CN108717862A - A kind of careful square evolution model of the intelligence based on machine learning - Google Patents
A kind of careful square evolution model of the intelligence based on machine learning Download PDFInfo
- Publication number
- CN108717862A CN108717862A CN201810317800.4A CN201810317800A CN108717862A CN 108717862 A CN108717862 A CN 108717862A CN 201810317800 A CN201810317800 A CN 201810317800A CN 108717862 A CN108717862 A CN 108717862A
- Authority
- CN
- China
- Prior art keywords
- drug
- tentative diagnosis
- ingredient
- probability
- grouping
- 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
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
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
-
- 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)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medicinal Chemistry (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Pharmacology & Pharmacy (AREA)
- Toxicology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a kind of intelligence based on machine learning to examine square evolution model, belong to the technical field of machine learning model, model is solved by sorting algorithm, and history prescription data and careful number formulary are obtained by the probability matrix of model according to study based on bayesian algorithm, intelligence, which is established, eventually by the probability matrix of drug and tentative diagnosis examines square model, by establishing the medication correlation analysis under same tentative diagnosis, establish drug relevance model, solve the correlation analysis between tentative diagnosis and drug, under the premise of tentative diagnosis is formulated in analysis, whether the drug of recommendation is to meet the tentative diagnosis;By the correlation analysis between tentative diagnosis and the drug ingredient of drug, to drug ingredient data normalization;After the standardization of user's symptom data and drug are applicable in disease data normalization, the relevance of analysis user's symptom and corresponding drug ingredient.
Description
Technical field
The invention belongs to the technical fields of machine learning model, and in particular to a kind of intelligence based on machine learning is careful just to be opened
Square model.
Background technology
With the extensive use of existing internet remote interrogation, patient's interrogation flow is substantially reduced to:Patient comes PC machine
It connects screen and the state of an illness is described by voice, doctor knows the symptom of the state of an illness and the severity of its state of an illness, and doctor is according to medicine
Relevant knowledge assigns a cause for an illness to obtain tentative diagnosis to know disease type, under then doctor suits the medicine to the illness according to symptom and disease type
Medicine selects corresponding related drug to form prescription from drug data bank.Then this prescription is issued pharmacist's progress prescription by system and is examined
Core, pharmacist complete Checking prescription according to the medication collocation of prescription, patient information description, tentative diagnosis, drug dose.
Under normal circumstances, user's referring physician describes the related datas such as the state of an illness (i.e. symptom) and allergies.Then doctor
Obtain tentative diagnosis, according to the symptom prescription of tentative diagnosis and user, and prescription generally comprises one or more drugs, and this
A little drugs all have attribute:Drug ingredient is applicable in symptom, drug taboo.Drug ingredient between different drugs needs to do ingredient
Conflict analysis, and the applicable symptom between different drug corresponds to user's symptom whether can do the whether applicable user's symptom of drug i.e. right
The analysis of disease prescribe medicine, and allergies and heredity medication history of the taboo of the drug of drug and user etc. may make up the verification of drug safety.
Due to product development need, a set of intelligence based on machine learning need to be established and examine square evolution system, for assisting doctor
Crude drug Shi Jinhang is quickly, accurately and efficiently diagnosing patient and service.
Invention content
In order to solve the above problem of the existing technology, present invention aims at provide a kind of intelligence based on machine learning
Square evolution model can be examined to reach for assisting doctor pharmacist to carry out mesh that is quick, being accurately and efficiently diagnosing patient and service
's.
