Traditional Chinese medicine formula evaluation system and method based on artificial intelligence

Abstract

The invention discloses a traditional Chinese medicine prescription evaluation system and method based on artificial intelligence, and belongs to the technical field of traditional Chinese medicine prescription evaluation. The data preprocessing module is used for preprocessing the data related to the Chinese herbal medicines to obtain preprocessed data; the feature extraction module is used for extracting key features of the preprocessed data to obtain a key feature data set; the model training module is used for constructing a neural network model, training the neural network model through the key feature data set, and generating an evaluation model; the evaluation module is used for acquiring the traditional Chinese medicine formula in real time and inputting the traditional Chinese medicine formula into an evaluation model to generate an evaluation result; and the result output module is used for visually displaying the evaluation result and generating a prescription optimization suggestion. Through the method, rapid, accurate and objective evaluation of the traditional Chinese medicine formula can be realized, and technical support is provided for modernization and internationalization of traditional Chinese medicines.

Description

Technical Field

[0001] The present invention belongs to the technical field of traditional Chinese medicine prescription evaluation, and in particular relates to a traditional Chinese medicine prescription evaluation system and method based on artificial intelligence. Background Art

[0002] Traditional Chinese medicine prescriptions are composed of a variety of herbs, each of which contains complex chemical components that interact with each other in the prescription to exert a therapeutic effect. However, the evaluation of traditional Chinese medicine prescriptions mainly relies on the experience of traditional Chinese medicine practitioners and lacks standardized and quantitative evaluation methods, which limits the application and development of traditional Chinese medicine prescriptions in the field of modern medicine.

[0003] The rapid development of artificial intelligence technology, especially the progress of machine learning and data mining technology, has provided new possibilities for the scientific evaluation of Chinese medicine prescriptions. At present, the application of artificial intelligence technology to the analysis of Chinese medicine prescriptions has the problems of incomplete or non-standard data sets, which makes it difficult to fully reflect the efficacy and chemical composition of herbs, and lacks an effective evaluation framework, making it impossible to accurately evaluate the effectiveness and safety of prescriptions. At the same time, the representation method of Chinese herbal medicines is too simple and cannot fully capture the complex characteristics of herbs. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a Chinese medicine prescription evaluation system and method based on artificial intelligence to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides a Chinese medicine formula evaluation system based on artificial intelligence, comprising:

[0006] Data collection module, used to collect data related to Chinese herbal medicine;

[0007] A data preprocessing module, used to preprocess the data related to the Chinese herbal medicine to obtain preprocessed data;

[0008] A feature extraction module, used to extract key features of the preprocessed data based on a machine learning algorithm to obtain a key feature data set;

[0009] A model training module is used to build a neural network model, train the neural network model through the key feature data set, and generate an evaluation model;

[0010] An evaluation module, used to obtain a Chinese medicine prescription in real time, input the Chinese medicine prescription into the evaluation model, and generate an evaluation result;

[0011] The result output module is used to visualize the evaluation results and generate prescription optimization suggestions.

[0012] Preferably, the data acquisition module includes:

[0013] Automated data crawling tools for obtaining Chinese herbal medicine data from public databases and literature libraries on the Internet;

[0014] The local data import module is used to obtain data files uploaded locally by users.

[0015] Preferably, the data preprocessing module includes:

[0016] A data cleaning unit, used to identify and correct inconsistencies and errors in the data related to the Chinese herbal medicine to obtain a corrected data set;

[0017] A missing value processing unit, used to process the missing data of the modified data set to obtain a missing value processing data set;

[0018] An outlier detection unit is used to process outliers in the missing processing data set to generate the preprocessed data.

[0019] Preferably, the feature extraction module comprises:

[0020] a statistical analysis unit to calculate the frequency of use, dosage distribution, and efficacy associations of herbal medicines;

[0021] a pattern recognition unit for identifying potential associations and interaction patterns between the herbs based on the usage frequency, dosage distribution and efficacy association of the herbs;

[0022] A dimension reduction unit is used to process the dimension of the feature space of the pre-processed data based on the potential association and interaction pattern between the herbs to generate the key feature data set.

