CN110289106A - A method of effect, which is analyzed, from Chinese medicine compound prescription corresponds to Chinese medicine and its pharmacological property compatibility relationship - Google Patents
A method of effect, which is analyzed, from Chinese medicine compound prescription corresponds to Chinese medicine and its pharmacological property compatibility relationship Download PDFInfo
- Publication number
- CN110289106A CN110289106A CN201910576633.XA CN201910576633A CN110289106A CN 110289106 A CN110289106 A CN 110289106A CN 201910576633 A CN201910576633 A CN 201910576633A CN 110289106 A CN110289106 A CN 110289106A
- Authority
- CN
- China
- Prior art keywords
- chinese medicine
- effect
- compound
- flavor
- nature
- 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.)
- Granted
Links
- 239000003814 drug Substances 0.000 title claims abstract description 227
- 150000001875 compounds Chemical class 0.000 title claims abstract description 112
- 230000000694 effects Effects 0.000 title claims abstract description 100
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000000144 pharmacologic effect Effects 0.000 title claims abstract description 32
- 239000000796 flavoring agent Substances 0.000 claims abstract description 86
- 235000019634 flavors Nutrition 0.000 claims abstract description 85
- 230000010415 tropism Effects 0.000 claims abstract description 59
- 229940079593 drug Drugs 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000013480 data collection Methods 0.000 claims abstract description 6
- 238000000605 extraction Methods 0.000 claims abstract description 5
- 235000019640 taste Nutrition 0.000 claims description 35
- 238000009826 distribution Methods 0.000 claims description 27
- 229940126678 chinese medicines Drugs 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 15
- 239000000203 mixture Substances 0.000 claims description 13
- 239000012141 concentrate Substances 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 235000009508 confectionery Nutrition 0.000 claims description 10
- 238000012986 modification Methods 0.000 claims description 5
- 230000004048 modification Effects 0.000 claims description 5
- 210000000056 organ Anatomy 0.000 claims description 4
- 231100000440 toxicity profile Toxicity 0.000 claims description 4
- 230000009471 action Effects 0.000 claims description 3
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 abstract description 12
- 238000005065 mining Methods 0.000 abstract description 7
- 238000007876 drug discovery Methods 0.000 abstract description 2
- 239000002547 new drug Substances 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000002474 experimental method Methods 0.000 description 7
- 230000003213 activating effect Effects 0.000 description 6
- 210000004185 liver Anatomy 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 210000000952 spleen Anatomy 0.000 description 6
- 241000205585 Aquilegia canadensis Species 0.000 description 5
- 240000004980 Rheum officinale Species 0.000 description 5
- 235000008081 Rheum officinale Nutrition 0.000 description 5
- 230000004089 microcirculation Effects 0.000 description 5
- 230000001737 promoting effect Effects 0.000 description 5
- 230000001502 supplementing effect Effects 0.000 description 5
- 206010004542 Bezoar Diseases 0.000 description 4
- 239000009636 Huang Qi Substances 0.000 description 4
- 208000004880 Polyuria Diseases 0.000 description 4
- 230000035619 diuresis Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 241000411851 herbal medicine Species 0.000 description 4
- 210000002429 large intestine Anatomy 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 230000008961 swelling Effects 0.000 description 4
- DTGKSKDOIYIVQL-WEDXCCLWSA-N (+)-borneol Chemical compound C1C[C@@]2(C)[C@@H](O)C[C@@H]1C2(C)C DTGKSKDOIYIVQL-WEDXCCLWSA-N 0.000 description 3
- REPVLJRCJUVQFA-UHFFFAOYSA-N (-)-isopinocampheol Natural products C1C(O)C(C)C2C(C)(C)C1C2 REPVLJRCJUVQFA-UHFFFAOYSA-N 0.000 description 3
- 244000144730 Amygdalus persica Species 0.000 description 3
- 244000197580 Poria cocos Species 0.000 description 3
- 235000008599 Poria cocos Nutrition 0.000 description 3
- 235000006040 Prunus persica var persica Nutrition 0.000 description 3
- 241000245665 Taraxacum Species 0.000 description 3
- 235000005187 Taraxacum officinale ssp. officinale Nutrition 0.000 description 3
- CKDOCTFBFTVPSN-UHFFFAOYSA-N borneol Natural products C1CC2(C)C(C)CC1C2(C)C CKDOCTFBFTVPSN-UHFFFAOYSA-N 0.000 description 3
- 229940116229 borneol Drugs 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 3
- DTGKSKDOIYIVQL-UHFFFAOYSA-N dl-isoborneol Natural products C1CC2(C)C(O)CC1C2(C)C DTGKSKDOIYIVQL-UHFFFAOYSA-N 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000000717 retained effect Effects 0.000 description 3
- 244000061520 Angelica archangelica Species 0.000 description 2
- 241000717739 Boswellia sacra Species 0.000 description 2
- 241000208671 Campanulaceae Species 0.000 description 2
- 235000003255 Carthamus tinctorius Nutrition 0.000 description 2
- 244000020518 Carthamus tinctorius Species 0.000 description 2
- 235000006965 Commiphora myrrha Nutrition 0.000 description 2
- 235000002991 Coptis groenlandica Nutrition 0.000 description 2
- 244000247747 Coptis groenlandica Species 0.000 description 2
- 239000004863 Frankincense Substances 0.000 description 2
- 244000111489 Gardenia augusta Species 0.000 description 2
- 235000018958 Gardenia augusta Nutrition 0.000 description 2
- 235000001287 Guettarda speciosa Nutrition 0.000 description 2
- 244000153234 Hibiscus abelmoschus Species 0.000 description 2
- 235000007265 Myrrhis odorata Nutrition 0.000 description 2
- UFWIBTONFRDIAS-UHFFFAOYSA-N Naphthalene Chemical compound C1=CC=CC2=CC=CC=C21 UFWIBTONFRDIAS-UHFFFAOYSA-N 0.000 description 2
- 244000046146 Pueraria lobata Species 0.000 description 2
- 235000010575 Pueraria lobata Nutrition 0.000 description 2
- 206010044565 Tremor Diseases 0.000 description 2
- 239000008280 blood Substances 0.000 description 2
- 210000004369 blood Anatomy 0.000 description 2
- 230000001914 calming effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 210000003734 kidney Anatomy 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 210000000582 semen Anatomy 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 235000014347 soups Nutrition 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 208000011580 syndromic disease Diseases 0.000 description 2
- 230000035922 thirst Effects 0.000 description 2
- 229940126680 traditional chinese medicines Drugs 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- DSSYKIVIOFKYAU-XCBNKYQSSA-N (R)-camphor Chemical compound C1C[C@@]2(C)C(=O)C[C@@H]1C2(C)C DSSYKIVIOFKYAU-XCBNKYQSSA-N 0.000 description 1
- 235000009051 Ambrosia paniculata var. peruviana Nutrition 0.000 description 1
- 235000003097 Artemisia absinthium Nutrition 0.000 description 1
- 240000001851 Artemisia dracunculus Species 0.000 description 1
- 235000017731 Artemisia dracunculus ssp. dracunculus Nutrition 0.000 description 1
- 235000003261 Artemisia vulgaris Nutrition 0.000 description 1
- 241000432824 Asparagus densiflorus Species 0.000 description 1
- 235000011274 Benincasa cerifera Nutrition 0.000 description 1
- 244000036905 Benincasa cerifera Species 0.000 description 1
- 241000258920 Chilopoda Species 0.000 description 1
- 241000723346 Cinnamomum camphora Species 0.000 description 1
- 240000001579 Cirsium arvense Species 0.000 description 1
- 235000005918 Cirsium arvense Nutrition 0.000 description 1
- 240000007311 Commiphora myrrha Species 0.000 description 1
- 241000721047 Danaus plexippus Species 0.000 description 1
- 235000013535 Dianthus superbus Nutrition 0.000 description 1
- 244000183914 Dianthus superbus Species 0.000 description 1
- 240000005717 Dioscorea alata Species 0.000 description 1
- 235000002722 Dioscorea batatas Nutrition 0.000 description 1
- 235000006536 Dioscorea esculenta Nutrition 0.000 description 1
- 240000001811 Dioscorea oppositifolia Species 0.000 description 1
- 235000003416 Dioscorea oppositifolia Nutrition 0.000 description 1
- 244000286838 Eclipta prostrata Species 0.000 description 1