The technical solution adopted in the present invention is:A kind of careful square evolution model of the intelligence based on machine learning, packet are provided
Include the following contents:
(1) it is A by all the components list of drug1, A2, A3....Aw, amount to w kind ingredients, for selected drug i,
It includes ingredient lists be { x1,x2.....,xk, xk∈Aw(k≤w);
(2) the correlation analysis model based on sorting algorithm, the relevance size are F (yj), using such as minor function public affairs
Formula:
Wherein, wiIndicate the cost coefficient and w of drug ii∈ [0,1], PijIndicate that drug i belongs to tentative diagnosis grouping j's
Probability size and Pij∈ [0,1];Then PkjIndicate drug ingredient xkThe probability size for belonging to tentative diagnosis grouping j, then obtain following
Function formula:
Wherein, PqjIndicate that drug ingredient q belongs to the probability size of tentative diagnosis grouping j;
Comprehensive (2-1) and (2-2), then obtain following function formula:
Its object function is:
The tentative diagnosis grouping j belongs to overall m kind tentative diagnosis grouping, then (1,2,3.......m) j ∈, each
The probability size that drug i belongs to each tentative diagnosis grouping j is Pij, then PijFor n × m dimension array, i.e.,:
Wherein (i≤n, j≤m); (2-5)
(3) it is by the process simplification of careful side:The relevance size of the drug i of selection and the tentative diagnosis grouping j of the disease
For F (yj), set its overall association size be as ε and ε (0,1] constant, as relevance size F (yj) be less than ε when indicate choosing
The drug i taken the and tentative diagnosis relationship grouping j of the disease is strong relationship, then prescription is to examining;As relevance size F (yj) be more than
Indicate that the tentative diagnosis relationship grouping j for the drug i and disease chosen is weak relationship when ε, then prescription is not to examining.
Further, the probability matrix PijCalculating steps are as follows:
(1) data initialization, for giving one group of tentative diagnosis Cj, split tentative diagnosis;
(2) prescription E is chosen in tentative diagnosisj, prescription EjIn include drug X1, X2......Xk;
(3) for drug XiAnd i≤k, and extract drug ingredient vector { x1,x2.....,xn, it is assumed that each drug ingredient it
Between independently of each other;If the drug is abandoned less than drug ingredient, is considered as invalid drug by extraction;
(4) drug ingredient is recorded to tentative diagnosis CjIn, if having there is the ingredient, go out occurrence in the record of the ingredient
Number plus 1, meanwhile, corresponding drug is recorded in tentative diagnosis CjIn occurrence number;
(5) record tentative diagnosis CjOccurrence number, total sample number adds 1;
(6) repeat step (2)-(5) until in prescription the study of all drugs finish to obtain tentative diagnosis list;
(7) tentative diagnosis list is traversed, drug ingredient is for tentative diagnosis C in calculating tentative diagnosis listjPosteriority it is general
Rate:
Wherein, xiFor drug ingredient and xi∈{x1,x2.....,xn};D is sample space;
(8) step (7) is repeated, until tentative diagnosis CjIn the calculating of all drug ingredients finish;
(9) all drugs in tentative diagnosis list are traversed;
(10) drug X is calculatediTo tentative diagnosis CjProbability:
With drug ingredient < x1,x2...xn> indicates drug Xi, due to drug < x1,x2...xnProbability p (the x of >1,
x2...xn) it is constant, normalized is done to it, according to Bayesian assumption, is unfolded to obtain using Bayesian formula:
(11) step (9)-(10) are repeated, until tentative diagnosis CjIn the probability calculations of all drugs finish;
(12) step (7)-(11) are repeated, until all drugs finish all tentative diagnosis efficiency calculations, obtain probability
Matrix Pij。
Further, the cost coefficient w in the formula (2-1)iThe preliminary applicability ω for being equal to drugi, wherein i tables
Show i-th kind of drug, according to formula:
Wherein, ωi=C 'i,C′i∈ (0,1), C 'iIndicate the applicability and C ' of drug iiHow many kinds of is appeared in equal to drug i
In different tentative diagnosis;And C 'i=[c1,c2,c3....,ci,....,cn], wherein ciIndicate that i-th kind of drug appears in how many kinds of
Different tentative diagnosis, wherein ci>=0, ciIt can learn to obtain by historical data.
Beneficial effects of the present invention are:
1. examine square evolution model for the intelligence provided by the invention based on machine learning, can by interrogation process simplification,
Mathematical model is extracted, corresponding model is established, model is solved by sorting algorithm;It will be gone through by being based on bayesian algorithm
History prescription data and careful number formulary obtain the probability matrix of model according to study, eventually by the probability matrix P of drug and tentative diagnosisij
It establishes the careful square model of intelligence and establishes drug relevance mould by establishing the medication correlation analysis under same tentative diagnosis
Type has achieved the purpose that intelligent careful side is open.