[0023] Preferably, the model training module includes:

[0024] A model selection unit, used to build a neural network model;

[0025] A hyperparameter optimization unit, used to train the neural network model based on the key feature data set, and adjust the parameters of the neural network model through an optimization algorithm to generate a trained neural network model;

[0026] A cross-validation unit is used to evaluate the generalization ability of the trained neural network model based on a cross-validation algorithm to generate the evaluation model.

[0027] Preferably, the evaluation module comprises:

[0028] An input verification unit is used to obtain a Chinese medicine prescription, check the integrity and rationality of the Chinese medicine prescription, and obtain the checked Chinese medicine prescription;

[0029] The real-time evaluation unit is used to input the checked Chinese medicine prescription into the evaluation model for scoring and analysis to generate the evaluation result.

[0030] Preferably, the result output module includes:

[0031] A visualization unit, used for visually displaying the evaluation results;

[0032] A report generating unit, configured to generate an evaluation report based on the evaluation result;

[0033] An optimization suggestion unit is used to generate a prescription optimization suggestion based on the evaluation report.

[0034] To achieve the above object, the present invention also provides a method for evaluating a Chinese medicine prescription based on artificial intelligence, comprising:

[0035] Collect data related to Chinese herbal medicines;

[0036] Preprocessing the data related to the Chinese herbal medicine to obtain preprocessed data;

[0037] Extracting key features of the preprocessed data based on a machine learning algorithm to obtain a key feature data set;

[0038] Constructing a neural network model, training the neural network model using the key feature data set, and generating an evaluation model;

[0039] Acquire a Chinese medicine prescription in real time, input the Chinese medicine prescription into the evaluation model, and generate an evaluation result;

[0040] The evaluation results are visualized and prescription optimization suggestions are generated.

[0041] Compared with the prior art, the present invention has the following advantages and technical effects:

[0042] The present invention discloses a Chinese medicine formula evaluation system and method based on artificial intelligence, which belongs to the technical field of Chinese medicine formula evaluation, and includes: a data acquisition module for collecting data related to Chinese herbal medicine; a data preprocessing module for preprocessing the data related to the Chinese herbal medicine to obtain preprocessed data; a feature extraction module for extracting key features of the preprocessed data based on a machine learning algorithm to obtain a key feature data set; a model training module for building a neural network model, training the neural network model through the key feature data set to generate an evaluation model; an evaluation module for real-time acquisition of Chinese medicine formulas, inputting the Chinese medicine formulas into the evaluation model, and generating evaluation results; a result output module for visualizing the evaluation results and generating formula optimization suggestions. The present invention can realize rapid, accurate and objective evaluation of Chinese medicine formulas, and provide technical support for the modernization and internationalization of Chinese medicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0044] Figure 1 Schematic diagram of a Chinese medicine formula evaluation system based on artificial intelligence according to an embodiment of the present invention;

[0045] Figure 2 This is a flow chart of a method for evaluating a Chinese medicine prescription based on artificial intelligence according to an embodiment of the present invention;

[0046] Figure 3 This is a flowchart of the self-attention mechanism operation of an embodiment of the present invention;

[0047] Figure 4 This is a NexGenPPI model architecture diagram of an embodiment of the present invention;

[0048] Figure 5 This is the ppi diagram of the traditional net medicine method when the confidence level is 0.9 according to the embodiment of the present invention;

[0049] Figure 6 This is the ppi diagram of the improved method when the confidence level is 0.9 according to the embodiment of the present invention. DETAILED DESCRIPTION

[0052] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0053] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0054] Embodiment 1

[0055] like Figure 1 As shown, this embodiment provides a Chinese medicine formula evaluation system based on artificial intelligence, including:

[0056] (a1) Data collection module, used to collect data related to Chinese herbal medicine from Chinese herbal medicine databases, clinical trial reports, scientific research literature, patent databases and expert knowledge bases;

