CN115050481B - Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network - Google Patents
Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network Download PDFInfo
- Publication number
- CN115050481B CN115050481B CN202210683366.8A CN202210683366A CN115050481B CN 115050481 B CN115050481 B CN 115050481B CN 202210683366 A CN202210683366 A CN 202210683366A CN 115050481 B CN115050481 B CN 115050481B
- Authority
- CN
- China
- Prior art keywords
- prescription
- chinese medicine
- traditional chinese
- graph
- efficacy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000003814 drug Substances 0.000 title claims abstract description 157
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 23
- 238000005516 engineering process Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 12
- 230000000694 effects Effects 0.000 claims abstract description 10
- 238000010586 diagram Methods 0.000 claims abstract description 9
- 238000009826 distribution Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000010276 construction Methods 0.000 claims abstract description 5
- 238000007418 data mining Methods 0.000 claims abstract description 4
- 238000003058 natural language processing Methods 0.000 claims abstract description 4
- 229940126680 traditional chinese medicines Drugs 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000005096 rolling process Methods 0.000 claims description 8
- 231100000419 toxicity Toxicity 0.000 claims description 8
- 230000001988 toxicity Effects 0.000 claims description 8
- 230000001105 regulatory effect Effects 0.000 claims description 6
- 230000010415 tropism Effects 0.000 claims description 6
- 230000005906 menstruation Effects 0.000 claims description 5
- 230000002776 aggregation Effects 0.000 claims description 4
- 238000004220 aggregation Methods 0.000 claims description 4
- 235000009508 confectionery Nutrition 0.000 claims description 4
- 229940079593 drug Drugs 0.000 claims description 4
- 210000002216 heart Anatomy 0.000 claims description 4
- 210000003734 kidney Anatomy 0.000 claims description 4
- 210000004072 lung Anatomy 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 210000000952 spleen Anatomy 0.000 claims description 4
- 238000006467 substitution reaction Methods 0.000 claims description 4
- 206010007247 Carbuncle Diseases 0.000 claims description 3
- 206010062717 Increased upper airway secretion Diseases 0.000 claims description 3
- 208000025865 Ulcer Diseases 0.000 claims description 3
- 230000004913 activation Effects 0.000 claims description 3
- 230000000740 bleeding effect Effects 0.000 claims description 3
- 239000008280 blood Substances 0.000 claims description 3
- 210000004369 blood Anatomy 0.000 claims description 3
- 230000029087 digestion Effects 0.000 claims description 3
- 230000004438 eyesight Effects 0.000 claims description 3
- 239000000796 flavoring agent Substances 0.000 claims description 3
- 235000019634 flavors Nutrition 0.000 claims description 3
- 210000000232 gallbladder Anatomy 0.000 claims description 3
- 230000001939 inductive effect Effects 0.000 claims description 3
- 210000002429 large intestine Anatomy 0.000 claims description 3
- 210000004185 liver Anatomy 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 210000005036 nerve Anatomy 0.000 claims description 3
- 210000003516 pericardium Anatomy 0.000 claims description 3
- 208000026435 phlegm Diseases 0.000 claims description 3
- 230000001737 promoting effect Effects 0.000 claims description 3
- 210000000813 small intestine Anatomy 0.000 claims description 3
- 210000002784 stomach Anatomy 0.000 claims description 3
- 231100000397 ulcer Toxicity 0.000 claims description 3
- 239000002253 acid Substances 0.000 claims description 2
- 239000000203 mixture Substances 0.000 description 7
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000006403 short-term memory Effects 0.000 description 3
- 230000000052 comparative effect Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 241000427159 Achyranthes Species 0.000 description 1
- 241000173529 Aconitum napellus Species 0.000 description 1
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 244000061520 Angelica archangelica Species 0.000 description 1
- 241000489492 Arisaema Species 0.000 description 1
- 241001061264 Astragalus Species 0.000 description 1
- 241000255789 Bombyx mori Species 0.000 description 1
- 244000037364 Cinnamomum aromaticum Species 0.000 description 1
- 235000014489 Cinnamomum aromaticum Nutrition 0.000 description 1
- 241000721047 Danaus plexippus Species 0.000 description 1
- 241000305492 Gastrodia Species 0.000 description 1
- 241000146384 Glaux Species 0.000 description 1
- 235000001287 Guettarda speciosa Nutrition 0.000 description 1
- 241000545744 Hirudinea Species 0.000 description 1
- 241000112528 Ligusticum striatum Species 0.000 description 1
- 241000239226 Scorpiones Species 0.000 description 1
- 229940023019 aconite Drugs 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 235000006533 astragalus Nutrition 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 229940126678 chinese medicines Drugs 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000007012 clinical effect Effects 0.000 description 1
- 239000000571 coke Substances 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003475 lamination Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007721 medicinal effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000003607 modifier Substances 0.000 description 1
- 231100000252 nontoxic Toxicity 0.000 description 1
- 230000003000 nontoxic effect Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000003997 social interaction Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 210000004233 talus Anatomy 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 239000009137 wuling Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/90—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to alternative medicines, e.g. homeopathy or oriental medicines
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H70/00—ICT specially adapted for the handling or processing of medical references
- G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Primary Health Care (AREA)
- Artificial Intelligence (AREA)
- Epidemiology (AREA)
- Pharmacology & Pharmacy (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Biomedical Technology (AREA)
- Animal Behavior & Ethology (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Pathology (AREA)
- Alternative & Traditional Medicine (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medicinal Chemistry (AREA)
- Toxicology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a traditional Chinese medicine prescription efficacy prediction method based on a graph convolution neural network, which comprises the following steps of: s1, a data preprocessing module preprocesses traditional Chinese medicine prescription data through a natural language processing technology, and builds a traditional Chinese medicine database and a prescription database which meet the data mining requirement; s2, a prescription diagram construction module, which is used for constructing prescription iso-graphs containing traditional Chinese medicine nodes and prescription nodes by representing prescription text data in a diagram form; s3, a node representation learning module learns node embedding representation in the prescription iso-graph through a graph convolution network for use by downstream tasks; s4, a prediction module is used for obtaining probability distribution of the traditional Chinese medicine prescription on the effect based on the final node embedded representation obtained through training of the traditional Chinese medicine prescription sample.