Description of the drawings
Fig. 1 is that the intelligence provided by the invention based on machine learning examines relevance size F (y in square evolution modelj) it is several
What meaning representation schematic diagram;
Fig. 2 is that the intelligence provided by the invention based on machine learning examines each drug ingredient A in square evolution model1,A2....An
With the relation schematic diagram between tentative diagnosis C.
Specific implementation mode
Below in conjunction with the accompanying drawings and specific embodiment the present invention is further elaborated.
The present invention provides a kind of intelligence based on machine learning to examine square evolution model comprising the following contents:
(1) drug ingredient is an attribute variable for drug, is A by all the components list of drug1, A2,
A3....Aw, amount to w kind ingredients, for selected drug i, it includes ingredient lists be { x1,x2.....,xk, xk∈Aw
(k≤w);
(2) the correlation analysis model based on sorting algorithm, the relevance size are F (yj), using such as minor function public affairs
Formula:
Wherein, wiIndicate the cost coefficient and w of drug ii∈ [0,1] is indicated if the cost caused by the wrong drug
Size, the more big then cost of cost coefficient are higher, wherein wipijIndicate validity of i-th of drug to jth kind tentative diagnosis, if
Validity>0, then it represents that drug i has facilitation to tentative diagnosis j.Conversely, then have inhibiting effect, PijIndicate that drug i belongs to
It is grouped the probability size and P of j in tentative diagnosisij∈ [0,1];Then PkjIndicate drug ingredient xkBelong to the general of tentative diagnosis grouping j
Rate size then obtains following function formula:
Wherein, PqjIndicate that drug ingredient q belongs to the probability size of tentative diagnosis grouping j;
Comprehensive (2-1) and (2-2), then obtain following function formula:
Its object function is:
The tentative diagnosis grouping j belongs to overall m kind tentative diagnosis grouping, then (1,2,3.......m) j ∈, each
The probability size that drug i belongs to each tentative diagnosis grouping j is Pij, then PijFor n × m dimension array, i.e.,:
Wherein (i≤n, j≤m); (2-5)
(3) process is issued for each prescription, can be considered and seeks from huge drug storage with the drug relevance most
The process simplification of careful side is by big drug:The relevance size of the drug i of selection and the tentative diagnosis grouping j of the disease
For F (yj), set its overall association size be as ε and ε (0,1] constant, as relevance size F (yj) be less than ε when indicate choosing
The drug i taken the and tentative diagnosis relationship grouping j of the disease is strong relationship, then prescription is to examining;As relevance size F (yj) be more than
Indicate that the tentative diagnosis relationship grouping j for the drug i and disease chosen is weak relationship when ε, then prescription is not to examining.
Relevance size F (yj) geometric meaning indicate as shown in Figure 1, region A indicates a kind of tentative diagnosis (in other words
One prescription), region B1, region B2, region B3, region Bi... the regions .. BnIndicate drug, region A is respectively in region B1, region
B2, region B3, region Bi... region BnIn size indicate probability.It degenerates if only recommending a kind of drug and is:Area
Domain A respectively with region B1, region B2, region B3, region Bi... region BnBetween repetition area.Drug BiTentatively examining
The bigger expression probability of area occurred in disconnected A is bigger.If all drugs are all related to A tentative diagnosis, it is considered as effective prescription, if
There is drug BiNot in the region where A tentative diagnosis, then it represents that drug BiIt is unrelated with tentative diagnosis A, or say drug BiIt is uncomfortable
With tentative diagnosis A.
Cost coefficient w in the formula (2-1)iThe preliminary applicability ω for being equal to drugi, wherein i i-th kind of medicine of expression
Product, according to formula:
Wherein, ωi=C 'i,C′i∈ (0,1), C 'iIndicate the applicability and C ' of drug iiHow many kinds of is appeared in equal to drug i
In different tentative diagnosis;And C 'i=[c1,c2,c3....,ci,....,cn], wherein ciIndicate that i-th kind of drug appears in how many kinds of
Different tentative diagnosis, wherein ci>=0, ciIt can learn to obtain by historical data.
