity of the molecule; the three-dimensional structure of the molecule in space, and the stereochemistry have an important influence on the efficacy and interaction of the molecule.

[0103] This stage usually uses a variety of chemical analysis techniques, such as nuclear magnetic resonance (NMR), mass spectrometry (MS), infrared spectroscopy (IR), X-ray crystallography, etc., to help accurately obtain the above-mentioned chemical structure characteristics; through detailed chemical structure analysis, we can have a deeper understanding of the molecular characteristics of the main components, providing support for subsequent pharmacological research and optimization; chemical structure characteristics such as functional groups and stereo configurations are crucial to the interaction between molecules and targets, helping to improve the accuracy of efficacy and side effect predictions.

[0104] Step 43: According to the chemical properties of the main component in step 42, select appropriate solvents and extraction methods (such as solvent extraction, ultrasonic extraction, frozen centrifugal extraction, etc.) to extract the main component in the compound. This process aims to obtain a high concentration of the target component; different components are separated from the extract by liquid chromatography, gas chromatography, gel filtration and other technologies. This step helps to separate the main component with higher purity from the complex Chinese medicine compound; crystallization, thin layer chromatography and other methods are used to further purify the separated main component. The purified component has higher pharmacodynamic stability and helps to more accurately evaluate its biological activity in the experiment; the extraction and separation purification process can ensure the acquisition of high-purity effective components and provide reliable samples for subsequent pharmacodynamic evaluation; the purified main component can not only better exert its pharmacodynamics, but also reduce the interference of other components, and improve the safety and efficacy of the drug; the purified Chinese medicine component can be used for the study of both drug and auxiliary, that is, it can not only play a therapeutic role, but also be applied to health care or other auxiliary treatment fields.

[0105] Through the processing from step 41 to step 43, the main components can be accurately identified from the complex Chinese medicine compound and their chemical structures can be further analyzed, providing strong support for optimizing drug efficacy, improving efficacy and reducing side effects; the process of extraction, separation and purification can not only obtain high-purity components, but also help to transform the valuable experience of traditional Chinese medicine into scientific and modern pharmaceutical technology, and promote innovation and standardization in the field of Chinese medicine.

[0106] In a preferred embodiment of the present invention, in the above step 4, based on the structural characteristics of the chemical components, obtaining the main component efficacy data, toxicity data and time-effect data to construct the efficacy index set includes the following steps:

[0107] Step 44, obtaining initial efficacy data of the main component based on the recognition result of the compound screening model; and obtaining actual proportion data of the main component in the traditional Chinese medicine compound based on the main component extraction to obtain actual efficacy data;

[0108] Step 45, based on the actual efficacy data, the human body system and the main component dose analysis are performed to obtain the main component toxicity analysis results of the Chinese medicine compound, and the toxicity analysis results are the efficacy and side effect data; according to the analysis results, the main component side effect information data is obtained to ensure the safe use of the Chinese medicine compound;

[0109] Step 46, through pharmacological analysis, to obtain the efficacy analysis results of the Chinese medicine compound for N time periods after opening, and N={n 1 、n 2 ,…,n i The drug effects at multiple time points of} are recorded and integrated into a time-effect data set; where n i Represents the i-th time node n.

[0110] In the embodiment of the present invention, step 44 first obtains preliminary efficacy data of the main components of the Chinese medicine compound according to the identification results of the compound screening model; this is usually data obtained through in vitro experiments or animal experiments, reflecting the influence of the main components on specific biomarkers, pathological conditions or cells; acquisition of actual proportion data of the main components: after extracting the main components in the compound, it is necessary to determine the actual content proportion of the main components in the compound through further analysis (such as chromatography analysis, mass spectrometry analysis, etc.). By determining the proportion of each main component in the compound, its efficacy performance can be accurately evaluated; based on the proportion data and efficacy data of the main components, the actual efficacy performance of the main components in the compound can be calculated. In other words, the actual efficacy data is the comprehensive efficacy result after considering the content of each component; by obtaining the preliminary efficacy data of the main components, it is possible to preliminarily evaluate whether the compound has the expected therapeutic effect; by accurately calculating the actual proportion of the main components, the efficacy performance of the compound can be more accurately predicted to avoid the effect deviation caused by unreasonable component proportions; this step can provide data support for subsequent efficacy optimization to ensure that the compound has good efficacy when applied.

[0111] Step 45 is to study the toxic effects of the main component at different doses through dose-response relationship analysis of the human system or animal model. It usually includes research on acute toxicity, chronic toxicity, teratogenicity, carcinogenicity, etc.; use in vitro cell experiments to evaluate the physiological responses and potential toxicity of the main component at different doses; determine the safety range and toxicity threshold of the main component through dose analysis; based on the results of toxicity analysis, identify and quantify the side effects of the main component to form "pharmacodynamic side effect data", which reflects the negative reactions (such as adverse reactions, organ damage, etc.) that may be caused by the main component at different doses; combine the efficacy and side effect data to obtain the side effect information data of the main component, which helps to clarify the safe use range and use restrictions of the Chinese medicine compound; toxicity analysis and side effect evaluation ensure the safety of the compound in clinical applications, help to formulate a reasonable dosage range, and reduce unnecessary drug toxic reactions; determining the side effects of the main component can provide a scientific basis for the safe use of the drug, ensuring that users will not experience adverse reactions due to overdose or improper use during use; side effect information provides a basis for subsequent drug supervision, formulation of medication guidelines, and patient education. Step 452, based on the target information of the main component, through