Description
Technical Field
The invention relates to the technical field of bioinformatics, in particular to a traditional Chinese medicine prescription efficacy prediction method based on a graph convolution neural network
Background
The Chinese medicine prescription is a main means for preventing and treating diseases by using Chinese medicine, is taken as a bridge for connecting the basis and clinic of the Chinese medicine, and is always the key point of the theory and clinic research of the Chinese medicine. The efficacy of the prescription is summarized by combining the compatibility theory and the clinical effect and characteristics of the prescription based on the theory of traditional Chinese medicine. The function of accurately grasping the compatibility of the traditional Chinese medicine composition plays a key role in clinical diagnosis and treatment effects of traditional Chinese medicine. Traditionally, the determination of efficacy of a prescription is carried out by taking into account the composition of the drug, the dosage ratio, the dialectical method and the like, and a great deal of manpower and material resources are required. If the efficacy of a prescription can be predicted by modern scientific technology, the result will provide valuable reference for subsequent clinical practice, and is also beneficial to the inheritance and innovation of traditional Chinese medicine theory.
The continuous emergence of intelligent information technology brings new opportunities for modern research of traditional Chinese medicine, and the deep fusion of artificial intelligence and traditional Chinese medicine leads the traditional Chinese medicine to be better inherited, developed and innovated. Existing methods for analyzing prescription efficacy based on artificial intelligence technology can be roughly divided into the following two types: (1) based on a topic model: wang et al treat traditional Chinese medicine as "words", prescription as "documents", and prescription efficacy as "subjects". The potential relation between the efficacy label and the traditional Chinese medicine is mined by using the topic model, a novel supervised topic model for predicting the efficacy of the prescription is provided, and the mining result of the prescription compatibility rule is brought into the learning process. However, as a typical bag-of-words model, the subject model has great limitations on short text data such as traditional Chinese medicine prescription, and it is difficult to understand the complex correlation between traditional Chinese medicine combination and prescription efficacy. (2) neural network-based: cheng et al propose an improved deep learning model consisting of a two-way long and short term memory neural network and a convolutional neural network, which predicts the efficacy of agents by building a plurality of two classifiers. However, such methods do not take into account the effect of the variation in the dosage of the drug in the formulation on the overall efficacy, and have certain limitations.
The graph convolutional neural network (Graph Neural Network, GCN) is taken as an important technology for information transmission and aggregation among nodes, can effectively apply the deep learning concept to unstructured data, and has remarkable effects in the fields of biomedicine, traffic logistics, electronic commerce, mobile social interaction and the like. In the formula process, besides the monarch, minister, assistant and guide structures of the prescription, the compatibility of medicines and the proportion of dosages are considered, and various factors such as the prescription, the traditional Chinese medicines, the medicine property, the dosages and the efficacy are involved, wherein the prescription, the traditional Chinese medicines, the medicine property and the efficacy have obvious space topological structures. Based on the method, on the basis of summarizing the defects of the prior methods, all prescription data are constructed into a complete heterogram, chinese medicinal attributes and common dosage ranges are introduced as external knowledge, and a prescription efficacy prediction method based on GCN is provided, so that contribution is made to the modern development of Chinese medicaments.
Disclosure of Invention
The invention aims to solve the technical problem of providing a traditional Chinese medicine prescription efficacy prediction method based on a graph convolution neural network, which is used for fully learning node embedding of traditional Chinese medicines and prescriptions through the graph convolution method and excavating the relation between the medicine composition and the prescription efficacy. In order to solve the problems, the technical scheme is as follows:
the invention discloses a traditional Chinese medicine prescription efficacy prediction method based on a graph convolution neural network, which comprises the following steps of:
s1, a data preprocessing module preprocesses traditional Chinese medicine prescription data through a natural language processing technology, and builds a traditional Chinese medicine database and a prescription database which meet the data mining requirement;
s2, a prescription diagram construction module, which is used for constructing prescription iso-graphs containing traditional Chinese medicine nodes and prescription nodes by representing prescription text data in a diagram form;
s3, a node representation learning module learns node embedding representation in the prescription iso-graph through a graph convolution network for use by downstream tasks;
s4, a prediction module is used for obtaining probability distribution of the traditional Chinese medicine prescription on the effect based on the final node embedded representation obtained through training of the traditional Chinese medicine prescription sample.