Based on NB Algorithm, to the probability matrix PijCalculating steps are as follows:
(1) data initialization, for giving one group of tentative diagnosis Cj, split tentative diagnosis;
(2) prescription E is chosen in tentative diagnosisj, prescription EjIn include drug X1, X2......Xk;
(3) for drug XiAnd i≤k, and extract drug ingredient vector { x1,x2.....,xn, it is assumed that each drug ingredient it
Between independently of each other;If the drug is abandoned less than drug ingredient, is considered as invalid drug by extraction;
(4) drug ingredient is recorded to tentative diagnosis CjIn, if having there is the ingredient, go out occurrence in the record of the ingredient
Number plus 1, meanwhile, corresponding drug is recorded in tentative diagnosis CjIn occurrence number;
(5) record tentative diagnosis CjOccurrence number, total sample number adds 1;
(6) repeat step (2)-(5) until in prescription the study of all drugs finish to obtain tentative diagnosis list;
(7) tentative diagnosis list is traversed, drug ingredient is for tentative diagnosis C in calculating tentative diagnosis listjPosteriority it is general
Rate:
Wherein, xiFor drug ingredient and xi∈{x1,x2.....,xn};D is sample space;
(8) step (7) is repeated, until tentative diagnosis CjIn the calculating of all drug ingredients finish;
(9) all drugs in tentative diagnosis list are traversed;
(10) drug X is calculatediTo tentative diagnosis CjProbability:
With drug ingredient < x1,x2...xn> indicates drug Xi, due to drug < x1,x2...xnProbability p (the < x of >1,
x2...xn> it is) constant, normalized is done to it, according to Bayesian assumption, is unfolded to obtain using Bayesian formula:
(11) step (9)-(10) are repeated, until tentative diagnosis CjIn the probability calculations of all drugs finish;
(12) step (7)-(11) are repeated, until all drugs finish all tentative diagnosis efficiency calculations, obtain probability
Matrix Pij。
For being Bayesian assumption in step (3), introduced in Bayes classifier following it is assumed that in given classification
In C, it is assumed that all properties are mutual indepedent, then have:
Each drug ingredient A of drug1,A2....AnBetween independent mutually, each drug ingredient A1,A2....AnWith tentatively examine
Relationship between disconnected C is illustrated in fig. 2 shown below, constant we indicated using normalization factor a.Its in the case of specified drug X
The posterior probability of given tentative diagnosis classification C is:
Specified classification ciIt should meet:
If classifying ciFor optimal classification, then also need to meet:
p(ci| < a1,…,an>) > p (cj| < a1,…,an), > i ≠ j
Wherein a1,…,anIndicate the n kind ingredients of drug.
The prior probability distribution of classification C can simply obtain its maximal possibility estimation from training set data, and maximum is seemingly
So estimation is equal to the frequency that occurs in data set of different classes of attribute, and computation complexity is O (| D |).
For being described as follows for Bayesian Classification Model:
Classification has rule-based classification (inquiry) and irregular classification (having guidance learning).Bayes's classification is irregular
Classification, it by training set (classified example collection) training summarizes grader, and (predicted variable is discrete is known as point
Class, continuous to be known as returning), and classified to non-classified data using grader.It is representative in Bayes classifier
Grader have Naive Bayes Classifier, BAYESIAN NETWORK CLASSIFIER and tree enhancing Naive Bayes Classification Model TAN etc..
Bayes's classification has following features:
A. Bayes's classification is not an example to be absolutely assigned to certain one kind, but belong to a certain by being calculated
The probability of class, the class with maximum probability are the classes belonging to the example.
B. all properties in Bayes's classification all directly or indirectly play a role under normal circumstances, i.e., all categories
Property be involved in classification, rather than one or several attributes determine classification.