- 241000246500 Gentianella aspera Species 0.000 description 1
- 241000825107 Hierochloe Species 0.000 description 1
- 235000015466 Hierochloe odorata Nutrition 0.000 description 1
- 241000830535 Ligustrum lucidum Species 0.000 description 1
- 240000009023 Myrrhis odorata Species 0.000 description 1
- 240000001659 Oldenlandia diffusa Species 0.000 description 1
- 235000003889 Paeonia suffruticosa Nutrition 0.000 description 1
- 240000005001 Paeonia suffruticosa Species 0.000 description 1
- 235000003143 Panax notoginseng Nutrition 0.000 description 1
- 241000180649 Panax notoginseng Species 0.000 description 1
- 241000405414 Rehmannia Species 0.000 description 1
- 235000008422 Schisandra chinensis Nutrition 0.000 description 1
- 240000006079 Schisandra chinensis Species 0.000 description 1
- 241000522620 Scorpio Species 0.000 description 1
- 208000037386 Typhoid Diseases 0.000 description 1
- 241001302463 Viola mandshurica Species 0.000 description 1
- 239000001138 artemisia absinthium Substances 0.000 description 1
- 235000019606 astringent taste Nutrition 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000001124 body fluid Anatomy 0.000 description 1
- 239000010839 body fluid Substances 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 229930008380 camphor Natural products 0.000 description 1
- 229960000846 camphor Drugs 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000001808 coupling effect Effects 0.000 description 1
- 239000006071 cream Substances 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 210000000232 gallbladder Anatomy 0.000 description 1
- 239000008838 gegenqinlian Substances 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 235000008216 herbs Nutrition 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 239000008629 longdanxiegan Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 210000004072 lung Anatomy 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- ALHUZKCOMYUFRB-UHFFFAOYSA-N muskone Natural products CC1CCCCCCCCCCCCC(=O)C1 ALHUZKCOMYUFRB-UHFFFAOYSA-N 0.000 description 1
- 239000002674 ointment Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002574 poison Substances 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000010791 quenching Methods 0.000 description 1
- 229910052957 realgar Inorganic materials 0.000 description 1
- 235000019991 rice wine Nutrition 0.000 description 1
- 239000012266 salt solution Substances 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000009490 scorpio Substances 0.000 description 1
- 210000002784 stomach Anatomy 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- KDYFGRWQOYBRFD-UHFFFAOYSA-N succinic acid Chemical compound OC(=O)CCC(O)=O KDYFGRWQOYBRFD-UHFFFAOYSA-N 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 231100000331 toxic Toxicity 0.000 description 1
- 230000002588 toxic effect Effects 0.000 description 1
- 231100000419 toxicity Toxicity 0.000 description 1
- 230000001988 toxicity Effects 0.000 description 1
- 201000008297 typhoid fever Diseases 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Data Mining & Analysis (AREA)
- Toxicology (AREA)
- Artificial Intelligence (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Medicinal Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Pharmacology & Pharmacy (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Medicines Containing Plant Substances (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a kind of to analyze the method that effect corresponds to Chinese medicine and its pharmacological property compatibility relationship from Chinese medicine compound prescription, this method comprises: (1) obtains pending data collection, data set includes Chinese Traditional Medicine information data set and compound message data set；(2) expansion of nature and flavor and channel tropism is carried out to pretreated compound message data set using pretreated Chinese Traditional Medicine information data set；(3) multi-dimensional relation between compound information data concentration Chinese medicine, effect and nature and flavor after expanding is taken out, building combines topic model, and solves the model parameter in conjunction with topic model；(4) the compound message data set after extraction section expands is trained as training set in conjunction with topic model, until reaching the number of iterations of setting, and the pro-bability value matrices of effect and Chinese medicine, nature and flavor and channel tropism is exported according to training result.The present invention can go out the multidimensional compatibility relationship recessiveness relationship that effect corresponds to Chinese medicine and its pharmacological property by mining analysis from a large amount of Chinese medicine compound prescriptions, can be widely used for instructing the automatic prescription of clinical application, computer and new drug discovery.
Description
Technical field
The present invention relates to text data digging analysis fields, and in particular to a kind of from analysis effect corresponds in Chinese medicine compound prescription
The method of medicine and its pharmacological property compatibility relationship.
Background technique
The effect of Western medicine and effect are mainly reflected in the different chemical composition of drug.In comparison, before being born in thousands of years
Chinese medicine be we ancestors in practice constantly repeatedly verify, summarize and in-depth obtain various Chinese herbal medicine sex knowledge.In
Medicine property theory passes through long-term development, has formd more comprehensive summary, thousands of singles to the pharmacological property knowledge of single medicinal material
The specific nature and flavor of Chinese medicine, channel tropism toxicity etc. form comprehensive detailed property of traditional Chinese knowledge base.But Chinese medicine is in use,
Single medicinal material exclusive use rate is very low, and majority of case all forms multiple efficacy Treatments diseases using multi-flavor medicine compatibility in compound
Card.According to Chinese medicine correlation theory, the whole nature and flavor that kinds of traditional Chinese medicines compatibility generates are the key that form its effect.
Enumerate several effects and kinds of traditional Chinese medicines of the compound in the Chinese medicine compound prescription specification that usually we see, and effect
Corresponding Chinese medicine and its pharmacological property compatibility relationship are as implicit knowledge.This implicit knowledge is exactly study of Chinese medicine, and person spends many decades
The key and problem of time comprehension, study.
With the development of the relevant technologies such as data mining, machine learning, analysis mining Chinese medicine is matched from a large amount of Chinese medicine compound prescriptions
5 rules, explore pharmacological property Compatibility Law, disclose effect and its corresponding drug matching, and to improving, drug matching is theoretical, guidance is clinical
Medication, new drug discovery and automatic side's group etc. all have important value.Current existing middle compound analysis mining method mostly uses
General method, such as classification, principal component analysis, Association Rule Analysis fail to combine the progress of Chinese medicine pharmacological property correlation theory above-mentioned
The mining analysis of multidimensional recessiveness relationship, resulting result popularity, practicability and accuracy are poor.
Summary of the invention
Goal of the invention: for overcome the deficiencies in the prior art, the present invention provides one kind and analyzes effect pair from Chinese medicine compound prescription
The method for answering Chinese medicine and its pharmacological property compatibility relationship, this method solve currently fail that Chinese medicine pharmacological property correlation theory is combined to carry out multidimensional
The problem of the mining analysis of recessive relationship and resulting result popularity, practicability and accuracy difference.
Technical solution: one kind of the present invention analyzes effect from Chinese medicine compound prescription and corresponds to Chinese medicine and its pharmacological property compatibility relationship
Method, this method comprises:
(1) pending data collection is obtained, the data set includes Chinese Traditional Medicine information data set and compound message data set, and right
The Chinese Traditional Medicine information data set and compound message data set are pre-processed respectively；
(2) nature and flavor are carried out to pretreated compound message data set using pretreated Chinese Traditional Medicine information data set and returned
The expansion of warp；
(3) the compound information data after taking out the expansion concentrates the multi-dimensional relation between Chinese medicine, effect and nature and flavor, structure
It builds and combines topic model, and solve the model parameter of the combination topic model；
(4) the compound message data set after extraction section expands as training set, to the combination topic model into
Row training until reaching the number of iterations of setting, and exports the probability value of effect and Chinese medicine, nature and flavor and channel tropism according to training result
Matrix.
Further, comprising:
Each compound information data that the compound information data is concentrated includes effect and drug composition, the Chinese Traditional Medicine information
Each Chinese Traditional Medicine information data in data set include medicine name, alias, nature and flavor and channel tropism, and the nature and flavor include four gas and the five tastes, institute
State four gas include: it is hot, warm, cold, cool, the five tastes include pungent, sweet, sour, bitter, salty.
Further, comprising:
In the step (1), Chinese medicine message data set is pre-processed, comprising:
(11) the nature and flavor description information in Chinese Traditional Medicine information is filtered and is replaced, only retained the five tastes and four gas and its add and repair
The nature and flavor of micro-, big, light, the small composition of excuse；
(12) only retain the description of channel tropism in Chinese Traditional Medicine information the direct statement word of the vital organs of the human body.
Further, comprising:
It is described that Chinese medicine message data set is pre-processed further include: will to include that micro-, light, small degree adverb is modified in nature and flavor
The Term co-occurrence frequency on the basis of, be denoted as 1；3 times of former word frequency are upgraded to comprising degree adverb big word frequency time；There is no degree adverb to repair
The word frequency of the nature and flavor of decorations is upgraded to 2 times of former word frequency.
Further, comprising:
In the step (3), the compound information data taken out after expanding is concentrated between Chinese medicine, effect and nature and flavor
Multi-dimensional relation specifically includes:
It is one-to-one recessive relationship between effect and medicine group；It is one-to-many recessive relationship between medicine group and Chinese medicine；In
There are the dominance relations of multi-to-multi between medicine and nature and flavor；There are the recessive relationships of indirect multi-to-multi between medicine group and nature and flavor；Effect
There are indirect multi-to-multi recessiveness relationships between nature and flavor.
Further, comprising:
In the step (3), building combines topic model, comprising:
Compound message data set after expansion regards D independent text generation processes, single compound information data conduct as
Total bag of words of one document d, JTM include three sub- bag of words, are respectively as follows: Chinese medicine bag of words, concentrate and occur comprising compound information data
All Chinese medicines；Chinese medicine nature and flavor bag of words, including four gas, the five tastes and toxicity profile；Channel tropism bag of words refer to Chinese medicine according to zang-fu differentiation
Theory action is in the statement word of human body；
One or more effects of the document d are regarded as the K theme label of document d, and K indicates effect type；Using to
Measure Λd=(t1,t2,...,tK) indicate effect t in a series of binary number instruction document diWhether occur, ti∈ { 0,1 }, 0
It indicates occur label l in d, does not occur l in 1 expression d；ΛdObedience priori is that the Bernoulli Jacob of Beta (γ) is distributed；ΛdDominant restraint
The theme distribution θ of document dd；θdThe effect of for document d ratio, be multinomial distribution, priori obeys Cray in Di that parameter is α
Distribution, it may be assumed that θd~Dirichlet (α)；zdnBy multinomial distribution θdIt generates, wherein zdn∈{0,…K-1}；In n-th of document d
Medicine hdn, with it for four gas gdn, five tastes fdn, channel tropism tdn, share effect zdn；WithIt is multinomial distribution, takes respectively
It is β from priorih、βfAnd βtDirichlet distribution, respectively indicate Chinese medicine under theme, nature and flavor and channel tropism distribution.
Further, comprising:
The generating probability formula of the document d indicates are as follows:
Wherein, NfIndicate the nature and flavor number of the n-th taste Chinese medicine in document d, NtIndicate the channel tropism of the n-th taste Chinese medicine in document d
Number, NhIndicate the sum of Chinese medicine in document d.
Further, comprising:
In the step (3), the model parameter of the combination topic model is solved, comprising: the k a certain type the effect of
Under, each parameter is expressed as:
Wherein, k ∈ { 1 ..., K }, | H |, | F | and | T | indicate that all Chinese medicines, nature and flavor and the channel tropism that occur in data set are total
Number,It indicates after removing current location i, Chinese medicine h is assigned the number of effect k, and i indicates that current location Chinese medicine ,-i indicate to remove
Current location i,Expression is gone out after the i of current location, and nature and flavor f is assigned the number of effect k,Expression is gone out current location i
Afterwards, channel tropism t is assigned the number of effect k.
Further, comprising:
In the step (4), the compound message data set cooperation after extraction section expands is training set, to the knot
Topic model is closed to be trained, comprising:
Input: compound message data set, JTM model hyper parameter α, β and training the number of iterations after expansion；
Step: (1) initialize: ergodic data concentrates every Chinese medicine h of each document ddn, from the Λ of the compounddCollection
An effect is randomly assigned in conjunction to number to hdn, while hdnCorresponding all nature and flavor fdnWith channel tropism tdnShare effect number；
(2) every Chinese medicine h of each compound d in data set is iterated overdnUntil the number of iterations of setting: to each
hdnIt calculates it and belongs to ΛdEach effect probability and carry out corresponding gibbs sampler.
Output: effect number two-dimensional matrix z corresponding to every Chinese medicine in the training set in all compounds.
Further, comprising:
The pro-bability value matrices that effect and Chinese medicine, nature and flavor and channel tropism are exported according to training result, comprising:
Calculate that gained institute is powerful and the pro-bability value matrices of Chinese medicine according to the resulting matrix z of training result and formula (3)
Wherein, NhThe unduplicated Chinese medicine number occurred for compound message data set.
Calculate that gained institute is powerful and the pro-bability value matrices of nature and flavor according to the resulting matrix z of training result and formula (4)
Wherein, NfThe unduplicated Chinese medicine nature and flavor number occurred for compound message data set.
Calculate that gained institute is powerful and the pro-bability value matrices of channel tropism according to the resulting matrix z of training result and formula (5)
Wherein, NtFor the number for the unduplicated meridian tropism that compound message data set occurs.
The utility model has the advantages that compared with prior art, the invention has the following advantages that (1) present invention combine Chinese medicine correlation theory,
It is put forward for the first time and analysis mining again after pharmacological property knowledge expansion is carried out to compound Chinese prescription, and then taken out that meet theory of traditional Chinese medical science actual
Multidimensional recessiveness relationship；(2) present invention construct in conjunction with topic model (JTM) to effect in compound Chinese prescription data, Chinese medicine, nature and flavor and
Channel tropism carries out multidimensional recessiveness relationship modeling, and the multi-dimensional relation analysis mining between them can be achieved at the same time；(3) existing side is compared
Method frequently with such as classification, principal component analysis, Association Rule Analysis, this method more suits theory of traditional Chinese medical science, resulting result
Popularity, practicability and accuracy are higher.
Detailed description of the invention
Fig. 1 is method flow diagram of the present invention；
Fig. 2 is that compound information data of the present invention concentrates Chinese medicine-pharmacological property-effect relation schematic diagram；
Fig. 3 is the graph model schematic diagram of JTM of the present invention.
Specific embodiment
The present invention will be describe below in further detail with reference to the accompanying drawings, it is clear that described embodiment is only this
Invention a part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art exist
All other embodiment obtained under the premise of creative work is not made, shall fall within the protection scope of the present invention.
The present invention provides a kind of method analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship, including
Once step:
S1 obtains pending data collection, and the data set includes Chinese Traditional Medicine information data set and compound message data set,
And the Chinese Traditional Medicine information data set and compound message data set are pre-processed respectively；
Pending data collection is collected in hundred TCM Databases of Chinese medicine e, hundred prescription database of Chinese medicine e, hundred proved recipe data of Chinese medicine e
Library, data center, traditional Chinese medicine section and prescription modern Application database, wherein Chinese Traditional Medicine information data set is from hundred Chinese medicine of Chinese medicine e
Database, compound message data set are collected in hundred prescription database of Chinese medicine e, hundred proved recipe database of Chinese medicine e, in traditional Chinese medicine section data
The heart and prescription modern Application database.
To the pretreated step of compound message data set are as follows: remove the prescription of missing efficacy information；Chinese medicine constructs special term
Table constitutes branch to recipe drug and segments；Lay down a regulation the various dosage informations removed in compound message data set composition,
Such as: 1 liang, 3-10g etc., and indicate the word of concocting method, and such as: raw, soup bubble, it is vinegar-fried；To all Chinese medicines medicine of the same name in prescription
Object carries out unified replacement；In order to guarantee enough co-occurrence informations, removal compound information data concentrates effect frequency of occurrence less than 5
All prescription data；In order to guarantee the multinomial distribution in model, removal efficiency is less than 2 or Chinese medicine constitutes the side less than 3 kinds
Agent.
Chinese Traditional Medicine information data set pre-treatment step are as follows:
In order to which vocabulary is stated unified, property of traditional Chinese taste description information is filtered and is replaced, only retained sour, bitter, sweet, pungent, salty
With it is cold, hot, warm, cool and its add micro-, the large and small composition of qualifier nature and flavor, such as: acrid flavour；Slightly warm in nature is pungent after replacement；Tepor；
In order to which vocabulary is stated unified, meridian tropism is described, only retains the directly statement word such as vital organs of the human body, such as: large intestine channel replaces
It is changed to large intestine.
The compound of compound message data set has been marked, and two or more correspond to effect；Comprising the above Chinese medicine of 3 tastes,
As shown in table 1.
1 Chinese medicine compound prescription example of table
S2 expands the nature and flavor information that compound information data concentrates every taste Chinese medicine using pretreated Chinese Traditional Medicine information data set
With channel tropism information；
Expand the four natures and five flavors of drug and channel tropism information of the nature and flavor of Chinese medicine in compound, four gas are main are as follows: hot, warm, cold, cool；The five tastes
Mainly are as follows: it is pungent, sweet, sour, bitter, salty, as shown in table 2.
2 compound Chinese prescription data nature and flavor of table expand example
S3 determines that the compound information data after the expansion concentrates the relationship between Chinese medicine, effect and nature and flavor, and building combines
Topic model, and solve the model parameter of the combination topic model；
Further, the S3 includes:
S31 building combines topic model (JTM): after compound message data set information expansion as shown in Table 2, one multiple
Side often have effects that one or several；A compound is by several taste Chinese medicinal compositions simultaneously, wherein every Chinese medicine has again
Meridian distribution of property and flavor alone.
Above-mentioned relation can be illustrated as shown in Figure 2.It can further analyze to obtain in Chinese medicine compound prescription by Fig. 2 and exist with ShiShimonoseki
System:
Corresponding relationship between effect and medicine group is one-to-one recessive relationship；
It is one-to-many recessive relationship between medicine group and Chinese medicine；
There are the dominance relations of multi-to-multi between Chinese medicine and pharmacological property；
There are the recessive relationships of indirect multi-to-multi between medicine group and pharmacological property；
There are indirect multi-to-multi recessiveness relationships between effect and pharmacological property.
Based on theme Modeling Theory, the combination theme of the Chinese medicine, nature and flavor and effect relationship that generate compound Chinese prescription is established
Model (Join Topic Model, JTM).
In JTM, Chinese medicine compound prescription data set regards D independent text generation processes as, and single Chinese medicine compound prescription is counted as one
A document d；Total bag of words of JTM are made of three seeds " bag of words " respectively:
First seed " bag of words " is Chinese medicine bag of words, includes all Chinese medicines occurred in data set；
Second seed " bag of words " is Chinese medicine nature and flavor bag of words, is referred to:
1) four gas (also known as four property), comprising hot, warm, cold, cool and four gas further according to degree use the words such as big, micro- into
The statement word of row refinement, such as: it is Great Cold, big heat, tepor, slightly cold；
2) five tastes mainly include pungent, sweet, sour, bitter, salty, and in addition there are other pharmacological properties such as derivative levelling；
3) toxicity profile；Third seed " bag of words " is that channel tropism " bag of words " are primarily referred to as Chinese medicine according to zang-fu differentiation theory action
In the statement word of human body.
Data after compound information data expansion nature and flavor and channel tropism model as shown in figure 3, using to it and generating model.
Each Chinese medicine compound prescription d effect is regarded as one or more demonstration theme labels of d by JTM；
Data set shares K theme label, it may be assumed that K kind effect；
The nature and flavor and channel tropism of prescription and expansion are regarded as the content of document d；
The generating process of document d is after determining K theme ratio, it may be assumed that K kind effect is extracted respectively in above-mentioned three kinds of bag of words
The process that three kinds of words combine, specific graph model indicate as shown in Figure 3.
All symbol meanings used in JTM graph model shown in Fig. 3 are as shown in table 3.JTM is the life for having supervision
At model, one or more effects of document d are regarded as the K theme label of document d, using vector Λd=(t1,t2,...,tK)
Effect t in document d is indicated for a series of binary numberiWhether occur, tiThere is label l in ∈ { 0,1 }, 0 expression d, 1 indicates d
In do not show l.ΛdObedience priori is that the Bernoulli Jacob of Beta (γ) is distributed；ΛdThe theme distribution θ of dominant restraint document dd。θdFor text
The effect of shelves d, (theme) ratio was multinomial distribution, and priori is obeyed Cray in Di that parameter is α and is distributed, it may be assumed that θd~
Dirichlet(α)；zdnBy multinomial distribution θdIt generates；N-th of h of document ddn(Chinese medicine), with it for gdn(four gas), fdn(five
Taste), tdn(channel tropism) shares effect zdn；It notices in Chinese medicine data, four gas and five tastes majority of case of single medicinal material are only
One, but some Chinese medicines have more than a kind of four gas or the five tastes.WithIt is multinomial distribution, obeying priori respectively is
βh、βfAnd βtDirichlet distribution, respectively indicate Chinese medicine under theme, nature and flavor and channel tropism distribution.
The generating probability of one first Chinese medicine compound prescription document d can be expressed as follows, and the generating process of data set is as shown in table 4:
3 JTM graph model of table indicates symbol meaning
The generating process of 4 JTM model data collection of table
S32JTM model parameter solves: in JTM model, hidden variableIt is similar to LDA model with z effect,
It is all the multinomial distribution that priori is Dirichlet distribution；And document effect (theme) is distributed θdBy dominant set ΛdConstraint, because
This, can be used the function of the Chinese medicine h of position i in the Collapsed Gibbs Sampling method approximate solution document d of similar LDA
Effect.