[0057] (b1) Data preprocessing module, used to perform data cleaning, missing value processing, outlier detection, data standardization, feature scaling and encoding conversion;

[0058] (c1) a feature extraction module for extracting key features from the processed data using statistical analysis, pattern recognition, dimensionality reduction techniques, and machine learning algorithms;

[0059] (d1) Model training module, used to train and evaluate models using machine learning algorithms such as deep learning, ensemble learning, and reinforcement learning;

[0060] (e1) an evaluation module, which is used to receive the input of the Chinese medicine formula and call the trained evaluation model to evaluate the effectiveness, safety and rationality;

[0061] (f1) Result output module, which is used to display the evaluation results in the form of visual charts, text reports and data files, and provide prescription optimization suggestions based on data mining technology.

[0062] The data acquisition module further comprises:

[0063] (a2) Automated data crawling tools are used to regularly crawl the latest Chinese herbal medicine data from public databases and literature repositories on the Internet;

[0064] (a3) A local data import module that allows users to upload local data files through a graphical user interface, including but not limited to CSV, XLSX, JSON and XML formats.

[0065] The data preprocessing module further comprises:

[0066] (b2) data cleaning unit, used to identify and correct inconsistencies and errors in the data set;

[0067] (b3) a missing value processing unit, used to process missing data in a data set by interpolating, deleting or predicting missing values;

[0068] (b4) An outlier detection unit, used to identify and process outliers through statistical tests or machine learning algorithms.

[0069] The feature extraction module further comprises:

[0070] (c2) statistical analysis unit to calculate the frequency of use, dosage distribution and efficacy association of herbal medicines;

[0071] (c3) pattern recognition unit, used to identify potential associations and interaction patterns among herbs;

[0072] (c4) Dimensionality reduction unit, which is used to reduce the dimension of the feature space through methods such as principal component analysis, linear discriminant analysis or autoencoder.

[0073] The model training module further includes:

[0074] (d2) a model selection unit, used to select the most appropriate model from a plurality of preset models, including but not limited to a random forest, a gradient boosting machine, a support vector machine, and a deep neural network;

[0075] (d3) a hyperparameter optimization unit, which is used to automatically adjust model parameters through methods such as grid search, random search or Bayesian optimization;

[0076] (d4) Cross-validation unit, used to evaluate the generalization ability of the model through k-fold cross-validation, leave-one-out cross-validation or bootstrap method.

[0077] The evaluation module further comprises:

[0078] (e2) Input verification unit, which is used to check the integrity and rationality of the input formula, including the herbal species, dosage and preparation method;

[0079] (e3) a real-time evaluation unit for quickly scoring and analyzing input formulas and providing immediate feedback;

[0080] (e4) Safety assessment unit, which is used to evaluate the potential risks and side effects of the formulation, including drug interactions and allergic reactions.

[0081] The result output module further comprises:

[0082] (f2) a visual display unit for displaying the evaluation results in the form of graphs or charts, including but not limited to bar graphs, line graphs, scatter graphs and heat maps;

[0083] (f3) Report generation unit, used to generate a detailed evaluation report, including an evaluation summary, detailed analysis and references;

[0084] (f4) Optimization suggestion unit, used to provide formulation optimization suggestions based on data mining technology, including herbal substitution, dosage adjustment and preparation method improvement.

[0085] Embodiment 2

[0086] like Figure 2 As shown, this embodiment provides a method for evaluating a Chinese medicine prescription based on artificial intelligence, comprising:

[0087] 1. Construction and optimization of own data sets:

[0088] Comprehensive data collection strategy: Implement a multi-source data collection strategy, including the chemical composition, efficacy, dosage, preparation method, historical use records, clinical trial data, expert knowledge and literature of Chinese herbal medicines. Pay special attention to the breadth and depth of the data to ensure that the data set covers all known Chinese herbal medicines and their properties.