C. the attribute of Bayes's classification example can be discrete, continuous, can also be mixing.
Its prior probability based on h gives the data itself observed the probability of different data under assuming and observed.
The prior probability of hypothesis h before no training data is indicated with P (h).P (h) is referred to as the prior probability of h
(priorprobability), indicate about h be the probability correctly assumed background knowledge.The instruction for indicating to observe with P (D)
The prior probability for practicing data D, i.e., in a certain probability for assuming D immediately of no determination.P (D | h) it indicates to assume the feelings that h is set up
The probability of data D under condition.The probability that h is set up when we need to find out given training data D, i.e. the posterior probability P (D | h) of h, then
The method for calculating posterior probability P (D | h) is acquired by Bayesian formula:
Maximum posteriori is assumed:In many study scenes, learner considers candidate hypothesis set H, and demand is given birth to wherein
At the maximum hypothesis h ∈ H of the possibility of data D.The hypothesis of maximum likelihood is referred to as maximum posteriori it is assumed that being denoted as:hMAP
Wherein, it since P (D) is the constant for not depending on h, so we remove P (D), is replaced with constant a.
In this example, it is assumed that A1,A2....AnIndicate drug characteristic attribute (herein value be whole drug at
Point, amount to n kinds ingredient), it is assumed that the total m classification of tentative diagnosis, C={ C1,C2,C3,…,Cm}.Give a specific drug
The attribute of X, drug ingredient are { x1,x2.....,xn, x hereiIt is attribute AiSpecific value, indicate that drug X's is all
Composition information.The drug belongs to classification CiPosterior probability be P (X | Ci), C (X) indicates affiliated point that drug final classification obtains
Class, that is, maximum probability class.
I.e. which classification prediction drug finally belongs to, i.e., under the classification finally in the case of given drug ingredient
Posterior probability it is maximum.
It is described as follows for the algorithm of Bayes:
NB Algorithm is based on Bayes formula, we introduce condition probability formula first.Time A
Under conditions of generation, the probability that time B occurs, referred to as event B given event A conditional probability (also referred to as posterior probability),
Conditional probability is expressed as P (B | A), and correspondingly, P (A) is known as unconditional probability (also referred to as prior probability).
In the case where event A occurs, the probability that event B occurs is
Wherein:
P (AB)=P (A) P (B | A)=P (B) P (A | B)
It is obtained in conjunction with above formula:
Wherein it is assumed that S is sample space, A1,A2,A3,…AmFor a division of sample space S.Wherein P (Aj) 0 (j=of >
1,2,3 ..., m) according to total probability formula:
It is in conjunction with Bayesian formula is obtained:
P (A in square model are examined for intelligencej| B) indicate that drug B belongs to tentative diagnosis AjPosterior probability. P(B|Aj) table
Show tentative diagnosis AjIn include the probability of drug B, according to central-limit theorem P (Aj) be a constant can pass through historical statistical data
It obtains.
The present invention is not limited to above-mentioned optional embodiment, anyone can show that other are various under the inspiration of the present invention
The product of form, however, make any variation in its shape or structure, it is every to fall into the claims in the present invention confining spectrum
Technical solution, be within the scope of the present invention.
Claims (3)
1. a kind of intelligence based on machine learning examines square evolution model, which is characterized in that it includes the following contents:
(1) it is A by all the components list of drug1, A2, A3....Aw, amount to w kind ingredients, for selected drug i, it includes
Ingredient lists be { x1,x2.....,xk, xk∈Aw(k≤w);
(2) the correlation analysis model based on sorting algorithm, the relevance size are F (yj), using following function formula:
Wherein, wiIndicate the cost coefficient and w of drug ii∈ [0,1], PijThe probability that expression drug i belongs to tentative diagnosis grouping j is big
Small and Pij∈ [0,1];Then PkjIndicate drug ingredient xkThe probability size for belonging to tentative diagnosis grouping j is then obtained with minor function public affairs
Formula:
Wherein, PqjIndicate that drug ingredient q belongs to the probability size of tentative diagnosis grouping j;
Comprehensive (2-1) and (2-2), then obtain following function formula:
Its object function is:
The tentative diagnosis grouping j belongs to overall m kind tentative diagnosis grouping, then (1,2,3.......m) j ∈, each drug i
The probability size for belonging to each tentative diagnosis grouping j is Pij, then PijFor n × m dimension array, i.e.,:
(3) it is by the process simplification of careful side:The drug i of selection and the relevance size of the tentative diagnosis grouping j of the disease are F
(yj), set its overall association size be as ε and ε (0,1] constant, as relevance size F (yj) be less than ε when indicate choose
The tentative diagnosis relationship grouping j of drug i and the disease be strong relationship, then prescription is to examining;As relevance size F (yj) it is more than ε
When indicate that the tentative diagnosis relationship grouping j of drug i and the disease chosen is weak relationship, then prescription is not to examining.