Further, the traditional Chinese medicine database comprises medicine names, aliases, sex flavors, menstruation, toxicity and common dosage range information, wherein the sex flavors comprise cold, heat, warmth, benign, flat, sour, bitter, sweet, pungent and salty, and the menstruation comprises lung, pericardium, heart, large intestine, triple-focus, small intestine, stomach, gall bladder, spleen, liver and kidney; the prescription database comprises prescription names, traditional Chinese medicine components, actual dosage and efficacy information; the efficacy information comprises a dryness treating prescription, a tonifying prescription, a blood regulating prescription, a wind dispelling prescription, a phlegm eliminating prescription, a carbuncle and ulcer prescription, a damp eliminating prescription, a qi regulating prescription, wen Lifang, a resolving prescription, a heat clearing prescription, an emergency prescription, a nerve soothing prescription, a eyesight improving prescription, a bleeding prescription, a resuscitation inducing prescription, a digestion promoting prescription, a astringing prescription and a exterior relieving prescription.
Further, in S1, preprocessing the prescription data of the traditional Chinese medicine includes:
s101, chinese medicine names are standardized, and the Chinese medicine names in the Chinese medicine prescription are subjected to homonymous substitution through a character string matching technology;
s102, modern conversion of dosage, namely splitting the name of the traditional Chinese medicine and the dosage, and converting calculation of different dosage units of the traditional Chinese medicine in the traditional Chinese medicine prescription into modern dosage unit grams.
Further, in S2, constructing a prescription iso-pattern, including:
s201, traditional Chinese medicine quantitative representation, namely converting traditional Chinese medicine attribute abstract description into specific numerical values as traditional Chinese medicine attribute vectorsd represents the attribute quantity of the traditional Chinese medicine; the Chinese medicinal properties include nature, meridian tropism and toxicity, and the nature and toxicity properties are quantified by index level, because these properties have modifier for describing degree, respectively 2 -1 、2 0 、2 1 The expression "slightly cold", "severe cold"; binary quantization is adopted for the meridian tropism attribute, wherein the existence of the attribute is represented by 1, and the absence of the attribute is represented by 0;
s202, quantitatively representing the Chinese medicinal prescription, namely matching Chinese medicaments contained in the Chinese medicinal prescription to corresponding Chinese medicament attribute vectors, and summing the Chinese medicament attribute vectors to obtain the Chinese medicinal prescription attribute vector
S203, normalizing the traditional Chinese medicine dosage according to the following relative dosage conversion formula:
wherein ,represents the relative dose of the Chinese medicine i in the prescription j, d ij Represents the actual dosage of the traditional Chinese medicine i in the prescription j, d min and dmax Respectively representing the minimum value and the maximum value of the traditional Chinese medicine common dosage range;
s204, calculating the similarity of the traditional Chinese medicines according to the following formula:
wherein ,xa and xb Attribute vectors respectively representing the traditional Chinese medicine a and the traditional Chinese medicine b;
s205, constructing a prescription iso-graph FH= (V, E), wherein V is a node set in the graph, E is an edge set in the graph, and a characteristic matrix of the graphCorresponding adjacency matrix->n represents n number of characteristicThe node, d, represents the node feature dimension, FH in undirected graph, and its adjacency matrix formalized representation is as follows:
wherein beta is the threshold value of the connecting edge between the nodes of the traditional Chinese medicines, when the similarity of the two traditional Chinese medicines is larger than beta, the connecting edge is established between the two traditional Chinese medicines, otherwise, the connecting edge is not established.
Further, in S3, nodes in the prescription iso-graph are learned through a graph rolling network to embed and represent, the graph rolling network captures neighbor information through one layer of convolution, and when the graph rolling network stacks multiple layers, high-order neighborhood information can be obtained;
inputting node attribute vectors at a first levelAccording to A i，j and Aj，i Obtaining an adjacent matrix A of FH, updating information of aggregation neighborhood nodes by using a traditional Chinese medicine prescription and traditional Chinese medicine nodes, and for a layer of graph rolling network, obtaining a k-dimensional node matrixFormalized representation of (c) is:
wherein ,is a normalized representation of the adjacency matrix a, +.>Representing the learned parameters, ρ being an activation function;
the multi-layer graph convolutional network is expressed as:
wherein l represents the number of layers, and H (0) ＝X。
Further, in S4, the traditional Chinese medicine prescription efficacy prediction module is specifically implemented as follows:
for each Chinese medicinal prescription, evaluating the distance between the predicted efficacy and the actual efficacy, setting the output dimension of the last layer of graph convolution network to be equal to the number of efficacy labels, and obtaining the final embedded representation of the Chinese medicinal prescription nodesAnd send it into Sigmoid classifier for learning to obtain probability distribution of Chinese medicine prescription on all effectsWherein C represents the number of efficacy tags, < + >>The calculation method is as follows:
further, in S4, the traditional Chinese medicine prescription efficacy prediction uses a multi-label cross entropy loss function as a loss function:
wherein n represents the number of training set samples, y (i) E {0,1} represents the actual efficacy label,representing the predicted value.
The traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network has the beneficial effects that:
1. the traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network provided by the invention combines the traditional Chinese medicine correlation theory to quantitatively express the traditional Chinese medicine and the prescription, thereby being beneficial to the information mining of the traditional Chinese medicine.
2. According to the traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network, provided by the invention, the influence of the dosage on the prescription efficacy is considered, and the normalization operation is carried out on the traditional Chinese medicine dosage through a relative dosage conversion formula, so that the dosage comparison among different traditional Chinese medicines is facilitated.
3. According to the traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network, prescription text data are represented in a graph mode, and the graph contains medicine composition, dosage and traditional Chinese medicine attribute information, so that node embedding with rich information can be learned easily.
4. The traditional Chinese medicine prescription efficacy prediction method based on the graph rolling neural network, provided by the invention, applies the graph rolling neural network, namely the artificial intelligence field leading edge technology, to prescription efficacy prediction scenes, and takes the relation between prescriptions and traditional Chinese medicines and the relation between traditional Chinese medicines into consideration, thus having a certain innovation in technology.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a general flow chart of a method for predicting efficacy of a traditional Chinese medicine prescription based on a graph roll-up neural network.
FIG. 2 is a diagram of a model for efficacy prediction of a graph roll-up neural network according to the present invention.
Detailed Description
In order to better understand the technical solution in the embodiments of the present invention and to make the above objects, features and advantages of the present invention more obvious, the following detailed description of the present invention will be given with reference to the accompanying drawings.
The description of these embodiments is provided to assist understanding of the present invention, but is not intended to limit the present invention. In addition, technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
Referring to fig. 1 and 2, the method for predicting efficacy of a traditional Chinese medicine prescription based on a graph convolution neural network in this embodiment includes the following steps:
s1, a data preprocessing module preprocesses traditional Chinese medicine prescription data through a natural language processing technology, and builds a traditional Chinese medicine database and a prescription database which meet the data mining requirement;
s2, a prescription diagram construction module, which is used for constructing prescription iso-graphs containing traditional Chinese medicine nodes and prescription nodes by representing prescription text data in a diagram form;
s3, a node representation learning module learns node embedding representation in the prescription iso-graph through a graph convolution network for use by downstream tasks;
s4, a prediction module is used for obtaining probability distribution of the traditional Chinese medicine prescription on the effect based on the final node embedded representation obtained through training of the traditional Chinese medicine prescription sample.
As a preferred embodiment, the traditional Chinese medicine database includes medicine name, alias, sex taste, menstruation, toxicity and common dose range information, the sex taste includes cold, heat, warmth, wellness, flatness, acid, bitter, sweet, pungent and salty, menstruation includes lung, pericardium, heart, large intestine, triple coke, small intestine, stomach, gall bladder, spleen, liver and kidney; the prescription database comprises prescription names, chinese medicinal composition, actual dosage and efficacy information; the efficacy information comprises a dryness treating prescription, a tonifying prescription, a blood regulating prescription, a wind dispelling prescription, a phlegm eliminating prescription, a carbuncle and ulcer prescription, a dampness eliminating prescription, a qi regulating prescription, wen Lifang, a relieving prescription, a heat clearing prescription, an emergency prescription, a nerve soothing prescription, an eyesight improving prescription, a bleeding prescription, a resuscitation inducing prescription, a digestion promoting prescription, a astringing prescription and a exterior relieving prescription.
Wherein, the Chinese medicine prescription data is shown in table 1, and the Chinese medicine prescription data is shown in table 2.
Table 1 examples of data for chinese medical formulas
Table 2 examples of chinese medicine data
Preferably, in S1, preprocessing the prescription data of the traditional Chinese medicine includes:
s101, chinese medicine names are standardized, and the Chinese medicine names in the Chinese medicine prescription are subjected to homonymous substitution through a character string matching technology;
s102, modern conversion of dosage, namely splitting the name of the traditional Chinese medicine and the dosage, and converting calculation of different dosage units of the traditional Chinese medicine in the traditional Chinese medicine prescription into modern dosage unit grams.