[0112] Step 46 analyzes the changes in the efficacy of the Chinese medicine compound at multiple time points after opening the package through long-term observation or experiments at different time points based on the efficacy and side effect data obtained in step 45. For example, the efficacy can be observed experimentally at different time points such as 1 hour, 3 hours, 6 hours, and 12 hours after drug administration to monitor the changes in efficacy over time; the performance of the efficacy in different time periods is recorded according to the experimental results. By recording the efficacy data at multiple time points, a time-effect data set is constructed. The time-effect data set contains the efficacy change data at multiple time points, such as N={n1,n2,...,ni}, where ni represents the efficacy at the ith time point; the data at different time points are integrated to analyze the trend of efficacy changes over time, and the persistence, intensity and attenuation of efficacy are evaluated through statistical methods (such as biopharmacological models); through time-effect analysis, data support can be provided for the clinical use of Chinese medicine compound prescriptions, helping doctors and patients understand the sustained effectiveness of drugs and the most suitable dosing interval; time-effect analysis helps understand the changing patterns of the efficacy of Chinese medicine compound prescriptions at different time points, and can evaluate the sustained effect of drugs and the best time for dosing; through the construction of time-effect data, the medication regimen of the compound can be optimized to ensure that the efficacy reaches the maximum effect within the critical time period; time-effect data can predict the attenuation or accumulation of efficacy, providing a scientific basis for formulating appropriate medication intervals and dosing frequencies.

[0113] The analysis in step 44 and step 45 not only helps optimize the efficacy of the drug by combining the efficacy data and toxicity data, but also ensures the safety of the compound and avoids abuse and side effects. The efficacy time analysis in step 46 provides important time dimension data for the clinical use of Chinese medicine compound, ensuring that the drug exerts its effectiveness at the best time and reducing the risk of efficacy attenuation.

[0114] In a preferred embodiment of the present invention, the above step 45 obtains the main component side effect information data according to the analysis results so as to ensure the safe use of the Chinese medicine compound, including the following steps:

[0115] Step 451, obtaining the component side effect dose values ​​of the general population and special populations, wherein the special populations include the metabolism, absorption, and excretion capacity data of pregnant women, lactating women, children, the elderly, and patients with liver and kidney dysfunction; using network pharmacology to analyze the action targets of the main components in the traditional Chinese medicine compound according to the component side effect dose values ​​of the special populations, wherein the action targets include enzymes, receptors, and signal pathways;

[0116] Step 452, based on the target information of the principal component, ; to obtain the interaction between the main component dose and the human system; where E represents the effect caused by the drug; Emax is the maximum effect that the drug can achieve; C is the concentration of the drug; is the drug concentration required to produce the h-potency effect; n is the Hill coefficient, describing the slope of the curve; the slope of the curve is related to the cooperativeness of the main component of the traditional Chinese medicine compound to bind to the receptor;

[0117] Step 453, based on the interaction between the main component dose and the human body system, the preset dose of the main component of the Chinese medicine compound is subjected to allergen screening to obtain the allergic factor of the preset dose of the main component; according to the allergic factor data combined with the interaction between the main component dose and the human body system, ; To obtain the risk assessment index of the main component drug dosage information for special populations;

[0118] Step 454, when RI>1, it means that the main component dose poses a risk to a special population, where RI represents the main component drug dose information risk assessment index; T represents the side effect parameter; Q represents the number of allergic factors; and S represents the safe dose limit.

[0119] In the embodiment of the present invention, in step 451, firstly, the side effect dose response of the general population and special populations (such as pregnant women, lactating women, children, the elderly, patients with liver and kidney dysfunction, etc.) to the main components of the Chinese medicine compound is collected. The physiological differences and metabolic characteristics of special populations (such as drug absorption, distribution, metabolism and excretion capacity) will affect their response to the drug, so these factors need to be specially considered; these data are obtained through clinical trials, literature surveys, laboratory data collection, etc., especially for the drug metabolism, effects and side effects of different populations at the same dose; safer and more effective drug doses can be tailored for different populations (especially special populations); by evaluating the metabolic differences of different groups, adverse reactions caused by improper doses can be avoided.

[0120] Step 452: The mechanism of action of a Chinese medicine compound often involves multiple targets, including enzymes, receptors, and signaling pathways. In step 452, based on the data obtained from step 451, analyze how the main components in the Chinese medicine compound interact with these targets; use pharmacological models to establish a dose-effect relationship model through parameters such as drug concentration (C), maximum effect (Emax), half-maximum effect concentration (EC50), and Hill coefficient to evaluate the physiological effects of the drug. This relationship reflects the connection between the concentration and effect of the drug in the body, which can help understand how different doses of drugs affect the functions of various systems in the human body; can accurately predict the relationship between drug dose and effect through mathematical models, and help formulate more appropriate treatment plans; clarify the effects of drugs on physiological systems at different concentrations to prevent overdose or inefficiency problems.

[0121] Step 453 further screens whether the ingredient triggers an allergic reaction at a specific dose by analyzing the allergenic factors of the main ingredient (such as the tendency of certain ingredients to cause allergic reactions), combined with the interaction between the dosage of the main ingredient and the human system in step 452; allergen screening involves analyzing the immunological response of the ingredients to assess which ingredients are more likely to trigger allergies in certain populations. Through this screening, early warning of potential allergic reactions can be provided to avoid the negative impact of drugs on patients; it refers to potential risk factors that may trigger immune system reactions in drugs or their ingredients, which may be related to the chemical structure, formula or genetic characteristics of the drug; allergic reaction warning: by identifying allergy risks in advance, personalized drug use recommendations can be provided to patients; reduce drug-induced allergic reactions and adverse events, and improve drug safety.