2. the intelligence according to claim 1 based on machine learning examines square evolution model, which is characterized in that the probability square
Battle array PijCalculating steps are as follows:
(1) data initialization, for giving one group of tentative diagnosis Cj, split tentative diagnosis;
(2) prescription E is chosen in tentative diagnosisj, prescription EjIn include drug X1, X2......Xk;
(3) for drug XiAnd i≤k, and extract drug ingredient vector { x1,x2.....,xn, it is assumed that phase between each drug ingredient
It is mutually independent;If the drug is abandoned less than drug ingredient, is considered as invalid drug by extraction;
(4) drug ingredient is recorded to tentative diagnosis CjIn, if having there is the ingredient, occurrence number adds in the record of the ingredient
1, meanwhile, corresponding drug is recorded in tentative diagnosis CjIn occurrence number;
(5) record tentative diagnosis CjOccurrence number, total sample number adds 1;
(6) repeat step (2)-(5) until in prescription the study of all drugs finish to obtain tentative diagnosis list;
(7) tentative diagnosis list is traversed, drug ingredient is for tentative diagnosis C in calculating tentative diagnosis listjPosterior probability:
Wherein, xiFor drug ingredient and xi∈{x1,x2.....,xn};D is sample space;
(8) step (7) is repeated, until tentative diagnosis CjIn the calculating of all drug ingredients finish;
(9) all drugs in tentative diagnosis list are traversed;
(10) drug X is calculatediTo tentative diagnosis CjProbability:
With drug ingredient < x1,x2...xn> indicates drug Xi, due to drug < x1,x2...xnProbability p (the x of >1,x2...xn)
For constant, normalized is done to it, according to Bayesian assumption, is unfolded to obtain using Bayesian formula:
(11) step (9)-(10) are repeated, until tentative diagnosis CjIn the probability calculations of all drugs finish;
(12) step (7)-(11) are repeated, until all drugs finish all tentative diagnosis efficiency calculations, obtain probability matrix Pij。
3. the intelligence according to claim 1 based on machine learning examines square evolution model, which is characterized in that the formula
Cost coefficient w in (2-1)iThe preliminary applicability ω for being equal to drugi, wherein i indicates i-th kind of drug, according to formula:
Wherein, ωi=C 'i,C′i∈ (0,1), C 'iIndicate the applicability and C ' of drug iiHow many kinds of difference is appeared in equal to drug i
In tentative diagnosis;And C 'i=[c1,c2,c3....,ci,....,cn], wherein ciIndicate that i-th kind of drug appears in how many kinds of difference
Tentative diagnosis, wherein ci>=0, ciIt can learn to obtain by historical data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810317800.4A CN108717862B (en) | 2018-04-10 | 2018-04-10 | Intelligent trial and development method based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810317800.4A CN108717862B (en) | 2018-04-10 | 2018-04-10 | Intelligent trial and development method based on machine learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108717862A true CN108717862A (en) | 2018-10-30 |
CN108717862B CN108717862B (en) | 2022-05-03 |
Family
ID=63898920
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810317800.4A Active CN108717862B (en) | 2018-04-10 | 2018-04-10 | Intelligent trial and development method based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108717862B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110706775A (en) * | 2019-09-29 | 2020-01-17 | 北京一药良心信息科技有限公司 | Remote intelligent party checking method and system |
CN111383732A (en) * | 2020-03-23 | 2020-07-07 | 平安医疗健康管理股份有限公司 | Medicine auditing method, device, computer system and readable storage medium based on mutual exclusion identification |