[0089] Data standardization process: Develop a set of standardized processes to convert data from different sources into a unified format, including terminology, measurement units, and coding systems, for easy analysis and processing.

[0090] Data richness and quality control: Ensure data accuracy and reliability through a combination of algorithms and manual review. Implement regular data update and maintenance plans to keep the database current and relevant.

[0091] 2. Feature extraction and vector representation technology:

[0092] Advanced data preprocessing algorithms: Use advanced data preprocessing algorithms, including missing value processing, outlier detection, and data transformation to improve data quality.

[0093] Feature Engineering Strategy: Use feature engineering strategies, including feature selection, feature construction, and dimensionality reduction techniques, to extract the most representative and predictive features.

[0094] Herb Vector Representation Model: Develop a herb vector representation model that uses deep learning techniques such as convolutional neural networks (CNN) or recurrent neural networks (RNN) to convert the textual descriptions and chemical components of herbs into vector representations in a high-dimensional space to capture the complex properties of herbs.

[0095] 3. Evaluation framework design and implementation:

[0096] Customized model training process: Design a customized model training process, including data partitioning, model selection, feature vector input, and hyperparameter tuning to build a high-performance evaluation model.

[0097] Multimodal evaluation model: Construct a multimodal evaluation model that integrates the chemical composition, efficacy information, and clinical data of herbal medicines to improve the accuracy and reliability of the evaluation.

[0098] Model validation and testing: Implement a rigorous model validation and testing process, including cross-validation, external validation, and model comparison, to ensure the generalization ability and evaluation accuracy of the model.

[0099] 4. Refined implementation of the evaluation module:

[0100] User-friendly input interface: Design an intuitive user input interface that supports multiple input methods, including text input, file upload, and graphic drag and drop, to improve user experience.

[0101] Real-time evaluation engine: Develop a real-time evaluation engine that can quickly process the Chinese medicine prescriptions input by users and provide instant feedback.

[0102] Safety and efficacy analysis: Implement the safety and efficacy analysis module to evaluate potential drug interactions, side effects and efficacy in the formulation.

[0103] 5. Result output and optimization suggestion system:

[0104] Interactive Visualization Tools: Develop interactive visualization tools that allow users to intuitively understand assessment results through graphs or charts.

[0105] Personalized report generator: Implement a personalized report generator to customize the content and format of the assessment report according to user needs.

[0106] Intelligent optimization suggestion engine: Build an intelligent optimization suggestion engine to provide personalized formulation optimization suggestions based on pharmacodynamics, pharmacokinetics and patient-specific conditions.

[0107] 6. System integration and user interaction optimization:

[0108] Modular system architecture: Design a highly modular system architecture to ensure system flexibility and scalability.

[0109] User interaction optimization: Continuously optimize user interaction design, collect user feedback, and iteratively improve the user interface and experience.

[0110] System testing and quality assurance: Implement comprehensive system testing, including unit testing, integration testing, performance testing, and security testing to ensure system stability, performance, and security.

[0111] Embodiment 3

[0112] This embodiment provides a method for evaluating a Chinese medicine prescription based on artificial intelligence, comprising:

[0113] A large amount of data was collected from classic ancient books, academic paper databases, and authoritative online pharmacology databases such as ETCM and TCMID, including 6192 prescription records, 4667 herbal information, 13099 chemical molecular data, and 14184 potential target-related data. After data cleaning, data reduction, and data integration, prescriptions, herbs, and related diseases were onehot encoded;

[0114] The digital statistics method is used to use a two-dimensional array to count the relationship between herbs and diseases, so as to predict the efficacy of the medicine. In this embodiment, each herb counted is used as a row of the two-dimensional array, and all the diseases counted are used as columns of the two-dimensional array. First, the values ​​of the two-dimensional array elements are all assigned to 0, and then according to Figure 1 The relevant data in the table performs corresponding operations on each element of the two-dimensional array and finally obtains the corresponding results.