[0122] Step 454 calculates the index based on the risk assessment index (RI) by combining multiple information such as drug dosage, allergic factors, and drug side effect parameters. The risk of the principal component to special populations (such as pregnant women, children, etc.) at a specific dose is evaluated through the formula; when RI>1, it means that the drug dosage has a certain risk for special populations, which usually means that the dose has a large side effect or allergic reaction to the special population, and it is necessary to adjust the dosage or consider other alternative treatment options; through the RI index, the potential risk of drug dosage to special populations can be discovered in time; if the RI value is found to be greater than 1, the drug dosage can be adjusted or the drug can be replaced in time to ensure the safety of the patient; according to the risk assessment index, a more personalized treatment plan can be provided for each patient, thereby improving the efficacy and reducing adverse reactions.

[0123] Through the combination of the above four steps, the safety and effectiveness of Chinese medicine compound prescriptions have been scientifically and systematically evaluated. The working principle and effect of each step help to optimize the dosage for different populations (especially special populations) to avoid side effects and allergic reactions; and understand the mechanism of action and dose-effect relationship of drugs through network pharmacology and mathematical models; and then ensure the safety and effectiveness of drugs through allergy screening and risk assessment; finally, provide the most appropriate drug dosage for special populations based on their metabolic and drug response characteristics.

[0124] In a preferred embodiment of the present invention, the above step 453 performs allergen screening on the preset dose of the main component of the Chinese medicine compound based on the interaction between the main component dose and the human body system to obtain the allergic factors of the preset dose of the main component, including the following steps:

[0125] Step 4531, obtaining the dosage range of the main component in the Chinese medicine compound, including the effective dosage or the dosage that causes adverse reactions; collating the known dose-response relationship data according to the side effect dosage of the main component in the Chinese medicine compound, including the activity threshold and allergic threshold of the component;

[0126] Step 4532, based on the dose-response relationship data, at different dose concentrations of the main component of the Chinese herbal compound, the main component is used to treat the immune cell model of the allergic constitution to verify the allergic reaction threshold and mechanism of action;

[0127] Step 4533, recording the incidence of allergic reactions in a small-scale clinical trial, and verifying the predicted allergic factors to obtain verification results;

[0128] Step 4534, modifying and updating the dose threshold and the category of allergic factors according to the verification results; extracting the key molecular features that cause allergic factors, wherein the key molecular features include molecular structure fragments and specific metabolites; and classifying and identifying them as direct antigens or indirect antigens;

[0129] Step 4535, combining the experimental and calculation results, adjusting the preset dosage range of the main component to reduce the risk of allergic reactions.

[0130] In the embodiment of the present invention, in step 4531, the effective dose and the dose of adverse reactions are determined by clarifying the dose range of the main component in the Chinese medicine compound. This is the basis for drug dose optimization; through clinical research, pharmacological data and known drug reaction models, the minimum effective dose of each component and the dose range that causes adverse reactions can be obtained; the effective dose refers to the dose that can play a therapeutic role, and the adverse reaction dose is the dose that may cause side effects (such as allergic reactions or toxic reactions); by determining the safe dose and effective dose range of each component, it can help to accurately adjust the drug dose in clinical practice and avoid the occurrence of side effects; provide a scientific basis for subsequent research and drug safety evaluation.

[0131] Step 4532 In an in vitro immune cell model, by exposing the main components of the Chinese herbal compound, study its effects on immune cells of allergic constitution (such as eosinophils, T cells, etc.); by observing the response of immune cells, the threshold of allergic reaction at different doses (that is, the immune system will have an allergic reaction after the dose exceeds this point) and the mechanism of action (such as the type of mediated immune response) can be determined; and it also involves technologies such as cytokine detection and immune response marker monitoring; it helps to clarify the critical dose of allergic reaction, that is, at which dose allergic reaction begins to occur, and provide experimental data for dose adjustment; revealing the mechanism of action of the drug or its components can guide further drug optimization and safety research.

[0132] Step 4533 In small-scale clinical trials, by testing different doses of the drug, monitor the allergic reactions of the participants, especially those with allergic constitutions; record the incidence of allergic reactions and compare it with the laboratory's predicted data to verify the accuracy and predictive ability of the allergic factors; through clinical trials, statistical analysis of allergic reactions at different doses can be performed to further confirm whether the prediction of the allergic factors is valid; through the accumulation of clinical data, further verify and revise the drug's allergic factors and dose-response model; provide real-world data to ensure the safety of the drug in different populations.

[0133] Step 4534 amends and updates the dose threshold according to the verification results in clinical trials. This means that in actual applications, the safe dose of the drug and the dose range for allergic reactions will be adjusted according to new evidence; further analyze and extract key molecular features that cause allergic reactions, such as molecular structure fragments of the drug, specific metabolites, etc., to identify potential direct or indirect antigens;

[0134] Direct antigen: refers to the molecule that the drug itself can directly induce an immune response.

[0135] Indirect antigens: molecules that indirectly trigger allergic reactions by changing the state of the immune system.

[0136] By continuously updating the dosage threshold, the risk of adverse drug reactions can be reduced and the treatment effect can be optimized; providing in-depth molecular mechanism data will help develop new drugs or improve existing drugs in the future, thereby reducing the occurrence of allergic reactions.

[0137] Step 4535 adjusts the preset dosage range of the main component based on the experimental and clinical verification results of step 4534, especially to avoid the occurrence of allergic reactions; combines different experimental data and calculation models to optimize the drug dosage to ensure that allergic reactions are minimized while providing therapeutic effects; after optimizing the dosage, it can ensure the safety of the drug, reduce the occurrence of allergic reactions, and improve the adaptability of the drug in different individual groups; through multi-angle data analysis, it helps to formulate personalized treatment plans and improve clinical efficacy.

[0138] The common goal of the above steps is to optimize the dosage range of the main ingredients of the traditional Chinese medicine compound, reduce the risk of allergic reactions, and improve the safety and effectiveness of the drug through systematic experimental and clinical data support. Each step provides data support and scientific basis for the final drug dosage optimization through different experiments, models and clinical verification, ensuring that the drug can take into account both efficacy and safety in actual application.

[0139] In a preferred embodiment of the present invention, the above step 4, based on the chemical component structure characteristics and the pharmacodynamic index set, evaluates the pharmacodynamic auxiliary added value of the component, including the following steps:

[0140] Step 47, based on the chemical component structure characteristics and the pharmacodynamic index set of the main components of the traditional Chinese medicine compound, by establishing a pharmacodynamic-structural association mechanism to analyze the quantitative relationship between the molecular structure characteristics of the chemical components and the pharmacodynamic index;

[0141] Step 48, pass , to obtain the enhancement effect of the main component and other components in the traditional Chinese medicine compound on the efficacy; where, is the enhancement effect index, It is the efficacy of the compound combination. It is the sum of the effects of the individual ingredients. >1, indicating that the combined effect is greater than the total effect of a single component, indicating an enhanced effect;

[0142] Step 49, according to the enhancement effect of the main component and other components on the efficacy, the main component is obtained to have an auxiliary added value of the efficacy, and the added value is ≥1.

[0143] In the embodiment of the present invention, in step 47, the pharmacodynamic mechanism of the main component is revealed through the quantitative relationship between the structural characteristics of the chemical components and the pharmacodynamics; specifically, the relationship between the chemical structural characteristics (such as functional groups, molecular weight, polarity, etc.) of the main components in the Chinese medicine compound and its pharmacodynamic index (such as anti-inflammatory, antioxidant, analgesic, etc.) is analyzed to establish a correlation mechanism between pharmacodynamics and molecular structure; the relationship between molecular structure and pharmacodynamics can be quantified through a quantitative structure-activity relationship (QSAR) model or chemical fingerprint analysis. The QSAR model usually uses mathematical equations to describe the quantitative relationship between molecular structure parameters and pharmacodynamics; by establishing a pharmacodynamic-structure correlation mechanism, it can help identify and screen potential pharmacodynamic components. It provides theoretical support for the discovery and optimization of new drugs; it can quantitatively analyze the complex components in the Chinese medicine compound, so as to accurately predict the pharmacodynamics of the main components in the compound, and help realize the personalized use of drugs; it provides a scientific basis for understanding and explaining the complex mechanism of action of the Chinese medicine compound, making drug research and development more systematic and precise.

[0144] Step 48 analyzes the interaction between the main component and other components in the Chinese medicine compound, especially their enhancement effect on the efficacy; specifically, by comparing the efficacy of the compound combination with the sum of the efficacy of the single components, the synergistic effect of the compound is evaluated. If the efficacy of the compound combination is greater than the sum of the efficacy of the single components, then the compound has an efficacy enhancement effect. When the value is greater than 1, it means that the efficacy of the compound is higher than the sum of the efficacy of the single components, indicating that there is a synergistic enhancement effect between the components of the compound; it helps to evaluate the synergistic effect between different components in the compound, revealing which components have the potential to enhance the efficacy; it can avoid duplication or ineffective superposition of the components of the compound, optimize the composition of the compound; provide a quantitative basis for the scientific compatibility of the Chinese medicine compound, and help maximize the efficacy.

[0145] Step 49 is a further analysis of step 48. According to the enhancement effect of the main component and other components, evaluate whether the main component has auxiliary added value of efficacy; added value refers to the synergistic effect of the main component in the compound. If the added value is ≥1, it means that the main component can not only play an independent efficacy, but also has an enhanced or auxiliary effect in the compound; through this analysis, it can be evaluated whether the role of the main component in the compound helps to improve the efficacy of other components, whether it has a synergistic or auxiliary effect; quantifying the added value of the main component can help determine the value of the main component in the compound and optimize the composition of the compound; determining the auxiliary effect of the main component can help to rationally design drug combinations and maximize the efficacy; provide a basis for the optimization and efficacy evaluation of the main component to ensure that the role of each component in the compound is effective, not redundant or conflicting.

[0146] The common goal of the above steps is to optimize the formula and efficacy of the Chinese medicine compound by quantifying the efficacy-structure relationship, evaluating the enhancement effect of different ingredients, and ultimately determining the added value of the main ingredient in the compound. Each step relies on systematic analysis and data support to achieve this goal:

[0147] Step 47 provides a scientific basis for predicting the efficacy of the main ingredient by constructing the association mechanism between efficacy and structure; Step 48 provides a quantitative analysis for compound compatibility by analyzing the enhancement effect of the main ingredient and other ingredients; Step 49 further verifies the added value of the main ingredient, which helps optimize the drug formula. The collaboration of the three steps helps to improve the efficacy and safety of traditional Chinese medicine compound, ensure the scientific nature of drug combination, and promote the modernization and standardization of traditional Chinese medicine.

[0148] In a preferred embodiment of the present invention, the characterization analysis of the component data in the traditional Chinese medicine compound data set in the above step 5 to obtain the dimensional accuracy of the compound components includes the following steps:

[0149] Step 51, based on the efficacy index set, ; To obtain the nonlinear relationship between the data items of the components;

[0150] In the formula, D represents the current molecular size of each data item of the Chinese medicine compound; D 0 Indicates the size value with the highest efficacy; K 1 , K 2 , K 3 , K 4 represents the model parameters, so that each data item is fitted and optimized; A represents the binding affinity parameter of each data item; S represents the efficacy enhancement value of the Chinese medicine compound; the size value with the highest efficacy is obtained by Get; where, Represents the shortest distance between the i-th atom and the j-th atom in a molecule;

[0151] Step 52, based on the nonlinear relationship between the data items of the components, ; To obtain the molecular size accuracy of each component when the Chinese medicine compound has the maximum efficacy; where, It represents the molecular size accuracy function of each component at the time of maximum efficacy.

[0152] In the embodiment of the present invention, step 51 explores the complex relationship between the data items by establishing a nonlinear relationship model. The relationship between drug efficacy and molecular size is usually not a linear relationship, so it needs to be fitted by a nonlinear model; specifically, the relationship between drug efficacy and molecular size involves the optimization and adjustment of molecular size to achieve the best relationship between it and drug efficacy; the model uses molecular spacing, that is, the shortest distance between the i-th atom and the j-th atom in the molecule, to estimate the molecular size, which is crucial to the relationship between molecular structure and drug efficacy;

[0153] Model optimization: In order to make the relationship between molecular size and drug efficacy as accurate as possible, the model will continuously optimize and adjust parameters (K 1 , K 2 , K 3 , K 4 ), gradually optimize the efficacy prediction results through experimental data or computational prediction; in this process, the ultimate goal is to find the molecular size that maximizes the efficacy (i.e., D0), and use this size to design a Chinese medicine compound with the best efficacy; accurately identify the molecular size with the best efficacy: by determining the molecular size with the highest efficacy (D0), it can be determined which molecular structures have the best correlation with efficacy, thereby providing guidance for drug design; by analyzing the nonlinear relationship between the molecular size and efficacy of each component, those components that may enhance the efficacy can be identified, thereby optimizing the compound formula and improving the overall efficacy; by optimizing the parameter K 1 , K 2 , K 3 , K 4 , the interaction between molecules can be precisely adjusted to achieve the optimal enhancement of drug efficacy and improve the effect of Chinese medicine compound.

[0154] Step 52 establishes a nonlinear relationship between the component data items based on step 51. Step 52 further calculates the molecular size of each component when the efficacy of the Chinese medicine compound is maximized based on these relationships; by optimizing the molecular size, it is ensured that the combination of the components in the compound can exert the maximum efficacy in the interaction; the molecular size accuracy function is used to quantify the accuracy of the molecular size of each component in the compound when the efficacy is maximized. This function reflects the precision requirement of the molecular size of the ingredients, that is, by adjusting the molecular size of each ingredient, the best efficacy can be achieved in the compound combination; through this function, researchers can obtain the ideal molecular size range of the Chinese medicine compound under the maximum efficacy state, and then guide the selection and compatibility of ingredients; in this step, through fine-tuning, the molecular size accuracy of each ingredient is maximized, so as to ensure that the Chinese medicine compound achieves the best balance in efficacy; this precision analysis includes not only the size of the molecule, but also may involve factors that affect the efficacy, such as molecular shape and electronic structure; by accurately calculating the molecular size accuracy of each ingredient at maximum efficacy, the ability to predict and optimize efficacy can be greatly improved; it helps to accurately control the efficacy of the compound by fine-tuning the molecular size, avoid ineffective superposition or adverse interactions between ingredients, and ensure maximum efficacy; the realization of this step is the core of the optimization of Chinese medicine compounds, which helps to combine the theory of traditional Chinese medicine with modern medicinal chemistry and pharmacodynamics research, and promote the scientific and standardized process of Chinese medicine.

[0155] The core working principle of these two steps is to explore the complex relationship between the molecular size and efficacy of each component in the traditional Chinese medicine compound through nonlinear relationship analysis, and to maximize the efficacy by optimizing these relationships. By accurately calculating the molecular size accuracy and efficacy enhancement value, it provides theoretical support for the scientific design and efficacy optimization of traditional Chinese medicine compound, which can help accurately design the combination of compound ingredients with the best efficacy; by adjusting the molecular size and affinity, the efficacy can be enhanced; it provides a quantifiable and systematic method for the modernization and standardization of traditional Chinese medicine, which is conducive to the precise application of traditional Chinese medicine compound.

[0156] In a preferred embodiment of the present invention, the size accuracy, structural characteristics and efficacy index set of the compound ingredients in step 5 are used to establish an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound ingredients, including the following steps:

[0157] Step 53, establishing an animal pharmacology experimental model based on the compound component size accuracy, component structure characteristics and efficacy index set; screening the Chinese medicine compound according to the compound component size accuracy, component structure characteristics and efficacy index set to obtain the screened Chinese medicine compound;

[0158] Step 54, the screened Chinese herbal compound is subjected to an animal pharmacology experimental model, and according to the acquired Chinese herbal compound information data, ; to obtain the synergistic effect between the components of the Chinese medicine compound; where I represents the overall drug interaction index; CI is the synergistic index of the single drug combination, which is obtained by comparing the ratio of the combined efficacy to the efficacy of each drug alone; m is the number of drugs in the combination;

[0159] Step 55, based on the synergistic effect between the ingredients of the Chinese medicine compound, the dosage combination of different ingredients and their contribution to the efficacy are evaluated, and the dosage of each ingredient is adjusted in real time; at the same time, the toxic and side effects of each ingredient are considered to optimize the formula to maximize the efficacy and minimize the side effects.