CN111640481A (en) * | 2020-05-22 | 2020-09-08 | 南京大学 | Multi-source data fusion-based Parkinson's disease medicine recommendation model |
CN116825364A (en) * | 2023-08-29 | 2023-09-29 | 江苏盛泰科技集团有限公司 | High-risk group health identification judgment system |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020042725A1 (en) * | 1994-10-28 | 2002-04-11 | Christian Mayaud | Computerized prescription system for gathering and presenting information relating to pharmaceuticals |
CN101313221A (en) * | 2005-09-21 | 2008-11-26 | Ccc诊断有限责任公司 | Comprehensive diagnostic testing procedures for personalized anticancer chemotherapy(pac) |
CN103365978A (en) * | 2013-07-01 | 2013-10-23 | 浙江大学 | Traditional Chinese medicine data mining method based on LDA (Latent Dirichlet Allocation) topic model |
WO2014061936A1 (en) * | 2012-10-16 | 2014-04-24 | An Hyeong Jun | System for searching for drug that minimizes individual side effect based on individual single nucleotide polymorphism, and method thereof |
CN104978474A (en) * | 2014-04-11 | 2015-10-14 | 中国中医科学院中医临床基础医学研究所 | Medicine effect evaluating method based on molecular network and medicine effect evaluating system |
CN106108846A (en) * | 2016-06-20 | 2016-11-16 | 中山大学 | A kind of intelligent drug risk monitoring method and system |
CN107302958A (en) * | 2016-04-19 | 2017-10-31 | 郭思东 | The Functional Design of Chinese medicine health-care series beverage and the method for formulation optimization |
CN107341345A (en) * | 2017-06-22 | 2017-11-10 | 厦门大学 | A kind of clinical rational drug use methods of risk assessment based on electronic health record big data |
CN107578818A (en) * | 2017-08-25 | 2018-01-12 | 广州宝荣科技应用有限公司 | A kind of auxiliary evolution method and device based on deep learning |
CN206932392U (en) * | 2017-07-24 | 2018-01-26 | 四川骏逸富顿科技有限公司 | A kind of micro- interrogation service platform audio frequency processing system |
-
2018
- 2018-04-10 CN CN201810317800.4A patent/CN108717862B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020042725A1 (en) * | 1994-10-28 | 2002-04-11 | Christian Mayaud | Computerized prescription system for gathering and presenting information relating to pharmaceuticals |
CN101313221A (en) * | 2005-09-21 | 2008-11-26 | Ccc诊断有限责任公司 | Comprehensive diagnostic testing procedures for personalized anticancer chemotherapy(pac) |
WO2014061936A1 (en) * | 2012-10-16 | 2014-04-24 | An Hyeong Jun | System for searching for drug that minimizes individual side effect based on individual single nucleotide polymorphism, and method thereof |
CN103365978A (en) * | 2013-07-01 | 2013-10-23 | 浙江大学 | Traditional Chinese medicine data mining method based on LDA (Latent Dirichlet Allocation) topic model |
CN104978474A (en) * | 2014-04-11 | 2015-10-14 | 中国中医科学院中医临床基础医学研究所 | Medicine effect evaluating method based on molecular network and medicine effect evaluating system |
CN107302958A (en) * | 2016-04-19 | 2017-10-31 | 郭思东 | The Functional Design of Chinese medicine health-care series beverage and the method for formulation optimization |
CN106108846A (en) * | 2016-06-20 | 2016-11-16 | 中山大学 | A kind of intelligent drug risk monitoring method and system |
CN107341345A (en) * | 2017-06-22 | 2017-11-10 | 厦门大学 | A kind of clinical rational drug use methods of risk assessment based on electronic health record big data |
CN206932392U (en) * | 2017-07-24 | 2018-01-26 | 四川骏逸富顿科技有限公司 | A kind of micro- interrogation service platform audio frequency processing system |
CN107578818A (en) * | 2017-08-25 | 2018-01-12 | 广州宝荣科技应用有限公司 | A kind of auxiliary evolution method and device based on deep learning |
Non-Patent Citations (1)
Title |
---|
张文娣: "抗生素相关性腹泻模型的构建", 《中国优秀博硕士学位论文全文数据库(硕士)医药卫生科技辑》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110706775A (en) * | 2019-09-29 | 2020-01-17 | 北京一药良心信息科技有限公司 | Remote intelligent party checking method and system |