[0115] For example, the formula "Ejiao Buxue Granule" contains the herbs "Ejiao, Rehmannia glutinosa, Codonopsis pilosula, Astragalus membranaceus, Lycium barbarum, Atractylodes macrocephala", and can treat the diseases "asthenia, cough, and blood-related diseases". In the two-dimensional array, the "asthenia, cough, and blood-related diseases" column of the row where "Ejiao, Rehmannia glutinosa, Codonopsis pilosula, Astragalus membranaceus, Lycium barbarum, Atractylodes macrocephala" is located will be increased by 1 accordingly.

[0116] By traversing all the data collected in this embodiment, the two-dimensional array finally obtained can reflect the relationship between the herbal medicine and the disease to a certain extent. The larger the element in the two-dimensional array, the greater the correlation between the herbal medicine and the treatment of the disease.

[0117] When a prescription is input in this embodiment, the model will calculate the treatment effect scores corresponding to all diseases based on the two-dimensional array, and output the diseases and scores according to the ranking of treatment effects.

[0118] In order to support more complex machine learning tasks and knowledge graph analysis, the team used the TransE_l2 model in the DGLKE (Deep Graph Library Knowledge Embedding) library to transform and embed the prescription data in version 0.5. This step converts the prescription entities and relationships in the knowledge graph into low-dimensional vector representations to improve the applicability and efficiency of the data in various machine learning algorithms. Through the TransE_l2 model, the herbal entities and prescription relationships in the prescription data are embedded in a low-dimensional vector space, and each herb and prescription is represented by a vector, thus forming a semantically rich knowledge graph.

[0119] At the same time, this embodiment divides the data set into a training set and a test set. The training set contains 1,123,592 data samples, and the test set contains 62,855 data samples.

[0120] After converting the herbs and prescriptions into vector representations, the prescriptions need to be trimmed and padded for input into machine learning tasks. Specifically, each herb in the prescription is represented as an embedding vector, which is assembled into a fixed-dimensional matrix (n256), where n is the number of herbs in the prescription. To ensure the dimensionality consistency of the input model, if the number of herbs in the prescription is less than 30, it is padded with zero vectors; if the dimension exceeds 30, it is trimmed to keep the dimension unchanged. This approach not only meets the standard requirements for data processing in machine learning, but also retains the integrity and consistency of the data, ensuring that the model can effectively process and analyze various prescription data.

[0121] This embodiment of the project adopts a hybrid model structure based on Transformer encoder and multi-layer perceptron (MLP). By introducing the self-attention mechanism, Transformer encoder can efficiently process input sequences in parallel, thereby improving the performance and training efficiency of the model in sequence data processing. Compared with traditional recurrent neural networks (RNN) and convolutional neural networks (CNN), Transformer avoids the information bottleneck problem in traditional models and can better capture long-distance dependencies in sequences.

[0122] The project's model combines the advantages of the Transformer encoder and MLP. The Transformer encoder effectively captures global dependencies through the self-attention mechanism and improves the representation of data features; while the MLP further optimizes feature representation through efficient computing and processing capabilities, ultimately achieving an effective solution to complex classification and prediction problems. This combination not only ensures the performance of the model when processing sequence data, but also shows good performance in smaller-scale data and real-time applications.

[0123] On the test set, this example evaluates the performance indicators of the model. The loss value of the model on the test set is 0.0055, and the root mean square error is 0.0738, indicating that the model has good accuracy and generalization ability in data prediction and classification tasks. The top 5 hit accuracy rate is more than 80%, and the top 20 accuracy rate is more than 95%.

[0124] Using this model, this embodiment can input a prescription, and the analysis model will output the top 8 diseases that the prescription can treat and the scores of the treatment effects.

[0125] For example, in this embodiment, if the formula "Licorice, Honey Ring Powder, Hibiscus Mutabilis Leaves, Chicken Intestines, Elsholtzia Cervifolia, Indigofera Speciosa, Parsnip" is input, the model will output "1:DiseaseName Cold with score 60.77036380767822" and other content. The formula has a high score for the treatment of colds, indicating that the formula can treat colds well; for other diseases, the formula has a low score for the treatment effect, indicating that the formula cannot treat other diseases.

[0126] Network pharmacology is based on the theory of systems biology, and it analyzes the network of biological systems and selects specific signal nodes for multi-point drug molecule design. Traditional Chinese medicine has multiple components, multiple targets, and the complexity of the interactions between them. Network pharmacology has the characteristics of integrity and systemicity, and has the interaction relationship between targets. The two concepts are consistent, so this embodiment can use network pharmacology to conduct some research on traditional Chinese medicine.

[0127] The steps adopted by traditional Chinese medicine network pharmacology are: 1. Obtain active ingredients from Chinese medicine / prescriptions; 2. Predict target genes based on active ingredients; 3. Obtain target genes for diseases / phenotypes; 4. Take the intersection of two target genes; 5. Construct a Chinese medicine / prescription active ingredient target gene disease network; 6. Construct an interaction network; 7. GO, KEGG enrichment analysis; 8. Molecular binding verification; 9. Experimental verification.

[0128] For step 6, traditional Chinese medicine network pharmacology usually uses data in the database to construct an interaction network in Cytoscape. However, there is currently no database that contains comprehensive protein interaction results. Only some protein-protein interactions can be found from a single database. Users must search across multiple databases and integrate data based on multiple results. However, multiple databases are not unified in terms of the number of protein interaction pairs found, displayed data structure, model language, annotations and evaluation criteria. This has affected the convenience and comprehensiveness of research data acquisition to a certain extent, leading to the emergence of research bias.

[0129] Secondly, when using the STRING database to explore protein interactions, genes such as Akt1 and TP53 often show a high degree of relationship with other proteins, leading to the homogenization of key proteins. The results of efficacy network screening may alleviate this homogenization trend to a certain extent, but big data prediction methods still need to continue to improve to provide more accurate results.

[0130] Based on this background, this embodiment creates the NexGenPPI deep learning architecture to improve this step and enhance the convenience and accuracy of this step.

[0131] NexGenPPI is a sequence-based deep learning model built based on a new generation of protein language models and recursive neural networks. It has the advantages of high accuracy, low training requirements, good versatility, and fast construction speed. The model architecture adopts the Siamese architecture, and the left and right networks share the same parameters. The input parameters are first embedded through ESM2 to generate a 2560-dimensional tensor. The tensor is input into the GRU network. The model contains 3 layers of GRU. Each layer of GRU is bidirectional, processing from the beginning to the end of the sequence and from the end to the beginning, using half of the gated units in each direction.

[0132] Since all sequences are padded to the same fixed size during training, each GRU layer is implemented as a dynamic GRU that processes only the actual sequence portion and skips the padded portion. This is achieved via sequence masking, which accepts a sequence mask as a secondary input.

[0133] Each GRU contains 256 units and outputs a 256-dimensional tensor, with 128 in one direction and 128 in the other direction. That is, the last layer of GRU only takes the output vector of the last time step, which represents the characteristics of the entire sequence. The two 256-dimensional GRU results are processed by Hadamard product to generate a tensor of the same dimension, which is input into the linear layer and DropOut. The last three linear layers have 256, 32 and 1 neurons respectively. The final output is a value between 0 and 1, which is used to describe the confidence level in the interaction, 0 represents low confidence and 1 represents high confidence.

[0134] This example uses a FASTA file containing protein sequences as input, mainly using the predict command, to obtain a .tsv file containing all predictions and a secondary .tsv file containing only positive interactions as output. Next, this example uses the NexGenPPI deep learning framework to improve the interaction network and perform subsequent GO and KEGG enrichment analysis and other operations.

[0135] Figure 3-6 The PPI graphs drawn using traditional Chinese medicine network pharmacology and the improved method of this embodiment are shown respectively at different confidence levels. It can be seen that at different confidence levels, the PPI graph drawn using the improved method of this embodiment is sparser than the PPI graph drawn using the traditional network medicine method, which shows that the deep model of this embodiment eliminates some false positive results through the algorithm and has higher accuracy.

[0136] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A Chinese medicine formula evaluation system based on artificial intelligence, characterized in that: include: Data collection module, used to collect data related to Chinese herbal medicine; A data preprocessing module, used to preprocess the data related to the Chinese herbal medicine to obtain preprocessed data; A feature extraction module, used to extract key features of the preprocessed data based on a machine learning algorithm to obtain a key feature data set; A model training module is used to build a neural network model, train the neural network model through the key feature data set, and generate an evaluation model; An evaluation module, used to obtain a Chinese medicine prescription in real time, input the Chinese medicine prescription into the evaluation model, and generate an evaluation result; The result output module is used to visualize the evaluation results and generate prescription optimization suggestions.

2. The artificial intelligence-based Chinese medicine formula evaluation system according to claim 1, characterized in that: The data acquisition module comprises: Automated data crawling tools for obtaining Chinese herbal medicine data from public databases and literature libraries on the Internet; The local data import module is used to obtain data files uploaded locally by users.

3. The Chinese medicine formula evaluation system based on artificial intelligence according to claim 1, characterized in that: The data preprocessing module comprises: A data cleaning unit, used to identify and correct inconsistencies and errors in the data related to the Chinese herbal medicine to obtain a corrected data set; A missing value processing unit, used to process the missing data of the modified data set to obtain a missing value processing data set; An outlier detection unit is used to process outliers in the missing processing data set to generate the preprocessed data.

4. The Chinese medicine formula evaluation system based on artificial intelligence according to claim 1, characterized in that: The feature extraction module comprises: a statistical analysis unit to calculate the frequency of use, dosage distribution, and efficacy associations of herbal medicines; a pattern recognition unit for identifying potential associations and interaction patterns between the herbs based on the usage frequency, dosage distribution and efficacy association of the herbs; A dimension reduction unit is used to process the dimension of the feature space of the pre-processed data based on the potential association and interaction pattern between the herbs to generate the key feature data set.

5. The Chinese medicine formula evaluation system based on artificial intelligence according to claim 1, characterized in that: The model training module includes: A model selection unit, used to build a neural network model; A hyperparameter optimization unit, used to train the neural network model based on the key feature data set, and adjust the parameters of the neural network model through an optimization algorithm to generate a trained neural network model; A cross-validation unit is used to evaluate the generalization ability of the trained neural network model based on a cross-validation algorithm to generate the evaluation model.

6. The Chinese medicine formula evaluation system based on artificial intelligence according to claim 1, characterized in that: The evaluation module includes: An input verification unit is used to obtain a Chinese medicine prescription, check the integrity and rationality of the Chinese medicine prescription, and obtain the checked Chinese medicine prescription; The real-time evaluation unit is used to input the checked Chinese medicine prescription into the evaluation model for scoring and analysis to generate the evaluation result.

7. The artificial intelligence-based Chinese medicine formula evaluation system according to claim 1, characterized in that: The result output module includes: A visualization unit, used for visually displaying the evaluation results; A report generating unit, configured to generate an evaluation report based on the evaluation result; An optimization suggestion unit is used to generate a prescription optimization suggestion based on the evaluation report.

8. The method of the traditional Chinese medicine formula evaluation system based on artificial intelligence according to any one of claims 1 to 7, characterized in that: The following steps are involved: Collect data related to Chinese herbal medicines; Preprocessing the data related to the Chinese herbal medicine to obtain preprocessed data; Extracting key features of the preprocessed data based on a machine learning algorithm to obtain a key feature data set; Constructing a neural network model, training the neural network model using the key feature data set, and generating an evaluation model; Acquire a Chinese medicine prescription in real time, input the Chinese medicine prescription into the evaluation model, and generate an evaluation result; The evaluation results are visualized and prescription optimization suggestions are generated.

Images