H, f and t share assigned same effect (theme) number in the same document.It is therefore not necessary to individually sample
Theme.In formula (3.2), i indicates that current location Chinese medicine ,-i indicate to remove current location i example,Indicate that document d removing is worked as
After the Chinese medicine of front position i, it is assigned the number of effect k；It indicates after removing current location i, Chinese medicine h is assigned time of effect k
Number；αkΛd,kAccording to dominant restraint Λd,kGuarantee that effect k whether there is Λ in document dd,k={ 0,1 }.
Similar to LDA, following parameter can finally be acquired according to multinomial distribution and Dirichlet conjugate property:
Formula (3.2) into formula (3.6), | H |, | F | and | T | indicate all Chinese medicines, nature and flavor and the channel tropism that occur in data set
Sum.
S4 trains JTM model；
Input: the compound message data set of expansion pharmacological property and channel tropism information as described in step 2；JTM model hyper parameter α, β
And training the number of iterations Inter.
Training process:
1) initialize: ergodic data concentrates every Chinese medicine h of each compound ddn, from the Λ of the compounddSet in
Machine distributes an effect and numbers to hdn, while hdnCorresponding all nature and flavor fdnWith channel tropism tdnShare effect number；
2) every Chinese medicine h of each compound d in data set is iterated overdnUntil the number of iterations Inter of setting:
To each hdnBelong to Λ according to what formula (3.2) recalculated itdEach effect probability and carry out corresponding
Gibbs sampler.
Output: effect number two-dimensional matrix z corresponding to every Chinese medicine in the data set in all compounds.
S5 exports effect according to training result and corresponds to Chinese medicine and pharmacological property.It can be according to public affairs according to the resulting matrix z of training result
Formula (3.4) statistics calculates the pro-bability value matrices of the powerful and Chinese medicine of gained institute
Wherein, NhThe unduplicated Chinese medicine number occurred for compound message data set.
Further, comprising:
According to the resulting matrix z of training result can according to formula (3.5) count calculate gained it is powerful general with nature and flavor
Rate value matrix
Wherein, NfThe unduplicated Chinese medicine nature and flavor number occurred for compound message data set.
Further, comprising:
According to the resulting matrix z of training result can according to formula (3.6) count calculate gained it is powerful general with channel tropism
Rate value matrix
Wherein, NtFor the number for the unduplicated meridian tropism that compound message data set occurs.
In order to illustrate the validity of the method for the invention, the present invention has done following experiment:
Experimental data be collected in " countries population and health science data sharing platform " Chinese medicine database (http: //
) and Chinese medicine special subject network station " Chinese medicine e hundred " (http://www.tcm100.com/ dbcenter.cintcm.com/
default.aspx).Data source and explanation are as shown in table 5.
Above data is mainly two class of Chinese herbal medicine and compound.Former data include that data information is more, and each data source is not united
One.Chinese herbal medicine data include after needing to screen according to this experiment: medicine name, alias, nature and flavor and channel tropism；Compound information include effect and
Drug forms two category informations, and the two example difference is as shown in tables 6 and 7.
5 data origin information of table
Examples of information after 6 Chinese medicine screening of table
Data instance after the screening of 7 compound of table
As exemplified in table 6, there are a large amount of alias for thousands of taste Chinese medicines, in 8123 kinds of Chinese Traditional Medicine informations of this experiment crawl,
Shared alias 21890, every kind of Chinese medicine has averagely reached 2.7 alias.Such as 7 example of table, compound medicine composition is write in the presence of more
The disunity situation of kind form, some prescriptions lack drug dose, some dosage are due to historical dynasty difference, unit disunity.
In order to examine the validity of JTM model.
In this experiment, following pretreatment has been carried out to above-mentioned compound message data set:
Remove the compound of missing efficacy information；
Dedicated vocabulary is constructed by 8123 kinds of Chinese medicines, branch is constituted to compound medicine and is segmented；
Lay down a regulation the various dosage informations removed in compound medicine composition, such as: 1 liang, 3-10g etc., and indicate processing side
The word of method, such as: raw, soup bubble, it is vinegar-fried etc.；
Unified replacement is carried out to all Chinese medicines drug of the same name in compound；
In order to guarantee enough co-occurrence informations, it is all less than 5 to remove effect frequency of occurrence in Chinese medicine compound prescription data set
Compound data；
In order to guarantee the multinomial distribution in model, removal efficiency is less than 2 or Chinese medicine constitutes the compound less than 3 kinds.
3488 compound data are obtained after carrying out the above pretreatment and screening.Compound statistical information about data set is such as
Shown in table 8.The channel tropism and nature and flavor statistical information of pretreatment 8123 kinds of Chinese medicine after reunification are as shown in table 9.
In addition, the nature and flavor and channel tropism to 8123 kinds of Chinese Traditional Medicine informations carry out following pretreatment or screening measure:
In order to which vocabulary is stated unified, property of traditional Chinese taste description information is filtered and is replaced, only retained sour, bitter, sweet, pungent, salty
With it is cold, hot, warm, cool and its add micro-, the large and small composition of qualifier nature and flavor, such as: acrid flavour；Slightly warm in nature is pungent after replacement；Tepor.
In order to which vocabulary is stated unified, meridian tropism is described, only retains the directly statement word such as vital organs of the human body, such as: large intestine channel replaces
It is changed to large intestine.
8 compound data set basic information of table statistics
The channel tropism and nature and flavor statistical information of 98 123 kinds of Chinese medicines of table
Experimental setup
JTM models coupling effect, Chinese medicine, nature and flavor and corresponding channel tropism, are the theme, to three semantic dimensions with dominant effect label
Degree combines modeling.In order to investigate JTM modelling technique performance, 90% will be randomly selected from experimental data and is used to train, remaining conduct
Test data.In experiment, model parameter α=0.01, β=K/50；K=749 is unique effect number in data set；
Inter=1000 is arranged in the number of iterations.
In addition, in an experiment due to including the degree adverbs such as micro-, light, small, the property modified by these adverbial words in Chinese medicine nature and flavor
Taste belongs to nature and flavor of the same race, such as: slight bitter and the bitter hardship belonged in five tastes of medicinal herb；Toxic and big poison all expresses the toxicity profile of medicine.
In order to embody their semantic general character and individual character, by the word comprising the degree adverbs modification such as micro-, light, small in pharmacological property taste in this experiment
It is denoted as 1 on the basis of the co-occurrence frequency, is upgraded to 3 times comprising the big frequency, other words are 2 times.
Experimental result and analysis；Table 10 enumerates 5 kinds of highest preceding 10 tastes of effect (topics, theme) probability in experimental result
Chinese medicine, preceding 3 nature and flavor and preceding 3 channel tropisms.
The probability distribution of five kinds of theme Chinese medicines of 10 JTM model training result of table, nature and flavor and channel tropism
Topic1: activating microcirculation and removing stasis medicinal.It is given in activating microcirculation and removing stasis medicinal classics in the works errors in Medicine Corrected of Qing Dynasty's name doctor's Wang Qingren
Medicine prescription is constituted are as follows: each three money of Radix Angelicae Sinensis, Radix Rehmanniae, four money of peach kernel, three money of safflower, each two money of Fructus Aurantii, RADIX PAEONIAE RUBRA, one money of radix bupleuri are sweet
Careless two money, one money of campanulaceae half, one money of Rhizoma Chuanxiong half, three money of radix achyranthis bidentatae；The party amounts to 11 taste medicines；And it is enumerated in JTM model learning result
Preceding 10 taste medicine in, covered 6 herbal medicines of the classics recipe；Radix Angelicae Sinensis, Rhizoma Chuanxiong, safflower, radix paeoniae rubra, peach kernel, Radix Glycyrrhizae；And it does not cover
The Radix Salviae Miltiorrhizae that covers, Radix Notoginseng, Caulis Spatholobi are the common drug of activating microcirculation and removing stasis medicinal.Activating microcirculation and removing stasis medicinal is mainly acted on promoting blood circulation, scattered silt
It is main, based on pungent and bitter, return liver and the heart channel of Hang-Shaoyin；JTM model training is the results show that preceding 2 nature and flavor are bitter, pungent；Front three channel tropism is
Liver, spleen and the heart.Experimental result reflect JTM model learning to theme activating blood and removing stasis drug composition meet traditional Chinese medical science prescription pharmacology
By.
Topic2: swelling and pain relieving.Chinese medicine traditional recipe detumescence ointment composition are as follows: Radix Aconiti 50g, Radix Aconiti Kusnezoffii 50g, olibanum 25g,
Myrrh 25g, peach kernel 90g, rheum officinale 100g, root of Dahurain angelica 75g, airpotato yam 75g, centipede 20g, scorpio 20g, ramulus cinnamomi 50g, Radix Angelicae Sinensis 50g, mountain
Naphthalene 180g, camphor 500g, Moschus (can use muskone generation), borneol are a little.10 taste pack before the swelling and pain relieving that JTM model learning arrives
7 taste medicines in the classics recipe contained: olibanum, borneol, rheum officinale, radix saposhnikoviae, the root of Dahurain angelica, myrrh, Moschus, swelling and pain relieving and activating microcirculation and removing stasis medicinal
There is similar property on nature and flavor and channel tropism, and first three of the result Topic1 of JTM study and Topic2 meridian distribution of property and flavor are homogeneous
Together.JTM has further been confirmed to pharmacological property channel tropism rule learning outcome reliability with higher.
Topic3: clearing heat and promoting diuresis.In the result of JTM study, front three channel tropism is liver, the heart, spleen, and nature and flavor are bitter, cold, sweet.
Syndrome of dampness-heat is broadly divided into liver and gallbladder and damp heat in the spleen and the stomach in Chinese medicine, damp and hot and show as dryness-heat, therefore JTM obtains clearing heat and promoting diuresis prescription
Whole channel tropism is that liver, the heart, spleen are more reasonable；According to theory of traditional Chinese medical science, bitter energy clearing damp, energy heat-clearing of trembling with fear, therefore JTM is to compound entirety
The nature and flavor front two provided is exactly bitter and cold, meets theory of traditional Chinese medical science reality.Has effects that the common classical Chinese patent drug of removing damp-heat
Having Longdan Xiegan wan, (prescription: rough gentian, radix bupleuri, radix scutellariae, cape jasmine (stir-fry), rhizoma alismatis, caulis akebiae, stir-baked SEMEN PLANTAGINIS with salt solution, prepared RADIX ANGELICAE SINENSIS with yellow rice wine, glutinous rehmannia, toast are sweet
Grass), ingredient includes that JTM learns four Chinese medicines (radix bupleuri, radix scutellariae, cape jasmine, Radix Glycyrrhizae) into clearing heat and promoting diuresis；And other Chinese medicines,
Such as: oriental wormwood, the coptis, Cortex Phellodendri clearing heat and promoting diuresis, rheum officinale heat-clearing, Poria cocos clearing damp calming heart.Result above reflects basic with theory of traditional Chinese medical science
It coincide.
Topic4: clearing heat and detoxicating.Common clearing heat and detoxicating Chinese patent drug such as detoxicating tablet of cow-bezoar (main component: cow-bezoar, realgar, stone
Cream, rheum officinale, radix scutellariae, campanulaceae, borneol, Radix Glycyrrhizae), honeysuckle particle (main component: honeysuckle, Fructus Forsythiae, radix scutellariae) etc..JTM is obtained
The Chinese medicine of clearing heat and detoxicating theme has covered the four traditional Chinese medicine object in detoxicating tablet of cow-bezoar: cow-bezoar, rheum officinale, Radix Astragali and Radix Glycyrrhizae.Compound Jin
The four traditional Chinese medicine object of honeysuckle flower particle main component all covers.The compatibility of drugs for coming from " Liu Fengwu Experience In Gynecology " Qingre Jiedu Tang is
1: 15 grams of Fructus Forsythiae, 15 grams of honeysuckle, 15 grams of dandelion, 15 grams of viola mandshurica, 9 grams of radix scutellariae, 9 grams of semen plantaginis, 9 grams of the root bark of tree peony, ground bone
9 grams of skin, 12 grams of dianthus superbus, plaque stores 12 grams, and 6 grams of radix paeoniae rubra, 30 grams of waxgourd seed, effect is clearing heat and detoxicating, dampness removing cool blood, swelling and pain relieving.
Fructus Forsythiae, honeysuckle, dandelion, radix scutellariae are in JTM model learning into clearing heat and detoxicating result in party's monarch drug in a prescription.For another example " typhoid fever
By " in heat-clearing class prescription gegen qinlian decoction constitute are as follows: pueraria lobata, radix glycyrrhizae preparata, radix scutellariae, the coptis are similar to the result that JTM is obtained.
For clearing heat and detoxicating theme, JTM learns to have arrived the whole nature and flavor of the prescription are as follows: bitter, cold, sweet；Channel tropism are as follows: liver, the heart
And spleen.It is cold with the medicine of a warm nature that treatment is spoken of in Shennong's Herbal, treats heat with medicine of trembling with fear；It is talked about in " being discussed when the hiding Plain Questions gas hair ": pungent to dissipate, is sour
Receipts, Gan Huan, hardship are hard, salty soft.Meet it can be seen that the clearing heat and detoxicating theme that JTM learns corresponds to drug matching, channel tropism and nature and flavor
Theory of traditional Chinese medical science.
Topic5: supplementing qi and nourishing yin.Supplementing qi and nourishing yin cures mainly the main attack supplementing qi and nourishing yin given for syndrome of deficiency of both qi and yin, name doctor Gu Zhendong,
Prescription are as follows: 30 grams of Radix Astragali, 20 grams of radix pseudostellariae, 15 grams of rhizoma polygonati, 12 grams of Rhizoma Atractylodis Macrocephalae, 10 grams of Poria cocos, 20 grams of Radix Rehmanniae, 20 grams of Radix Ophiopogonis, asparagus fern
15 grams, 18 grams of Eclipta prostrata, 15 grams of the fruit of glossy privet, 30 grams of oldenlandia diffusa, 30 grams of Sculellaria barbata, 30 grams of dandelion, 15 grams of field thistle, Radix Glycyrrhizae
5 grams learn to being overlapped in ten taste medicines before supplementing qi and nourishing yin to be Radix Astragali, rhizoma polygonati, Rhizoma Atractylodis Macrocephalae, Poria cocos, Radix Glycyrrhizae with JTM.In addition, JTM learns
Result in include common spleen-Qi nourishing Chinese herbs Radix Astragali, Chinese yam, Radix Glycyrrhizae；Schisandra chinensis astringency inducing, nourishing generate fluid, kidney tonifying are peaceful
The heart；Radix scrophulariae enriching yin；Pueraria lobata promotes the production of body fluid to quench thirst, and cures mainly the deficiency of Yin and quenches one's thirst.The main tonifying spleen lung kidney of supplementing qi and nourishing yin, calming heart.
Therefore the channel tropism result reasonable that JTM is obtained.And it is sweet, flat, bitter can boosting qi and nourishing yin, nature and flavor collocation meets Chinese medicine
It is theoretical.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Finally it should be noted that: the above embodiments are merely illustrative of the technical scheme of the present invention and are not intended to be limiting thereof, to the greatest extent
Invention is explained in detail referring to above-described embodiment for pipe, it should be understood by those ordinary skilled in the art that: still
It can be with modifications or equivalent substitutions are made to specific embodiments of the invention, and without departing from any of spirit and scope of the invention
Modification or equivalent replacement, should all cover within the scope of the claims of the present invention.
Claims (10)
1. a kind of method analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship, which is characterized in that the party
Method includes:
(1) pending data collection is obtained, the data set includes Chinese Traditional Medicine information data set and compound message data set, and to described
Chinese Traditional Medicine information data set and compound message data set are pre-processed respectively；
(2) nature and flavor and channel tropism are carried out to pretreated compound message data set using pretreated Chinese Traditional Medicine information data set
Expand；
(3) multi-dimensional relation between compound information data concentration Chinese medicine, effect and nature and flavor after expanding is taken out, building combines master
Model is inscribed, and solves the model parameter of the combination topic model；
(4) the compound message data set after extraction section expands instructs the combination topic model as training set
Practice, until reaching the number of iterations of setting, and exports the probability value square of effect Yu Chinese medicine, nature and flavor and channel tropism according to training result
Battle array.
2. the method according to claim 1 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, each compound information data that the compound information data is concentrated includes effect and drug composition, the Chinese medicine
Each Chinese Traditional Medicine information data that information data is concentrated include medicine name, alias, nature and flavor and channel tropism, and the nature and flavor include four gas and five
Taste, four gas include: it is hot, warm, cold, cool, the five tastes include pungent, sweet, sour, bitter, salty.
3. the method according to claim 2 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, being pre-processed in the step (1) to Chinese medicine message data set, comprising:
(11) the nature and flavor description information in Chinese Traditional Medicine information is filtered and is replaced, only retain the five tastes and four gas and its add qualifier
The nature and flavor of micro-, big, light, small composition；
(12) only retain the description of channel tropism in Chinese Traditional Medicine information the direct statement word of the vital organs of the human body.
4. the method according to claim 3 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, described pre-process Chinese medicine message data set further include: will include micro-, light, small degree adverb in nature and flavor
On the basis of the Term co-occurrence frequency of modification, it is denoted as 1；3 times of former word frequency are upgraded to comprising degree adverb big word frequency time；There is no degree pair
The word frequency of the nature and flavor of word modification is upgraded to 2 times of former word frequency.
5. the method according to claim 1 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, in the step (3), the compound information data after taking out the expansion concentrate Chinese medicine, effect and nature and flavor it
Between multi-dimensional relation, specifically include:
It is one-to-one recessive relationship between effect and medicine group；It is one-to-many recessive relationship between medicine group and Chinese medicine；Chinese medicine and property
There are the dominance relations of multi-to-multi between taste；There are the recessive relationships of indirect multi-to-multi between medicine group and nature and flavor；Effect and nature and flavor
Between there are indirect multi-to-multi recessiveness relationships.
6. the method according to claim 5 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, building combines topic model in the step (3), comprising:
Compound message data set after expansion regards D independent text generation processes as, and single compound information data is as one
Total bag of words of document d, JTM model include three sub- bag of words, are respectively as follows: Chinese medicine bag of words, concentrate and occur comprising compound information data
All Chinese medicines；Chinese medicine nature and flavor bag of words, including four gas, the five tastes and toxicity profile；Channel tropism bag of words refer to Chinese medicine according to zang-fu differentiation
Theory action is in the statement word of human body；
One or more effects of the document d are regarded as the K theme label of document d, and K indicates effect type；Using vector Λd
=(t1,t2,...,tK) indicate effect t in a series of binary number instruction document diWhether occur, ti∈ { 0,1 }, 0 indicates d
In there is label l, do not occur l in 1 expression d；ΛdObedience priori is that the Bernoulli Jacob of Beta (γ) is distributed；ΛdDominant restraint document d
Theme distribution θd；θdThe effect of for document d ratio, be multinomial distribution, priori is obeyed Cray in Di that parameter is α and is distributed,
That is: θd~Dirichlet (α)；zdnBy multinomial distribution θdIt generates, wherein zdn∈{0,…K-1}；N-th of Chinese medicine h of document ddn,
With it for four gas gdn, five tastes fdn, channel tropism tdn, share effect zdn；WithIt is multinomial distribution, obeys priori respectively
For βh、βfAnd βtDirichlet distribution, respectively indicate Chinese medicine under theme, nature and flavor and channel tropism distribution.
7. the method according to claim 6 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, the generating probability formula of the document d indicates are as follows:
Wherein, NfIndicate the nature and flavor number of the n-th taste Chinese medicine in document d, NtIndicate the channel tropism number of the n-th taste Chinese medicine in document d, Nh
Indicate the sum of Chinese medicine in document d.
8. the method according to claim 7 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, solving the model parameter of the combination topic model in the step (3), comprising: a certain type the effect of
Under k, each parameter is expressed as:
Wherein, k ∈ { 1 ..., K }, | H |, | F | and | T | indicate all Chinese medicines occurred in data set, nature and flavor and channel tropism sum,
It indicates after removing current location i, Chinese medicine h is assigned the number of effect k, and i indicates that current location Chinese medicine ,-i indicate to remove present bit
I is set,Expression is gone out after the i of current location, and nature and flavor f is assigned the number of effect k,Expression goes out to return after the i of current location
The number of effect k is assigned through t.
9. the method according to claim 8 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, the compound message data set cooperation after extraction section expands is training set, to institute in the step (4)
It states and is trained in conjunction with topic model, comprising:
Input: compound information data set, JTM model hyper parameter α, β and training the number of iterations after expansion；
Step: (1) initialize: ergodic data concentrates every Chinese medicine h of each document ddn, from the Λ of the compounddSet in
An effect is randomly assigned to number to hdn, while hdnCorresponding all nature and flavor fdnWith channel tropism tdnShare effect number；
(2) every Chinese medicine h of each compound d in data set is iterated overdnUntil the number of iterations of setting: to each hdnIt calculates
It belongs to ΛdEach effect probability and carry out corresponding gibbs sampler.
Output: effect number two-dimensional matrix z corresponding to every Chinese medicine in the training set in all compounds.
10. the method according to claim 9 analyzed effect from Chinese medicine compound prescription and correspond to Chinese medicine and its pharmacological property compatibility relationship,
It is characterized in that, the pro-bability value matrices for exporting effect and Chinese medicine, nature and flavor and channel tropism according to training result, comprising:
Calculate that gained institute is powerful and the pro-bability value matrices of Chinese medicine according to the resulting matrix z of training result and the formula 3
Wherein, NhThe unduplicated Chinese medicine number occurred for compound message data set.
Calculate that gained institute is powerful and the pro-bability value matrices of nature and flavor according to the resulting matrix z of training result and formula 4
Wherein, NfThe unduplicated Chinese medicine nature and flavor number occurred for compound message data set.
Calculate that gained institute is powerful and the pro-bability value matrices of channel tropism according to the resulting matrix z of training result and formula 5
Wherein, NtFor the number for the unduplicated meridian tropism that compound message data set occurs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910576633.XA CN110289106B (en) | 2019-06-28 | 2019-06-28 | Method for analyzing compatibility relation of traditional Chinese medicines with corresponding efficacy and drug properties in traditional Chinese medicine compound |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910576633.XA CN110289106B (en) | 2019-06-28 | 2019-06-28 | Method for analyzing compatibility relation of traditional Chinese medicines with corresponding efficacy and drug properties in traditional Chinese medicine compound |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110289106A true CN110289106A (en) | 2019-09-27 |
| CN110289106B CN110289106B (en) | 2023-06-16 |
Family
ID=68019557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910576633.XA Active CN110289106B (en) | 2019-06-28 | 2019-06-28 | Method for analyzing compatibility relation of traditional Chinese medicines with corresponding efficacy and drug properties in traditional Chinese medicine compound |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110289106B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110970115A (en) * | 2019-10-28 | 2020-04-07 | 广西科技大学 | Informatization representation method for nature, taste and meridian tropism of traditional Chinese medicine prescription |
| CN111383740A (en) * | 2020-03-03 | 2020-07-07 | 杭州师范大学 | Traditional Chinese medicine efficacy prediction method based on multitask deep neural network |
| CN111681717A (en) * | 2020-05-28 | 2020-09-18 | 北京银谷芳香科技有限公司 | Artificial intelligence-based traditional Chinese medicine quality control and natural substance efficacy prediction model and construction and use methods thereof |
| CN112133382A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Learning method and system for medicine analysis by using algorithm model |
Citations (1)
| 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 |
-
2019
- 2019-06-28 CN CN201910576633.XA patent/CN110289106B/en active Active
Patent Citations (1)
| 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 |
Non-Patent Citations (3)
| Title |
|---|
| 李味味等: "基于BP神经网络中药复方功效的预测研究", 《中医药导报》 * |
| 胡建军: "中药复方配伍多维分析系统中数据库的设计", 《计算机工程与设计》 * |
| 胡雅凌等: "基于数据挖掘对小柴胡汤类方性味归经功效属性关系的发现研究", 《中国实验方剂学杂志》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110970115A (en) * | 2019-10-28 | 2020-04-07 | 广西科技大学 | Informatization representation method for nature, taste and meridian tropism of traditional Chinese medicine prescription |
| CN110970115B (en) * | 2019-10-28 | 2023-03-21 | 广西科技大学 | Informatization representation method for nature, taste and meridian tropism of traditional Chinese medicine prescription |
| CN111383740A (en) * | 2020-03-03 | 2020-07-07 | 杭州师范大学 | Traditional Chinese medicine efficacy prediction method based on multitask deep neural network |
| CN111681717A (en) * | 2020-05-28 | 2020-09-18 | 北京银谷芳香科技有限公司 | Artificial intelligence-based traditional Chinese medicine quality control and natural substance efficacy prediction model and construction and use methods thereof |
| CN111681717B (en) * | 2020-05-28 | 2023-09-15 | 北京银谷芳香科技有限公司 | Traditional Chinese medicine quality control and natural matter efficacy prediction model based on artificial intelligence and construction and use methods thereof |
| CN112133382A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Learning method and system for medicine analysis by using algorithm model |
| CN112133382B (en) * | 2020-09-24 | 2024-02-20 | 南京泛泰数字科技研究院有限公司 | Learning method and system for medical analysis by using algorithm model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110289106B (en) | 2023-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110289106A (en) | A method of effect, which is analyzed, from Chinese medicine compound prescription corresponds to Chinese medicine and its pharmacological property compatibility relationship | |
| CN106803012B (en) | Prescription function prediction method based on probability topic model and Chinese medicine essential attribute | |
| Lee et al. | Systematic review on herb-induced liver injury in Korea | |
| Yang et al. | Chinese herbal medicines for induction of remission in advanced or late gastric cancer | |
| Yi et al. | An overview of traditional Chinese herbal formulae and a proposal of a new code system for expressing the formula titles | |
| CN112992378B (en) | Multi-dimensional traditional Chinese medicine prescription recommendation method and recommendation system | |
| He et al. | Does the couse of astragalus‐containing chinese herbal prescriptions and radiotherapy benefit to non‐small‐cell lung cancer treatment: a meta‐analysis of randomized trials | |
| Yang et al. | Research on the medication rules of Chinese herbal formulas on treatment of threatened abortion | |
| CN108014303A (en) | A kind of Chinese medicine preparation for treating Ru Repair-in and preparation method thereof | |
| CN108079260A (en) | A kind of Chinese medicine preparation for treating the proliferation of mammary gland and preparation method thereof | |
| CN105749226A (en) | Traditional Chinese medicine composition for treating rheumatism, rheumatoid and early osteoma | |
| Hijikata | Analgesic treatment with Kampo prescription | |
| CN105362656A (en) | Traditional Chinese medicine decoction capable of relaxing bowels | |
| CN105434864A (en) | Traditional Chinese medicine composition for treating malignant tumor | |
| Li et al. | An analysis and research of type-2 diabetes TCM records based on text mining | |
| Qiao et al. | Analysis of Clinical Medication Rule of Trigonellae Semen Based on Data Mining | |
| Wan-li et al. | Study on the medication rule of Wang Xun in the treatment of dysentery based on Data Mining. | |
| Yin et al. | Discussion on the relationship between symptoms and drugs in the treatment of fever with modified Zhuye Shigao Decoction based on data mining | |
| Gao et al. | Analysis of Professor Lei Zhang’s Medical Records of Banxia XieXin Decoction Analogous Prescriptions Based on Principle of attribute partial ordering | |
| Zheng et al. | Data Mining-Based Analysis of Chinese Medicinal Herb Formulae in Metabolic associated fatty liver Disease Treatment | |
| CN116340480A (en) | Traditional Chinese medicine dialectical system based on document theme generation model and storage medium | |
| CN104208445A (en) | Traditional Chinese medicine composition for treating dizziness | |
| Cheng et al. | A Support Vector Machine Learning for the Upward and Downward Tendency Theory of Traditional Chinese Medicine | |
| Dou et al. | Exploring medication rules in Chinese randomized controlled trials of type 2 diabetes based on data mining technology | |
| CN105853868A (en) | Medicine for prolonging life |
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 | ||
| CP03 | Change of name, title or address |
Address after: 223005 A12-2, high tech Industrial Park, three East seven street, Hongze District, Huaian, Jiangsu (Hongze technology transfer center Hongze sub center) Patentee after: HUAIYIN INSTITUTE OF TECHNOLOGY Country or region after: China Patentee after: THE SECOND PEOPLE'S HOSPITAL OF HUAI'AN Address before: 223005 A12-2, high tech Industrial Park, three East seven street, Hongze District, Huaian, Jiangsu (Hongze technology transfer center Hongze sub center) Patentee before: HUAIYIN INSTITUTE OF TECHNOLOGY Country or region before: China Patentee before: HUAI'AN SECOND PEOPLES HOSPITAL (HUAI'AN RENCI HOSPITAL) |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20240924 Address after: 150000 Building 1, Unit 2502, Mingyue Romantic City Community, Harmony Road, Songbei District, Harbin City, Heilongjiang Province Patentee after: Harbin Huilan Pharmaceutical Technology Co.,Ltd. Country or region after: China Address before: 223005 A12-2, high tech Industrial Park, three East seven street, Hongze District, Huaian, Jiangsu (Hongze technology transfer center Hongze sub center) Patentee before: HUAIYIN INSTITUTE OF TECHNOLOGY Country or region before: China Patentee before: THE SECOND PEOPLE'S HOSPITAL OF HUAI'AN |