CN111383732A (en) * | 2020-03-23 | 2020-07-07 | 平安医疗健康管理股份有限公司 | Medicine auditing method, device, computer system and readable storage medium based on mutual exclusion identification |
CN111383732B (en) * | 2020-03-23 | 2023-04-18 | 平安医疗健康管理股份有限公司 | Medicine auditing method, device, computer system and readable storage medium based on mutual exclusion identification |
CN111640481A (en) * | 2020-05-22 | 2020-09-08 | 南京大学 | Multi-source data fusion-based Parkinson's disease medicine recommendation model |
CN111640481B (en) * | 2020-05-22 | 2023-06-20 | 南京大学 | Parkinson's disease drug recommendation model based on multi-source data fusion |
CN116825364A (en) * | 2023-08-29 | 2023-09-29 | 江苏盛泰科技集团有限公司 | High-risk group health identification judgment system |
Also Published As
Publication number | Publication date |
---|---|
CN108717862B (en) | 2022-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Jaiswal et al. | Classification of the COVID-19 infected patients using DenseNet201 based deep transfer learning | |
US11669965B2 (en) | AI-based label generating system and methods for use therewith | |
Basiri et al. | A novel method for sentiment classification of drug reviews using fusion of deep and machine learning techniques | |
US11551795B2 (en) | AI-based multi-label heat map generating system and methods for use therewith | |
US11636309B2 (en) | Systems and methods for modeling probability distributions | |
Heidari et al. | The COVID-19 epidemic analysis and diagnosis using deep learning: A systematic literature review and future directions | |
CN108717862A (en) | A kind of careful square evolution model of the intelligence based on machine learning | |
CN108596199A (en) | Unbalanced data classification method based on EasyEnsemble algorithms and SMOTE algorithms | |
CN107895168A (en) | The method of data processing, the device of data processing and computer-readable recording medium | |
Tiwari et al. | A knowledge infused context driven dialogue agent for disease diagnosis using hierarchical reinforcement learning | |
CN115050442B (en) | Disease category data reporting method and device based on mining clustering algorithm and storage medium | |
CN112201359A (en) | Artificial intelligence-based critical illness inquiry data identification method and device | |
Liu et al. | Knowledge-aware deep dual networks for text-based mortality prediction | |
Chen et al. | Using data mining strategies in clinical decision making: a literature review | |
Spanier et al. | A new method for the automatic retrieval of medical cases based on the RadLex ontology | |
Teimouri et al. | Detecting diseases in medical prescriptions using data mining tools and combining techniques | |
CN112991079B (en) | Multi-card co-occurrence medical treatment fraud detection method, system, cloud end and medium | |
CN112635050B (en) | Diagnosis recommendation method, electronic equipment and storage device | |
CN115240811A (en) | Construction method and application of implicit relation drug recommendation model based on graph neural network | |
Bickel | Learning under differing training and test distributions | |
Segev et al. | Internet as a knowledge base for medical diagnostic assistance | |
Wei et al. | A Bayesian network under strict chain model for computing flow risks in smart city | |
Ganjdanesh et al. | Multi-modal genotype and phenotype mutual learning to enhance single-modal input based longitudinal outcome prediction | |
Arya et al. | Performance analysis of deep transfer learning for manifestation of COVID-19 using chest X-ray | |
Mary et al. | Design and development of big data-based model for detecting fraud in healthcare insurance industry |
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 |