Preferably, in S2, constructing a prescription iso-pattern, including:
s201, traditional Chinese medicine quantitative representation, namely converting traditional Chinese medicine attribute abstract description into specific numerical values as traditional Chinese medicine attribute vectorsd represents the attribute quantity of the traditional Chinese medicine; the Chinese medicinal materials have properties including nature, taste, channel tropism and toxicity, and the nature, taste and toxicity are quantified by index level, respectively 2 -1 、2 0 、2 1 The expression "slightly cold", "severe cold"; binary quantization is adopted for the meridian tropism attribute, wherein the existence of the attribute is represented by 1, and the absence of the attribute is represented by 0; wherein, the quantitative representation of the traditional Chinese medicine is shown in the table 3, for example, ginseng is slightly warm, slightly bitter and sweet in taste, enters lung, heart, spleen and kidney meridians, is nontoxic, and the ginseng can be quantitatively represented as:
x ginseng radix ＝(0，0，0.5，0，0，0，0.5，1，0，0，1，0，1，0，0，0，0，0，0，1，0，1，0)。
Table 3 examples of quantized representations of chinese medicine
S202, quantitatively representing the Chinese medicinal prescription, namely matching Chinese medicaments contained in the Chinese medicinal prescription to corresponding Chinese medicament attribute vectors, and summing the Chinese medicament attribute vectors to obtain the Chinese medicinal prescription attribute vector
S203, normalizing the traditional Chinese medicine dosage according to the following relative dosage conversion formula:
wherein ,represents the relative dose of the Chinese medicine i in the prescription j, d ij Represents the actual dosage of the traditional Chinese medicine i in the prescription j, d min and dmax Respectively representing the minimum value and the maximum value of the traditional Chinese medicine common dosage range; taking Wuling powder as an example, the relative dose conversion results of the traditional Chinese medicines in the prescription are shown in Table 4:
table 4 relative dose transition examples
S204, calculating the similarity of the traditional Chinese medicines according to the following formula:
wherein ,xa and xb Attribute vectors respectively representing the traditional Chinese medicine a and the traditional Chinese medicine b;
s205, constructing a prescription iso-graph FH= (V, E), wherein V is a node set in the graph, E is an edge set in the graph, and a characteristic matrix of the graphCorresponding adjacency matrix->n represents n nodes with characteristics, d represents node characteristic dimension, FH is represented in an undirected graph, and the adjacency matrix is represented as follows:
wherein beta is the threshold value of the connecting edge between the nodes of the traditional Chinese medicines, when the similarity of the two traditional Chinese medicines is larger than beta, the connecting edge is established between the two traditional Chinese medicines, otherwise, the connecting edge is not established.
Preferably, in S3, the node embedding representation in the prescription iso-graph is learned through a graph convolution network, the graph convolution network captures neighbor information through a layer of convolution, and when the graph convolution network stacks multiple layers, high-order neighborhood information can be obtained;
inputting node attribute vectors at a first levelAccording to A i，j and Aj，i Obtaining an adjacent matrix A of FH, updating information of aggregation neighborhood nodes by using a traditional Chinese medicine prescription and traditional Chinese medicine nodes, and for a layer of graph rolling network, obtaining a k-dimensional node matrixFormalized representation of (c) is:
wherein ,is a normalized representation of the adjacency matrix a, +.>Representing the learned parameters, ρ being an activation function;
the multi-layer graph convolutional network is expressed as:
wherein l represents the number of layers, and H (0) ＝X。
Preferably, in S4, the traditional Chinese medicine prescription efficacy prediction module is specifically implemented as follows:
for each Chinese medicinal prescription, evaluating the distance between the predicted efficacy and the actual efficacy, setting the output dimension of the last layer of graph convolution network to be equal to the number of efficacy labels, and obtaining the final embedded representation of the Chinese medicinal prescription nodesAnd send it into Sigmoid classifier for learning to obtain probability distribution of Chinese medicine prescription on all effectsWherein C represents the number of efficacy tags, < + >>The calculation method is as follows:
preferably, in S4, the traditional Chinese medicine prescription efficacy prediction uses a multi-label cross entropy loss function as the loss function:
wherein n represents the number of training set samples, y (i) E {0,1} represents the actual efficacy label,representing the predicted value.
In order to verify the effectiveness of the traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network, experiments are carried out on 2274 first prescription, the prescription contains 539 Chinese medicines, and the basic information statistics of the prescription data set are shown in table 5. The prescription data were randomly divided into training and testing sets at a ratio of 8:2, and 20% of the training set was randomly selected as the validation set. A predictive model was built using two layers of GCN, the first layer GCN node embedding dimension set to 800 and the second layer GCN node embedding dimension set to 19. Setting β to 0.6, learning rate to 0.01, training the model for up to 500 rounds using Adam optimizer, preventing model over-fitting with early shutdown, and terminating training if the validation loss for consecutive 10 rounds is not reduced.
Table 5 formulation dataset basic information statistics
| Efficacy of | Traditional Chinese medicine | |
| Minimum number | 1 | 1 |
| Maximum number | 7 | 50 |
| Average number | 2.67 | 9.66 |
| Standard deviation of | 1.04 | 3.81 |
In the invention, three commonly used multi-label model evaluation indexes are selected to measure the model performance, namely, precision (P), F1 fraction (F1-score) and Hamming Loss (HL), and the calculation formulas of the indexes are as follows:
where m represents the number of test set samples, C represents the number of efficacy labels,the j-th label representing the i-th sample. In general, the higher the Precision and F1-score, the lower the Hamming Loss, representing the better the model prediction.
Comparative experiments were performed using the following models:
multilayer Perceptron (MLP): the layers are in full connection to transfer information, so that complex data such as nonlinearity, unbalance, small samples and the like can be processed.
Long Short-Term Memory (LSTM): the time cyclic neural network extracts text features in time, solves the problems of gradient disappearance and gradient explosion of the cyclic neural network, and has higher calculation efficiency.
Bi-directional Long Short-Term Memory (Bi-LSTM): the system consists of a bidirectional LSTM, solves the difficulty that the traditional LSTM model can not capture the context information due to the problem of serialization processing, and is often used for classifying tasks.
To minimize the effect of randomness on the experimental results, we performed 10 experiments on the data set of the chinese prescription and reported the mean plus-minus standard deviation, and the comparative experimental results are shown in table 5.