[0160] In the embodiment of the present invention, step 53 establishes an animal pharmacology experimental model based on the size accuracy, component structure characteristics and efficacy index set of the compound components in step 51 and step 52; the core of this step is to build an animal pharmacology experimental model, which needs to comprehensively consider three key factors:

[0161] Size accuracy of compound ingredients: The molecular size accuracy of each component in the compound has been accurately calculated in the previous steps. These data provide a structural basis for subsequent animal pharmacology experiments;

[0162] Structural characteristics of ingredients: The molecular structure of each Chinese medicine ingredient has a great influence on its efficacy, so it is necessary to consider structural characteristics such as molecular shape, chemical bonds, functional groups, etc. to infer its biological activity in animals;

[0163] Efficacy index set: Efficacy index sets usually include physiological and pathological data related to the disease, such as biomarker data such as anti-inflammatory, antioxidant, and immunomodulatory. Through these index sets, the comprehensive efficacy of compound ingredients can be evaluated.

[0164] A rat ulcerative colitis (UC) model was established, and immunohistochemistry, molecular biology and other methods were used to screen the activity of the extracts of petroleum ether, chloroform, ethyl acetate, n-butanol and other substances, and the effective substance site group was determined. After analyzing and identifying its chemical components, the next step was to use the data of the Chinese medicine compound for screening. This screening process evaluates the overall effect of the compound based on the size accuracy, structural characteristics and efficacy index set of each component; by comparing with the animal model, the Chinese medicine compound with the greatest potential is screened out to ensure that its components can exert the best pharmacological effect in the body; by combining the size accuracy, structural characteristics and efficacy index of the components, the Chinese medicine compound with high efficacy can be accurately screened from multiple dimensions; the screening results are verified by using animal pharmacology experimental models, so that the theoretical efficacy speculation can be confirmed by experimental data, ensuring the scientificity and reliability of the screening process; the screened compound can better adapt to specific clinical needs and provide data support for subsequent drug development and practical applications.

[0165] Step 54 is a comprehensive indicator for measuring the interaction of the entire compound drug, representing whether the interaction of the compound components enhances the overall efficacy; the synergy index is defined by comparing the effect of the drug combination with that of a single drug. Generally, when the synergy index value is less than 1, it indicates that the efficacy of the combined drug is higher than that of a single drug, and vice versa, it indicates that the combined effect is not as good as that of a single drug. The index is obtained by calculating the ratio of the combined efficacy to the efficacy of each drug alone; the number of components in the compound affects the overall synergistic effect of the drug. The combination of more components may lead to more complex interactions, so it is necessary to comprehensively evaluate the contribution of each component to the efficacy; through animal experimental models, researchers can simulate the use of the compound and monitor the interaction and synergistic effect between the components of the compound. For example, in some cases, there may be an enhancement effect (synergistic effect) between the components, while in other cases, the components may inhibit or conflict with each other (antagonistic effect); through experimental data, the overall drug interaction index (I) of the compound and the synergy index (CI) of the single drug combination are calculated. This calculation can help scientists understand which ingredients in the compound have the most significant effect and provide a basis for subsequent dosage optimization; identify drug interactions: it can quantify the synergistic effects between ingredients in the compound and evaluate whether the combination of different ingredients is conducive to enhancing drug efficacy and reducing ineffective ingredients or side effects; improve the accuracy of drug efficacy prediction: through quantitative analysis of synergistic effects, researchers can predict the potential effects of different combinations of ingredients in clinical practice, thereby optimizing the ingredients and dosage of the compound; it can discover the antagonistic effects between ingredients in advance, provide scientific guidance for compound optimization, and avoid situations in which the drug efficacy is weakened or offset each other in the formula.

[0166] Step 55 After determining the synergistic effect of each component, the next step is to evaluate the effect of different dose combinations on the efficacy. This process considers the contribution of each component to the efficacy at different doses. For example, some components may have the best efficacy at low doses, while other components may require higher doses to produce effects; by optimizing the dosage of the components, it is possible to ensure that the contribution of each component's efficacy in the compound is maximized to avoid excessive or insufficient amounts of certain components; at the same time, it is necessary to consider the possible toxic and side effects of each component. For example, some Chinese herbal medicine components may cause side effects such as liver and kidney damage and gastrointestinal discomfort at high doses, so safety must be comprehensively considered when optimizing the dosage; by evaluating the toxic and side effects of each component, the dosage can be adjusted in a timely manner to ensure a balance between the safety and efficacy of the prescription; based on the evaluation of efficacy and side effects, real-time adjustments can be made to achieve the best efficacy and the least side effects. For example, based on clinical trials or individualized treatment needs, the dosage of compound ingredients can be adjusted individually to meet the needs of specific patients; the dosage of each ingredient can be accurately controlled to maximize the efficacy of the compound while avoiding unnecessary side effects caused by excessive use of certain ingredients; by considering the side effects of each ingredient and adjusting the dosage according to the patient's specific situation, the safety of the compound can be ensured and protection can be provided for clinical application; according to the patient's specific situation (such as age, weight, condition, etc.), the dosage of the Chinese medicine compound can be adjusted individually to provide tailor-made treatment plans and further improve the treatment effect.

[0167] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A method for screening Chinese medicine compound prescriptions based on machine learning, characterized in that: The following steps are involved: Collect the data of spleen-tonifying ingredients in traditional Chinese medicine compound to obtain the traditional Chinese medicine compound dataset; Integrate and encode each data item in the Chinese medicine compound data set into sequence data, train the sequence data to construct a compound screening model, and output the results of identifying the efficacy of the components and the effective parts of the components of the Chinese medicine compound; Input at least one data item in the newly entered Chinese medicine compound data set into the compound screening model to output the corresponding Chinese medicine compound component efficacy and component effective part identification results after screening; Based on the recognition results of the compound screening model, the chemical component structure characteristics of the effective ingredient part data in the traditional Chinese medicine compound data set are extracted; The effective component part data is the main component data, based on the chemical component structure characteristics, to obtain the main component efficacy data, toxicity data and time-effect data; the main component efficacy data, toxicity data and time-effect data are constructed into an efficacy index set; according to the chemical component structure characteristics and the efficacy index set, the auxiliary added value of the component efficacy is evaluated; Characterize and analyze the component data in the traditional Chinese medicine compound data set to obtain the size accuracy of the compound components; The size accuracy, structural characteristics and efficacy index set of the compound ingredients are used to establish an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound ingredients; based on the experimental results, the proportion of the compound Chinese medicine ingredients is adjusted to maximize the efficacy of each component data.

2. A method for screening Chinese medicine compound prescriptions based on machine learning according to claim 1, characterized in that: Collect the data of spleen-tonifying ingredients in traditional Chinese medicine compound prescriptions to obtain the traditional Chinese medicine compound prescription dataset, including: Collect multiple component data of Chinese medicine compound prescriptions and extract the chemical component data of Chinese medicine compound prescriptions to retain the effective Chinese medicine component data; Performing physical analysis on the effective Chinese medicine component data, and recording the smell, dosage and ratio of multiple components to obtain Chinese medicine compound component information data; According to the Chinese medicine compound ingredient information data, combined with the input clinical trial data, to obtain the basic efficacy information data of the Chinese medicine compound; Integrate the effective Chinese medicine ingredient data, Chinese medicine compound ingredient information data and Chinese medicine compound basic efficacy information data to obtain the Chinese medicine compound dataset.

3. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 2, characterized in that: Integrate and encode each data item in the Chinese medicine compound data set into sequence data, train the sequence data to construct a compound screening model, and output the results of identifying the efficacy and effective parts of the components of the Chinese medicine compound, including: The data items in the Chinese medicine compound data set are supplemented or eliminated without duplication or missing items to obtain the pre-processed data items; the pre-processed data items are formatted in a unified format, and characteristic data with predictive power of the Chinese medicine compound are extracted, wherein the characteristic data with predictive power are Chinese medicine ingredients, proportions, medicinal properties and meridians; The effective Chinese medicine component data except Bletilla striata is extracted and separated by a system solvent method to obtain effective component part information data of the Chinese medicine compound components; Encoding the characteristic data of each Chinese medicine compound data item and the effective part information data of the components into a characteristic vector sequence data; to obtain a plurality of sequence data; performing normalization processing on the plurality of sequence data to obtain a sequence data set; The sequence data set is input into the compound screening model; the compound screening model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and the intermediate representation data of multiple hidden layers are transmitted to the output layer, and the output layer outputs the identification results of the efficacy of the ingredients and the effective parts of the ingredients of the screened Chinese medicine compound.

4. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 3, characterized in that: Based on the recognition results of the compound screening model, the chemical component structure characteristics of the principal component data in the traditional Chinese medicine compound data set are extracted, including: According to the recognition result of the compound screening model, principal component analysis and recognition are performed to obtain the principal component data of the Chinese medicine compound; the principal component data of the Chinese medicine compound is Bletilla striata; Obtaining chemical structure characteristic data of the main components of the traditional Chinese medicine compound; the chemical structure characteristic data includes chemical molecular formula, molecular weight, functional group and stereo configuration; According to the chemical structure characteristic data of the main components, the main components are extracted, separated and purified so that the components of the Chinese herbal medicine compound can play the role of both medicine and auxiliary.

5. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 4, characterized in that: Based on the structural characteristics of the chemical components, the efficacy data, toxicity data and time-effect data of the main components are obtained, including: According to the identification result of the compound screening model, the initial efficacy data of the main component is obtained; and based on the main component extraction, the actual proportion data of the main component in the traditional Chinese medicine compound is obtained to obtain the actual efficacy data; Based on the actual efficacy data, the human body system and the main component dose analysis are performed to obtain the main component toxicity analysis results of the Chinese medicine compound, and the toxicity analysis results are the efficacy and side effect data; according to the analysis results, the main component side effect information data is obtained to ensure the safe use of the Chinese medicine compound; Through pharmacological analysis, the efficacy analysis results of the Chinese medicine compound in N time periods after opening are obtained, and N={n1, n2, …, n i The drug effects at multiple time points of} are recorded and integrated into a time-effect data set; where n i Represents the i-th time node n.

6. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 5, characterized in that: According to the analysis results, the main component side effect information data is obtained to ensure the safe use of Chinese medicine compound, including: Obtaining the dosage values ​​of side effects of ingredients for the general population and special populations, wherein the special populations include the metabolism, absorption, and excretion capacity data of pregnant women, lactating women, children, the elderly, and patients with liver and kidney dysfunction; using network pharmacology to analyze the targets of the main ingredients in the traditional Chinese medicine compound according to the dosage values ​​of side effects of ingredients for the special populations, wherein the targets include enzymes, receptors, and signaling pathways; Based on the target information of the principal component, ; to obtain the interaction between the main component dose and the human system; where E represents the effect caused by the drug; E max is the maximum effect that the drug can achieve; C is the concentration of the drug; is the drug concentration required to produce the h-potency effect; n is the Hill coefficient, describing the slope of the curve; the slope of the curve is related to the cooperativeness of the main component of the traditional Chinese medicine compound to bind to the receptor; Based on the interaction between the main component dose and the human body system, the preset dose of the main component of the traditional Chinese medicine compound is subjected to allergen screening to obtain the allergic factor of the preset dose of the main component; according to the allergic factor data combined with the interaction between the main component dose and the human body system, ; To obtain the risk assessment index of the main component drug dosage information for special populations; When RI>1, it means that the main ingredient dose poses a risk to special populations. In the formula, RI represents the main ingredient drug dose information risk assessment index; T represents the side effect parameter; Q represents the number of allergic factors; and S represents the safe dose limit.

7. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 6, characterized in that: Based on the interaction between the main component dose and the human body system, the preset dose of the main component of the traditional Chinese medicine compound is subjected to allergen screening to obtain the allergenic factors of the preset dose of the main component, including: Obtain the dosage range of the main ingredients in the Chinese medicine compound, including the effective dose or the dose that causes adverse reactions; organize the known dose-response relationship data based on the side effect dose of the main ingredients in the Chinese medicine compound, including the activity threshold and allergic threshold of the ingredients; Based on the dose-response relationship data, the main components of the Chinese herbal compound were used to treat the immune cell model of allergic constitution at different dose concentrations to verify the allergic reaction threshold and mechanism of action; Record the incidence of allergic reactions in small-scale clinical trials and validate the predicted allergenic factors to obtain validation results; According to the validation results, the dose threshold and the category of allergic factors are revised and updated; the key molecular features that cause allergic factors are extracted, and the key molecular features include molecular structure fragments and specific metabolites; they are classified and identified as direct antigens or indirect antigens; Based on the experimental and calculation results, the preset dosage range of the main ingredients is adjusted to reduce the risk of allergic reactions.

8. The method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 7, characterized in that: According to the chemical composition structure characteristics and efficacy index set, the auxiliary added value of the ingredient efficacy is evaluated, including: Based on the chemical composition structure characteristics and efficacy index set of the main components of the traditional Chinese medicine compound, the quantitative relationship between the molecular structure characteristics of the chemical components and the efficacy index is analyzed by establishing the efficacy-structure association mechanism; pass , to obtain the enhancement effect of the main component and other components in the traditional Chinese medicine compound on the efficacy; where, is the enhancement effect index, It is the efficacy of the compound combination. It is the sum of the effects of the individual ingredients. >1, indicating that the combined effect is greater than the total effect of a single component, indicating an enhanced effect; According to the enhancing effect of the main component and other components on the efficacy, it is obtained that the main component has an auxiliary added value of efficacy, and the added value is ≥1.

9. The method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 8, characterized in that: Characterize and analyze the component data in the traditional Chinese medicine compound data set to obtain the size accuracy of the compound components, including: Based on the efficacy index set, ; To obtain the nonlinear relationship between the data items of the components; Wherein, D represents the current molecular size of each data item of the Chinese medicine compound; D0 represents the size value with the highest efficacy; K1, K2, K3, K4 represent model parameters to optimize the fitting of each data item; A represents the binding affinity parameter of each data item; S represents the efficacy enhancement value of the Chinese medicine compound; the size value with the highest efficacy is obtained by Get; where, Represents the shortest distance between the i-th atom and the j-th atom in a molecule; Based on the nonlinear relationship between the data items of the components, ; To obtain the molecular size accuracy of each component when the Chinese medicine compound has the maximum efficacy; where, A function representing the molecular size accuracy of each component at maximum efficacy.

10. A method for screening Chinese medicinal compound prescriptions based on machine learning according to claim 9, characterized in that: The size accuracy, structural characteristics and efficacy index set of the compound ingredients are used to establish an animal pharmacology experimental model to obtain the synergistic effect and efficacy contribution value between the compound ingredients, including: Based on the size accuracy, structural characteristics and efficacy index set of compound ingredients, an animal pharmacology experimental model is established; the Chinese medicine compound is screened according to the size accuracy, structural characteristics and efficacy index set of compound ingredients to obtain the screened Chinese medicine compound; The screened Chinese herbal compound is tested through animal pharmacology experimental models, and according to the obtained Chinese herbal compound information data, ; to obtain the synergistic effect between the components of the Chinese medicine compound; where I represents the overall drug interaction index; CI is the synergistic index of the single drug combination, which is obtained by comparing the ratio of the combined efficacy to the efficacy of each drug alone; m is the number of drugs in the combination; Based on the synergistic effects among the ingredients of traditional Chinese medicine compound, the dosage combination of different ingredients and their contribution to the efficacy are evaluated, and the dosage of each ingredient is adjusted in real time. At the same time, the toxic and side effects of each ingredient are taken into consideration to optimize the formula to maximize the efficacy and minimize the side effects.

Images