TABLE 5 experimental results of Chinese prescription data on different baseline models
| model | MLP | LSTM | Bi-LSTM | TCMGCN |
| P↑ | 0.7338±0.0138 | 0.7072±0.0209 | 0.7050±0.0149 | 0.7503±0.0114 |
| F1↑ | 0.6041±0.0074 | 0.6022±0.0109 | 0.6057±0.0094 | 0.6275±0.0130 |
| HL↓ | 0.0967±0.0020 | 0.0992±0.0022 | 0.0987±0.0022 | 0.0936±0.0030 |
Table 5 shows the experimental results of four methods on the data set of the Chinese medicine prescription, wherein TCMGCN is the method proposed by the invention. It can be seen that the TCMGCN achieves the best effect of multi-label classification over the other three baseline models in terms of the evaluation index accuracy, F1 score, and hamming loss evaluation index. Specifically, TCMGCN increased 2.25% over the strongest baseline MLP in the accuracy index, 3.60% over the strongest baseline Bi-LSTM in the F1 score index, and 3.21% over the strongest baseline MLP in the hamming loss index. The MLP, LSTM, bi-LSTM deep learning network models all achieve a considerable effect, and the fact that the traditional Chinese medicine prescription efficacy prediction task is regarded as a multi-label classification task is reasonable is explained. The reason why Bi-LSTM is slightly better than LSTM in F1 score and Hamming loss index is that Bi-LSTM can receive the context information of the sequence, and the features of deeper storage hierarchy and stronger resolution are stored. The accuracy and hamming loss index of the MLP are better than those of LSTM and Bi-LSTM, probably because the MLP targets the training process with updated network parameters, and the output is closer to the actual situation. The reason why TCMGCN is excellent may be that the prescription iso-graph can capture the relationship between the prescription and the traditional Chinese medicine, and the traditional Chinese medicine, and the graph volume lamination layer enables the label to be transmitted to the whole graph by aggregating the label node information and the label-free node information. In conclusion, the traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network has certain effectiveness.
Table 6 shows the efficacy predictions of TCMGCN for two specific examples, with bold fonts indicating the correct efficacy for the predictions. The first example inputs prescription information of astragalus (20 g), cassia twig (30 g), angelica (15 g), radix rehmanniae (15 g), achyranthes (30 g), ligusticum wallichii (15 g), radix salviae miltiorrhizae (15 g) and leech (10 g), and the method provided by the invention hits all efficacy labels. The second example inputs prescription information of aconite (30 g), grassleaf sweelflag rhizome (30 g), thinleaf milkwort root-bark (30 g), tall gastrodia tuber (30 g), scorpion (30 g), notopterygium root (30 g), stiff silkworm (30 g), arisaema tuber (30 g), and the method provided by the invention hits all efficacy labels except damp-dispelling. Therefore, the TCMGCN method provided by the invention has good practical application value, and can be used for reasonably predicting the efficacy of the traditional Chinese medicine prescription.
Table 6 case analysis
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. It will be apparent to those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.
Claims (6)
1. The traditional Chinese medicine prescription efficacy prediction method based on the graph convolution neural network is characterized by comprising the following steps of:
s1, a data preprocessing module preprocesses traditional Chinese medicine prescription data through a natural language processing technology, and builds a traditional Chinese medicine database and a prescription database which meet the data mining requirement;
s2, a prescription diagram construction module, which is used for constructing prescription iso-graphs containing traditional Chinese medicine nodes and prescription nodes by expressing prescription text data in a diagram form, wherein the construction of the prescription iso-graphs comprises the following steps:
S201，the Chinese medicine quantization representation converts the abstract description of Chinese medicine attribute into concrete numerical value as Chinese medicine attribute vectord represents the attribute quantity of the traditional Chinese medicine; the Chinese medicinal materials have properties including nature, taste, channel tropism and toxicity, and the nature, taste and toxicity are quantified by index level, respectively 2 -1 、2 0 、2 1 The expression "slightly cold", "severe cold"; binary quantization is adopted for the meridian tropism attribute, wherein the existence of the attribute is represented by 1, and the absence of the attribute is represented by 0;
s202, quantitatively representing the Chinese medicinal prescription, namely matching Chinese medicaments contained in the Chinese medicinal prescription to corresponding Chinese medicament attribute vectors, and summing the Chinese medicament attribute vectors to obtain the Chinese medicinal prescription attribute vector
S203, normalizing the traditional Chinese medicine dosage according to the following relative dosage conversion formula:
wherein ,represents the relative dose of the Chinese medicine i in the prescription j, d ij Represents the actual dosage of the traditional Chinese medicine i in the prescription j, d min and dmax Respectively representing the minimum value and the maximum value of the traditional Chinese medicine common dosage range;
s204, calculating the similarity of the traditional Chinese medicines according to the following formula:
wherein ,xa and xb Attribute vectors respectively representing the traditional Chinese medicine a and the traditional Chinese medicine b;
s205, constructing a prescription iso-graph FH= (V, E), wherein V is a node set in the graph, E is an edge set in the graph, and a characteristic matrix of the graphCorresponding adjacency matrix->n represents n nodes with characteristics, d represents node characteristic dimension, FH is represented in an undirected graph, and the adjacency matrix is represented as follows:
wherein beta is the threshold value of the connecting edge between the nodes of the traditional Chinese medicines, when the similarity of the two traditional Chinese medicines is larger than beta, the connecting edge is established between the two traditional Chinese medicines, otherwise, the connecting edge is not established;
s3, a node representation learning module learns node embedding representation in the prescription iso-graph through a graph convolution network for use by downstream tasks;
s4, a prediction module is used for obtaining probability distribution of the traditional Chinese medicine prescription on the effect based on the final node embedded representation obtained through training of the traditional Chinese medicine prescription sample.
2. The method of claim 1, wherein the database of traditional Chinese medicines comprises drug names, aliases, sex flavors including cold, heat, warm, benign, flat, acid, bitter, sweet, pungent and salty, menstruation including lung, pericardium, heart, large intestine, triple-focus, small intestine, stomach, gall bladder, spleen, liver and kidney;
the prescription database comprises prescription names, traditional Chinese medicine components, actual dosage and efficacy information; the efficacy information comprises a dryness treating prescription, a tonifying prescription, a blood regulating prescription, a wind dispelling prescription, a phlegm eliminating prescription, a carbuncle and ulcer prescription, a damp eliminating prescription, a qi regulating prescription, wen Lifang, a resolving prescription, a heat clearing prescription, an emergency prescription, a nerve soothing prescription, a eyesight improving prescription, a bleeding prescription, a resuscitation inducing prescription, a digestion promoting prescription, a astringing prescription and a exterior relieving prescription.
3. The prediction method according to claim 2, wherein in S1, the pretreatment of the prescription data of the traditional Chinese medicine comprises:
s101, chinese medicine names are standardized, and the Chinese medicine names in the Chinese medicine prescription are subjected to homonymous substitution through a character string matching technology;
s102, modern conversion of dosage, namely splitting the name of the traditional Chinese medicine and the dosage, and converting calculation of different dosage units of the traditional Chinese medicine in the traditional Chinese medicine prescription into modern dosage unit grams.
4. A prediction method according to claim 3, wherein in S3, the node embedding representation in the prescription iso-graph is learned by a graph convolution network, the graph convolution network captures neighbor information by a layer of convolution, and when the graph convolution network stacks multiple layers, high-order neighborhood information can be obtained;
inputting node attribute vectors at a first levelAccording to A i，j and Aj，i Obtaining an adjacent matrix A of FH, updating information of aggregation neighborhood nodes by using a traditional Chinese medicine prescription and traditional Chinese medicine nodes, and regarding a layer of graph rolling network, K-dimensional node matrix +.>Formalized representation of (c) is:
wherein ,is a normalized representation of the adjacency matrix a, +.>Representing the learned parameters, ρ being an activation function;
the multi-layer graph convolutional network is expressed as:
wherein l represents the number of layers, and H (0) ＝X。
5. The prediction method according to claim 4, wherein in S4, the traditional Chinese medicine prescription efficacy prediction module is specifically implemented as follows:
for each Chinese medicinal prescription, evaluating the distance between the predicted efficacy and the actual efficacy, setting the output dimension of the last layer of graph convolution network to be equal to the number of efficacy labels, and obtaining the final embedded representation of the Chinese medicinal prescription nodesAnd send it into Sigmoid classifier for learning to obtain probability distribution of Chinese medicine prescription on all effectsWherein C represents the number of efficacy tags, < + >>The calculation method is as follows:
6. the method according to claim 5, wherein in S4, the traditional Chinese medicine prescription efficacy prediction uses a multi-label cross entropy loss function as the loss function:
wherein n represents the number of training set samples, y (i) E {0,1} represents the actual efficacy label,representing the predicted value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210683366.8A CN115050481B (en) | 2022-06-17 | 2022-06-17 | Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210683366.8A CN115050481B (en) | 2022-06-17 | 2022-06-17 | Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115050481A CN115050481A (en) | 2022-09-13 |
| CN115050481B true CN115050481B (en) | 2023-10-31 |
Family
ID=83160800
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210683366.8A Active CN115050481B (en) | 2022-06-17 | 2022-06-17 | Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115050481B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115472241B (en) * | 2022-09-23 | 2023-07-14 | 湖南中医药大学 | Chinese medicine component cluster menstruation determining method based on chemical structure topology index comparison and chromatographic imprinting measurement |
| CN116525100B (en) * | 2023-04-26 | 2024-08-23 | 脉景（杭州）健康管理有限公司 | Traditional Chinese medicine prescription reverse verification method and system based on label system |
| CN117079762B (en) * | 2023-09-25 | 2024-01-23 | 腾讯科技（深圳）有限公司 | Drug effect prediction model training method, drug effect prediction method and device thereof |
| CN118098435B (en) * | 2024-02-04 | 2024-09-20 | 中央民族大学 | Method and system for predicting efficacy of medicine |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106803012A (en) * | 2016-12-29 | 2017-06-06 | 杭州师范大学钱江学院 | Prescription function prediction method based on probability topic model and Chinese medicine base attribute |
| CN110981021A (en) * | 2019-12-23 | 2020-04-10 | 中新国际联合研究院 | Intelligent dosing system and method for advanced wastewater oxidation treatment based on fuzzy BP neural network |
| CN111312356A (en) * | 2020-01-17 | 2020-06-19 | 四川大学 | Traditional Chinese medicine prescription generation method based on BERT and integration efficacy information |
| CN113539412A (en) * | 2021-07-19 | 2021-10-22 | 闽江学院 | Chinese herbal medicine recommendation system based on deep learning |
| CN114255851A (en) * | 2021-12-22 | 2022-03-29 | 闽江学院 | Chinese herbal medicine prescription recommendation method based on hypergraph neural network |
| CN114547298A (en) * | 2022-02-14 | 2022-05-27 | 大连理工大学 | Biomedical relation extraction method, device and medium based on combination of multi-head attention and graph convolution network and R-Drop mechanism |
| CN114582497A (en) * | 2022-03-04 | 2022-06-03 | 闽江学院 | Traditional Chinese medicine syndrome type classification method based on graph convolution neural network |
| CN114613425A (en) * | 2022-03-10 | 2022-06-10 | 中国石油大学(华东) | Drug-target interaction prediction algorithm based on graph volume and similarity |
-
2022
- 2022-06-17 CN CN202210683366.8A patent/CN115050481B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106803012A (en) * | 2016-12-29 | 2017-06-06 | 杭州师范大学钱江学院 | Prescription function prediction method based on probability topic model and Chinese medicine base attribute |
| CN110981021A (en) * | 2019-12-23 | 2020-04-10 | 中新国际联合研究院 | Intelligent dosing system and method for advanced wastewater oxidation treatment based on fuzzy BP neural network |
| CN111312356A (en) * | 2020-01-17 | 2020-06-19 | 四川大学 | Traditional Chinese medicine prescription generation method based on BERT and integration efficacy information |
| CN113539412A (en) * | 2021-07-19 | 2021-10-22 | 闽江学院 | Chinese herbal medicine recommendation system based on deep learning |
| CN114255851A (en) * | 2021-12-22 | 2022-03-29 | 闽江学院 | Chinese herbal medicine prescription recommendation method based on hypergraph neural network |
| CN114547298A (en) * | 2022-02-14 | 2022-05-27 | 大连理工大学 | Biomedical relation extraction method, device and medium based on combination of multi-head attention and graph convolution network and R-Drop mechanism |
| CN114582497A (en) * | 2022-03-04 | 2022-06-03 | 闽江学院 | Traditional Chinese medicine syndrome type classification method based on graph convolution neural network |
| CN114613425A (en) * | 2022-03-10 | 2022-06-10 | 中国石油大学(华东) | Drug-target interaction prediction algorithm based on graph volume and similarity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115050481A (en) | 2022-09-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115050481B (en) | Traditional Chinese medicine prescription efficacy prediction method based on graph convolution neural network | |
| Lin et al. | User-level psychological stress detection from social media using deep neural network | |
| CN107016438B (en) | System based on traditional Chinese medicine syndrome differentiation artificial neural network algorithm model | |
| CN111414393A (en) | Semantic similar case retrieval method and equipment based on medical knowledge graph | |
| CN109903854A (en) | A kind of core drug recognition methods based on TCM Literature | |
| Jiang et al. | A hybrid intelligent model for acute hypotensive episode prediction with large-scale data | |
| Yao et al. | Discovering treatment pattern in Traditional Chinese Medicine clinical cases by exploiting supervised topic model and domain knowledge | |
| CN109947901B (en) | Prescription efficacy prediction method based on multilayer perceptron and natural language processing technology | |
| Liu et al. | A novel transfer learning model for traditional herbal medicine prescription generation from unstructured resources and knowledge | |
| Hasan et al. | Integrating text embedding with traditional nlp features for clinical relation extraction | |
| Yu et al. | Identification of pediatric respiratory diseases using a fine-grained diagnosis system | |
| CN110827990B (en) | Typhoid fever syndrome differentiation reasoning system based on knowledge graph | |
| Cheng et al. | An improved deep learning model: S-TextBLCNN for traditional Chinese medicine formula classification | |
| CN114822874B (en) | Prescription efficacy classification method based on characteristic deviation alignment | |
| CN114141361B (en) | Traditional Chinese medicine prescription recommendation method based on symptom term mapping and deep learning | |
| CN113539412B (en) | Deep learning-based Chinese herbal medicine recommendation system | |
| Ren et al. | The intelligent experience inheritance system for Traditional Chinese medicine | |
| Zhao et al. | PreGenerator: TCM prescription recommendation model based on retrieval and generation method | |
| CN115631852B (en) | Certificate type recommendation method and device, electronic equipment and nonvolatile storage medium | |
| Yang et al. | Multi-label rhinitis prediction using ensemble neural network chain with pre-training | |
| CN116682536A (en) | Traditional Chinese medicine prescription recommendation method based on improved Bert4Rec | |
| Yan et al. | Graph-based medicine embedding learning via multiple attentions | |
| Zhu et al. | Herbkg: Constructing a herbal-molecular medicine knowledge graph using a two-stage framework based on deep transfer learning | |
| Zhang et al. | A hierarchical symptom-herb topic model for analyzing traditional Chinese medicine clinical diabetic data | |
| Xie et al. | Predicting the risk of stroke based on imbalanced data set with missing data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |