CN108573758A - A kind of intelligent medical big data service system and application process - Google Patents
A kind of intelligent medical big data service system and application process Download PDFInfo
- Publication number
- CN108573758A CN108573758A CN201810386146.2A CN201810386146A CN108573758A CN 108573758 A CN108573758 A CN 108573758A CN 201810386146 A CN201810386146 A CN 201810386146A CN 108573758 A CN108573758 A CN 108573758A
- Authority
- CN
- China
- Prior art keywords
- hospital
- cloud platform
- user
- data
- health
- 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
- G16H80/00—ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring
-
- 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
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Pathology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a kind of intelligent medical big data service systems.Including subscription client, subscription client is connect with hospital client;The subscription client includes the private clound health data center for storing the illness information of user, diagnosis and therapy recording, health data and disease model, and private clound health data center is connect with data acquisition and monitoring unit;Hospital's client includes hospital's cloud platform, and hospital's cloud platform is connect with more than one doctor's sub-platform;The private clound health data center of hospital's cloud platform and user connects.The present invention can realize that medical resource is shared, reduce medical services spending, alleviate hospital's pressure, improve diagnosis efficiency, mitigate outpatients mental state burden.
Description
Technical field
The present invention relates to medical information field, especially a kind of intelligent medical big data service system and application process.
Background technology
After 10 years high speed developments of Medical and health information system and biological gene technology, medical field enters " big number
According to the epoch ", explosive growth is presented in data.The digital research report that International Data Corporation (IDC) provides shows that the whole world was created in only 2011
The data volume built and replicated is more than 1.8ZB, if data growth trend follows new Moore's Law, i.e. global metadata every two years turns over
, then the year two thousand twenty data volume will be up to 30ZB.By taking the prediction address of Mai Kenxi as an example, the data in medical services place in 2012
Total amount is about 5000PB, the medical imaging filing of storage in 2014, electronic health record, medical research information, the files such as record of being hospitalized
Storage reaches nearly 10000PB.However, so huge data utilization rate is but less than 20%, many data are deposited as " historical relic "
Storage be nobody shows any interest in " data museum ", this not only causes great trouble for the data storage of hospital, while also bringing
The waste of data, prevents the data from playing its effect well.
The distribution of today's society medical resource is uneven, and good medical services resource concentrates on a line city, rural medical treatment
Resource is seriously deficient.The shortage of healthcare givers causes a doctor to be met daily to hundreds of patients, and huge patient's diagnosis and treatment need
The huge contradiction that summation medical resource relative deficiency is formed becomes the principal element of today's society conflict between doctors and patients.According to statistics,
2002~2012 years, increase by 23% every year on average for the incident of violence of medical worker, every hospital there are 27 to attack every year on average
The event for hitting medical worker occurs.Even in the U.S. of medical field prosperity, also there is nearly half people by inappropriate every year
Treatment;Nosocomial infection is suffered from more than 2,000,000 people;More than the complication that 1,000,000 people are disabled in surgical operation, and wherein
Half is avoidable.
Aging of population is that medical field brings new challenge.It is expected that the year two thousand twenty and the year two thousand fifty, 60 years old Chinese or more old man
2.34 hundred million people and 4.37 hundred million people will be respectively reached, shared population ratio will be respectively more than 16% and 30%;Aging of population is brought
Chronic disease, healthcare givers's shortage cause medical expense to occupy the significant proportion and cumulative year after year of social total expenditure, in advance
Meter causes 2010~the year two thousand thirty of medical expense actually to increase about 5.2% every year.According to the report of United States Medicine institute, nowadays medical treatment & health
The 1/3 of expenditure is wasted without being used to improve medical treatment.These wastes include unnecessary service, administrative waste, expensive doctor
Treatment expense, medical treatment fraud and the chance for missing prevention.
Invention content
The object of the present invention is to provide a kind of intelligent medical big data service systems.The present invention can realize medical money
Source is shared, reduces medical services spending, alleviates hospital's pressure, improve diagnosis efficiency, mitigates outpatients mental state burden.
Technical scheme of the present invention:A kind of intelligent medical big data service system, including subscription client, subscription client
It is connect with hospital client;The subscription client includes for storing the illness information of user, diagnosis and therapy recording, health data
With the private clound health data center of disease model, private clound health data center is connect with data acquisition and monitoring unit;It is described
Hospital's client include hospital's cloud platform, hospital's cloud platform is connect with more than one doctor's sub-platform;Hospital's cloud
The private clound health data center of platform and user connect.
Intelligent medical big data service system above-mentioned further include medicine enterprise cloud platform, medicine look forward to cloud platform respectively with the private of user
There is cloud health data center to be connected with hospital cloud platform.
The application process of intelligent medical big data service system above-mentioned, hospital's cloud platform establish pathological model, will be described
Pathological model is stored in the private clound health data center of subscription client;Illness information, diagnosis and therapy recording and the data of user are adopted
The private clound health data center of the health data deposit user for the user that monitor set unit acquires in real time, the private clound of user are strong
Health data center is matched according to the collected health data pathological model corresponding in cloud platform of institute, to the real-time of user
Health condition is analyzed;When analysis result is health, periodically to user feedback health and fitness information.
In the application process of intelligent medical big data service system above-mentioned, the analysis result is illness or may
When illness, private clound health data center is diagnosed and is provided treatment advice, and treatment advice is effective, then feeds back treatment results
To the private clound health data center of user, data acquisition and monitoring unit is continued through later, user health is supervised in real time
It surveys.
In the application process of intelligent medical big data service system above-mentioned, when the treatment advice is invalid, private clound
The pathological information that diagnosis obtains is transmitted to hospital's cloud platform of hospital's client by health data center, and hospital's cloud platform is according to disease
Reason information is classified to user's illnesses and is assigned on doctor's sub-platform of corresponding section office, and doctor passes through hospital's cloud platform
The therapeutic scheme of previous case or the patient for once suffering from this kind of disease are analyzed to assist diagnosis to obtain doctor
Treatment advice, later, the private clound health data center that medical treatment opinion is sent to user by hospital's cloud platform are controlled for user
It treats.
In the application process of intelligent medical big data service system above-mentioned, the foundation of the pathological model is specifically,
By Flume tools by collected medical data be transferred to hospital's cloud platform or by by database data import hospital
The mode of cloud platform stores medical data;By high performance parallel computation platforms of the MapReduce based on cluster to doctor
It treats data to be handled, by association rule algorithm, to treated, medical data carries out analyzing the pass found out between disease later
Connection property, the pathological model of corresponding medicine is established using decision tree.
In the application process of intelligent medical big data service system above-mentioned, it is described by association rule algorithm to processing
Medical data afterwards carries out analyzing the relevance found out between disease, and specific implementation process is using in association rule algorithm
FP-growth algorithms excavate the frequent item set in treated medical data, after obtaining frequent item set, according to minimum support and
Min confidence obtains correlation rule.
In the application process of intelligent medical big data service system above-mentioned, medicine enterprise's cloud platform is for obtaining user
The partial data at private clound health data center, hospital's cloud platform medical data, by cloud platform to the analyses of data with
Rear line private clound health data center, hospital's cloud platform send drug information.
Advantageous effect
Compared with prior art, the present invention monitor in real time acquisition user health data be transmitted in private clound health data
The heart analyzes the health data of user by private clound health data center, realizes the real time monitoring to user health;
When the analysis result of user is illness, private clound health data center is diagnosed and is provided treatment advice, and treatment advice has
Treatment results, then is fed back to the private clound health data center of user by effect;In this way, private clound health data can be passed through
The disease that the treatment advice that center provides is cured no longer needs to hospital be treated, and hospital's pressure has been effectively relieved, has reduced doctor
Service spending is treated, and makes the treatment of disease more convenient.
When the treatment advice that the present invention provides at private clound health data center cannot cure disease, by private clound health number
The pathological information that diagnosis obtains is transmitted to hospital's cloud platform of hospital's client according to center, hospital's cloud platform is according to pathological information
Classify to user's illnesses and be assigned on doctor's sub-platform of corresponding section office, doctor is by hospital's cloud platform to previous
Case or once suffered from this kind of disease patient therapeutic scheme analyzed with assist diagnosis obtain medical treatment meaning
See, medical treatment opinion is sent to the private clound health data center of user for user's treatment by hospital's cloud platform later.Pass through
This method, one side doctor can first understand conditions of patients before the patient that sees and treat patients, additionally it is possible to obtain previous case or
Person once suffered from the therapeutic scheme auxiliary diagnosis of the patient of this kind of disease, effectively increased diagnosis efficiency and the reduction of doctor in this way
Misdiagnosis rate;On the other hand, patient can select the hospital for more properly treating the disease to upload pathological information, best to obtain
Therapeutic scheme greatly reduces medical facilities construction cost in this way, realizing the shared of high-quality medical resource.
In conclusion the present invention can realize that medical resource is shared, medical services spending is reduced, alleviates hospital's pressure, carries
High diagnosis efficiency, mitigating outpatients mental state burden, (going hospital's medical treatment excessively to constrain for many people may bring very greatly
Psychological pressure, especially patients with chronic diseases), so that user is avoided the disease that can prevent as possible, analyze disease data in real time
Diagnosis and treatment suggestion is provided in real time.
Description of the drawings
Fig. 1 is the service flow diagram of the present invention;
Fig. 2 is the logic diagram of the present invention;
Fig. 3 is that Flume transmits form;
Fig. 4 is the system assumption diagram of MapReduce;
Fig. 5 is the FP trees of storage compression frequent mode information;
Fig. 6 is the High relevancy figure between the disease of part;
Fig. 7 is the structural schematic diagram of the present invention.
Specific implementation mode
The present invention is further illustrated with reference to the accompanying drawings and examples, but be not intended as to the present invention limit according to
According to.
Embodiment 1.A kind of intelligent medical big data service system, structure such as Fig. 7 institutes, including subscription client, Yong Huke
Family end is connect with hospital client;The subscription client includes illness information, diagnosis and therapy recording, the health for storing user
The private clound health data center of data and disease model, private clound health data center is connect with data acquisition and monitoring unit;
Hospital's client includes hospital's cloud platform, and hospital's cloud platform is connect with more than one doctor's sub-platform;The doctor
The private clound health data center of institute's cloud platform and user connect.The data acquisition and monitoring unit can be wearable intelligence
Equipment.
Intelligent medical big data service system above-mentioned further include medicine enterprise cloud platform, medicine look forward to cloud platform respectively with the private of user
There is cloud health data center to be connected with hospital cloud platform.
The application process of intelligent medical big data service system above-mentioned:Hospital's cloud platform is analyzed by association rule algorithm
Relevance between disease establishes pathological model using decision tree, and the pathological model is stored in the private clound of subscription client
Health data center;The health for the user that the illness information of user, diagnosis and therapy recording and data acquisition monitoring unit are acquired in real time
Data are stored in the private clound health data center of user, and user private clound health data center is according to the collected health data of institute
Pathological model corresponding in cloud platform is matched, and is analyzed the real time health situation of user;Analysis result is health
When, periodically to user feedback health and fitness information.
When analysis result above-mentioned is illness or possible illness, private clound health data center, which is diagnosed and provided, to be controlled
Opinion is treated, treatment advice is effective, then treatment results are fed back to the private clound health data center of user, continue through number later
User health is monitored in real time according to acquisition monitoring unit.
When treatment advice above-mentioned is invalid, the pathological information that diagnosis obtains is transmitted to hospital by private clound health data center
Hospital's cloud platform of client, hospital's cloud platform classify to user's illnesses according to pathological information and are assigned to corresponding section
On doctor's sub-platform of room, treatment of the doctor by hospital's cloud platform to previous case or the patient for once suffering from this kind of disease
Scheme is analyzed to assist diagnosis to obtain medical treatment opinion, and later, hospital's cloud platform sends medical treatment opinion
It is treated to the private clound health data center of user for user.
User can also pass through the relevance and correspondence of oneself previous case and disease in private clound health data center
The relevance of symptom and disease carries out checking oneself for disease.
The foundation of pathological model above-mentioned is specifically to be transferred to collected medical data acquisition by Flume tools
Hospital's cloud platform is stored medical data in such a way that data in database are imported hospital's cloud platform;Pass through
High performance parallel computation platforms of the MapReduce based on cluster, i.e., the distributed data processing method of parallel computation is to medical number
According to being handled, by association rule algorithm, to treated, medical data carries out analyzing the association found out between disease later
Property, the pathological model of corresponding medicine is established using decision tree.Medical data includes from wearable smart machine, electronics disease
Go through, medical image, clinical examination, medical literature, doctors and patients' behavior, medical insurance industry, pharmaceutical industry, the ground such as medicine marketing enterprise
The medical data of side.
It stores for medical data acquisition is transferred to hospital's cloud platform by Flume, passes through in the present embodiment
Flume comes data from the data source collection in Fig. 2, and the data being collected into are sent to specified Sink (destination).
Flume is a high availability, the acquisition of reliable and distributed massive logs, polymerization and Transmission system, design
Principle, which is also based on, collects data flow in storage to the pooled storages such as HDFS, HBase from various Website servers.
Most basic unit of the event (event) as Flume internal data transfers, flows in the transmission process of entire data
That dynamic is event.The byte arrays and an optional head (Header) that event (event) reprints data by one are constituted.
The minimum stand-alone unit of Flume is Agent, and Agent itself is an independent finger daemon JVM, it from
Client or others Agent receive data, and the data of acquisition are then transmitted to next destination node Sink or Agent.
Agent is mainly made of three Source (source), Channel (channel), Sink (destination) components, and Agent is by managing it
Component complete process of the flow of event from an external source to destination.Source receives data from number generator, and will connect
The data of receipts pass to one or more Channel with the event formats of Flume;Channel is that an of short duration storage is held
Device can be first by the data buffer storage of the event formats received from Source before being sent to Sink, and pending data really reaches
After Sink is consumed, Flume deletes the data of oneself caching again.Channel can be with any number of Source and Sink chains
It connects.Sink stores data into pooled storage (such as HBase and HDFS), it from Channel consumption datas (event) and by its
Pass to another Sink or HDFS or HBase.Shown in transmission form such as Fig. 3 (a) and Fig. 3 (b) of Flume.
After collected data are stored by approach shown in Fig. 3 by MapReduce based on parallel computation point
Cloth data processing method is handled.MapReduce is a high performance parallel computation platform based on cluster, it allows city
It includes tens of, hundreds of distributions and parallel computing trunking to many thousands of nodes that common commercial server, which constitutes one, on field.It
Host-guest architecture is taken, there are one control node and multiple working nodes in a MapReduce cluster.When cluster is run,
All working nodes periodically can send heartbeat message to control node, report this node current state.After receiving heartbeat message,
Control node can send command information according to the state of current working condition and working node itself to working node.Work section
Point completes corresponding actions according to the command information received.In MapReduce frames, the base for the data processing work that user carries out
Our unit is " operation ".In MapReduce clusters, " operation " is divided into two stages of Map and Reduce to execute.And every
A stage, and there are multiple tasks executing parallel.These tasks are assigned on multiple working nodes and execute, and complete basic number
According to processing work.MapReduce uses Master/Slave (M/S) framework, it is mainly made of following component:
Client, JobTracker, TaskTracker and Task.The architecture of MapReduce is as shown in Figure 4.What user write
MapReduce programs are submitted to the ends JobTracker by client;Meanwhile user can be connect by some provided of Client
Mouth checks job run state.One MapReduce program may correspond to several operations, and each operation be broken down into it is several
A Map/Reduce tasks.JobTracker is mainly responsible for monitoring resource and job scheduling in figure.JobTracker monitoring is all
TaskTracker and operation health status, corresponding task can be transferred to other nodes after finding failure scenarios;
The information such as implementation progress, the resource usage amount of the tracing task of JobTracker meetings simultaneously, and tell these information to task scheduling
Device, and task dispatcher can select suitable task to use these resources when resource occurs idle.In Hadoop, task
Scheduler is a pluggable module, and user can design corresponding scheduler according to the needs of oneself.TaskTracker meetings
The operation progress of the service condition of resource on this node and task is reported to periodically by Heartbeat
JobTracker, while receiving the order that JobTracker is sent and executing corresponding operation (as started new task, killing task
Deng).TaskTracker is used
" slot " equivalent divides the stock number on this node." slot " represents computing resource (CPU, memory etc.).One
Task just has an opportunity to run after getting a slot, and the effect of Hadoop schedulers be exactly will be on each TaskTracker
Idle slot distribute to Task use.It is Mapslot and two kinds of Reduceslot that slot, which is divided to, respectively for MapTask and
ReduceTask is used.TaskTracker limits the concurrency of Task by slot numbers (configurable parameter).Task points are
MapTask and two kinds of ReduceTask, is started by TaskTracker.The processing unit of MapReduce is split.split
It is a logical concept, it is only comprising the section where some metadata informations, such as data start, data length, data
Point etc..Its division methods are defined by user oneself.Each split can transfer to a MapTask processing, MapTask first will
Split iterative resolutions call user-defined map () function to be handled, finally successively at key/value pairs one by one
Result is stored on local disk, wherein ephemeral data is divided into several partition (fragment), each partition
By a ReduceTask processing.
It is above-mentioned that by association rule algorithm, to treated, medical data carries out analyzing the relevance found out between disease,
Specific implementation process is to utilize the frequent episode in the medical data that excavates that treated of the FP-growth algorithms in association rule algorithm
Collection, after obtaining frequent item set, correlation rule is obtained according to minimum support and min confidence.
To obtain correlation rule, need first to find out frequent item set.Utilize the tool of FP-growth algorithm Mining Frequent Itemsets Baseds
Body realizes that process is as follows:
The excavation of frequent item set is to excavate institute by the comparison of the support and threshold value of item collection from given data set
Some frequent item sets.If I={ I1, I2..., ImBe item set, give a transaction database D, wherein each affairs T is I
Nonvoid subset, i.e. each transaction is corresponding with one unique identifier TID (Transaction ID).If A is one
Item collection, affairs T include A, and if only ifCorrelation rule be shaped likeImplication, whereinA≠
φ, B ≠ φ, and A ∩ B=φ.RuleSet up in transaction set D, have support s, wherein s be in D include A ∪ B
The percentage of (i.e. the union of set A and B), i.e. P (A ∪ B).RuleThere is confidence level c in transaction set D, wherein c is D
In include A affairs while also include B affairs percentage, i.e. conditional probability P (A Shu B).I.e.:
Once finding out frequent item set by the affairs in database D, so that it may (strong to close directly to there is them to generate Strong association rule
Connection rule meets minimum support and min confidence), the calculating about confidence level can be as obtained by following formula:
The method for digging of frequent item set utilizes FP-growth algorithms, it is assumed that the Transaction Information of certain diseases of certain hospital
As shown in table 1, first time scanning is carried out to affairs first, exports the set of 1 item collection, and obtain their support counting, if
Most ramuscule degree is 2.Obtained result set L={ { I2: 7 }, { I1: 6 }, { I3: 6 }, { I4: 2 } { I5: 2 } }.Then FP trees are constructed,
First, the root node for creating tree, is marked with null.Second of scan database D.Item in each affairs is handled by L order
(sorting according to the support counting to successively decrease), and a branch is created to each affairs.In general, when an affairs consider to increase
When bonus point branch, the counting of each node on common prefix increases by 1, is that the item after prefix creates node and link.In order to
Facilitate traversal of tree, creates an item head table, each item is made to be directed toward its position in tree by a node chain.Scanning is all
Affairs (scanning n-th obtain the set of the n-th item collection) after obtain FP trees, as shown in Figure 5.And then the frequent mode generated
As shown in table 2.Assuming that minimal confidence threshold is 95%, then there was only { I2, I5(confidence level 100%) and I1, I5{ I1, I2}
(confidence level 100%) is strong correlation.By the method, we can reach to being predicted disease according to symptom in disease
The effect treated in advance before occurring.Fig. 6 be the relevance between the disease of part excavated by FP-growth algorithms, wherein
The connection of two node a lines represents between corresponding two kinds of diseases that there are High relevancies.
Table 1
Table 2
The present embodiment is used for carrying out clinical disease auxiliary diagnosis by the decision tree in sorting algorithm, from clinical database
Diagnostic rule is extracted, rate of correct diagnosis is improved.Decision tree is the machine learning for having supervision, i.e. the study of grader is being apprised of often
A trained tuple, which belongs to, to be carried out under " supervision " of which class.This method is implemented as follows:
The construction of decision tree uses C4.5 algorithms, if node N is represented or the tuple of storage subregion D.According to information gain-ratio
To obtain split point.
The required expectation information of tuple classification in D is given by:
Wherein, PiIt is that arbitrary tuple belongs to class C in DiNonzero probability is used in combination | Ci,D|/D estimates.Use the logarithm bottom of for 2
Because of information binary coding when function.Info (D) is the required average information of class label for identifying tuple in D, again
The referred to as entropy of D.Assuming that dividing the tuple in D by certain attribute A, wherein attribute A has v different value according to training data observation
{a1,a2,…,av}.If A is centrifugal pump, these values correspond to the v output tested on A.D can be divided with attribute A
For v subregion or subset { D1,D2,…,Dv, wherein DjIncluding the tuple in D, their A values are aj.These subregions correspond to from
The branch that node N grows out.It is calculate by the following formula and divides the expectation information arrived required for the tuple classification of D by A:
Its middle termServe as the weight of j-th of subregion.The expectation information needed is smaller, and subregion degree is higher.
Information gain is defined as the difference between original information requirement and new information requirement.I.e.:
Gain (A)=Info (D)-InfoA(D)
Gain (A) teaches that we have obtained how many by the division on A.
Division information is defined as follows:
SplitInfoA(D) it represents and is divided into the information that v subregion of corresponding attribute A tests generates by training dataset D.
Ratio of profit increase is defined as:
Select the attribute of maximum gain ratio as Split Attribute.It then proceedes to repeat the mistake in child node after cleaving
Journey finally obtains decision tree.
When decision tree creates, due to the noise and outlier in data, many branch reflections are different in training set
Often.By handling this excessive fitting problems with the method for beta pruning, C4.5 uses the method (PEP) of pessimistic beta pruning, specific algorithm
It is as follows:
Assuming that e (t) is error at t;I is covering TtLeaf;NtFor subtree TtLeaf number;N (t) is to be instructed at node t
Practice the number of example.
If e ' (t)≤e ' (Tt)+Se(e′(Tt)) set up, then TtIt should be tailored.
During the beta pruning of PEP algorithms, every stalk tree in tree at most needs to access once, in the worst case,
Its calculating time complexity is also only linear with the nonleaf node number of non-beta pruning tree, so PEP algorithms are considered working as
One of higher algorithm of precision in preceding decision tree post pruning method.
Medicine above-mentioned enterprise cloud platform be used to obtain the private clound health data center of user partial data (needed when acquisition through
Cross user license, while the personal information of user need to pass through concealmentization handle after can submit data to medicine enterprise cloud platform), cure
The medical data of institute's cloud platform, by cloud platform to the analysis of data with rear line private clound health data center, hospital's cloud
Platform sends drug information.The effect of medicine enterprise obtains drug by user private clound health data center and hospital cloud platform with
And the relevant information of side effect, the adjustment of drug is carried out, meanwhile, current certain diseases are obtained by the analysis of hospital's cloud platform data
The aggregation of sick frequency of disease development and disease patient, then carry out the marketing of targetedly drug production and drug.
In addition, medicine enterprise can provide corresponding drug information to hospital and user, carry out the popularization of drug, realize medicine enterprise and hospital end with
And the intercommunication of user terminal.
Claims (8)
1. a kind of intelligent medical big data service system, which is characterized in that including subscription client, subscription client and hospital visitor
Family end connects;The subscription client includes for storing the illness information of user, diagnosis and therapy recording, health data and disease mould
The private clound health data center of type, private clound health data center is connect with data acquisition and monitoring unit;The hospital visitor
Family end includes hospital's cloud platform, and hospital's cloud platform is connect with more than one doctor's sub-platform;Hospital's cloud platform and use
The private clound health data center at family connects.
2. intelligent medical big data service system according to claim 1, it is characterised in that:The system further includes medicine enterprise cloud
Platform, medicine enterprise cloud platform are connect with the private clound health data center of user and hospital's cloud platform respectively.
3. a kind of application process of intelligent medical big data service system as claimed in claim 1 or 2, it is characterised in that:Doctor
Institute's cloud platform establishes pathological model, and the pathological model is stored in the private clound health data center of subscription client;By user
Illness information, the private clound of diagnosis and therapy recording and the health data of the user that acquires in real time of data acquisition monitoring unit deposit user
Health data center, the private clound health data center of user is according to the collected health data of institute disease corresponding in cloud platform
Reason model is matched, and is analyzed the real time health situation of user;It is periodically strong to user feedback when analysis result is health
Health information.
4. the application process of intelligent medical big data service system according to claim 3, it is characterised in that:Point
When analysis result is illness or possible illness, private clound health data center is diagnosed and is provided treatment advice, treatment advice
Effectively, then the private clound health data center that treatment results are fed back to user, continues through data acquisition and monitoring unit later
User health is monitored in real time.
5. the application process of intelligent medical big data service system according to claim 4, it is characterised in that:Described controls
When treatment opinion is invalid, hospital's cloud that the pathological information that diagnosis obtains is transmitted to hospital's client by private clound health data center is put down
Platform, hospital's cloud platform classify to user's illnesses according to pathological information and are assigned to doctor's sub-platform of corresponding section office
On, doctor by hospital's cloud platform to the therapeutic scheme of previous case or the patient for once suffering from this kind of disease analyzed with
Auxiliary diagnosis obtains medical treatment opinion, and later, medical treatment opinion is sent to the private clound of user by hospital's cloud platform
It is treated for user at health data center.
6. the application process of intelligent medical big data service system according to claim 3, it is characterised in that:The disease
Reason model foundation be specifically, by Flume tools by collected medical data be transferred to hospital's cloud platform or pass through by
The mode of data importing hospital cloud platform stores medical data in database;Pass through height of the MapReduce based on cluster
Performance parallel computing platform handles medical data, and by association rule algorithm, to treated, medical data carries out later
The relevance between disease is found out in analysis, and the pathological model of corresponding medicine is established using decision tree.
7. the application process of intelligent medical big data service system according to claim 6, it is characterised in that:Described is logical
Crossing association rule algorithm, medical data carries out analyzing the relevance found out between disease to treated, and specific implementation process is profit
The frequent item set in treated medical data is excavated with the FP-growth algorithms in association rule algorithm, obtains frequent item set
Afterwards, correlation rule is obtained according to minimum support and min confidence.
8. the application process of intelligent medical big data service system according to claim 3, it is characterised in that:The medicine
Enterprise's cloud platform is used to obtain the medical data of the partial data at the private clound health data center of user, hospital's cloud platform, passes through
Cloud platform sends drug information to the analysis of data with rear line private clound health data center, hospital's cloud platform.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810386146.2A CN108573758B (en) | 2018-04-26 | 2018-04-26 | Intelligent medical big data service system and application method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810386146.2A CN108573758B (en) | 2018-04-26 | 2018-04-26 | Intelligent medical big data service system and application method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108573758A true CN108573758A (en) | 2018-09-25 |
CN108573758B CN108573758B (en) | 2021-10-08 |
Family
ID=63575338
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810386146.2A Active CN108573758B (en) | 2018-04-26 | 2018-04-26 | Intelligent medical big data service system and application method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108573758B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109088782A (en) * | 2018-11-01 | 2018-12-25 | 郑州云海信息技术有限公司 | The log collecting method and device of distributed system |
CN109411093A (en) * | 2018-10-16 | 2019-03-01 | 烟台翰宁信息科技有限公司 | A kind of intelligent medical treatment big data analysis processing method based on cloud computing |
CN109599155A (en) * | 2018-12-10 | 2019-04-09 | 上海新储集成电路有限公司 | A kind of intelligent service system and method applied to medical data center |
CN110164552A (en) * | 2019-05-27 | 2019-08-23 | 苏州嘉展科技有限公司 | A kind of system and method for extensive screening diabetic and tracking treatment effect |
CN110336706A (en) * | 2019-07-23 | 2019-10-15 | 中国工商银行股份有限公司 | Network message method for transmission processing and device |
CN110969534A (en) * | 2018-09-30 | 2020-04-07 | 上海笛乐护斯健康科技有限公司 | Intelligent health management and life safety insurance system and implementation method thereof |
CN111312409A (en) * | 2018-12-11 | 2020-06-19 | 康泰医学系统(秦皇岛)股份有限公司 | Medical data sharing system and method |
CN111785372A (en) * | 2020-05-14 | 2020-10-16 | 浙江知盛科技集团有限公司 | Collaborative filtering disease prediction system based on association rule and electronic equipment thereof |
CN112309564A (en) * | 2019-07-26 | 2021-02-02 | 深圳百诺明医说科技有限公司 | Artificial intelligence diagnostic system and intelligent robot |
CN117349030A (en) * | 2023-12-04 | 2024-01-05 | 深圳本贸科技股份有限公司 | Medical digital system, method and equipment based on cloud computing cluster |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105760705A (en) * | 2016-05-20 | 2016-07-13 | 陕西科技大学 | Medical diagnosis system based on big data |
CN105808946A (en) * | 2016-03-08 | 2016-07-27 | 贵州省邮电规划设计院有限公司 | Remote mobile medical system based on cloud computing |
CN106326623A (en) * | 2015-07-06 | 2017-01-11 | 薛海强 | Health information processing method and system |
CN106951691A (en) * | 2017-03-06 | 2017-07-14 | 宁波大学 | Mobile telemedicine management method based on cloud platform |
CN107633876A (en) * | 2017-10-31 | 2018-01-26 | 郑宇� | A kind of internet medical information processing system and method based on mobile platform |
-
2018
- 2018-04-26 CN CN201810386146.2A patent/CN108573758B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106326623A (en) * | 2015-07-06 | 2017-01-11 | 薛海强 | Health information processing method and system |
CN105808946A (en) * | 2016-03-08 | 2016-07-27 | 贵州省邮电规划设计院有限公司 | Remote mobile medical system based on cloud computing |
CN105760705A (en) * | 2016-05-20 | 2016-07-13 | 陕西科技大学 | Medical diagnosis system based on big data |
CN106951691A (en) * | 2017-03-06 | 2017-07-14 | 宁波大学 | Mobile telemedicine management method based on cloud platform |
CN107633876A (en) * | 2017-10-31 | 2018-01-26 | 郑宇� | A kind of internet medical information processing system and method based on mobile platform |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110969534A (en) * | 2018-09-30 | 2020-04-07 | 上海笛乐护斯健康科技有限公司 | Intelligent health management and life safety insurance system and implementation method thereof |
CN109411093A (en) * | 2018-10-16 | 2019-03-01 | 烟台翰宁信息科技有限公司 | A kind of intelligent medical treatment big data analysis processing method based on cloud computing |
CN109411093B (en) * | 2018-10-16 | 2022-03-18 | 国康中健(北京)健康科技有限公司 | Intelligent medical big data analysis processing method based on cloud computing |
CN109088782A (en) * | 2018-11-01 | 2018-12-25 | 郑州云海信息技术有限公司 | The log collecting method and device of distributed system |
CN109599155A (en) * | 2018-12-10 | 2019-04-09 | 上海新储集成电路有限公司 | A kind of intelligent service system and method applied to medical data center |
CN111312409A (en) * | 2018-12-11 | 2020-06-19 | 康泰医学系统(秦皇岛)股份有限公司 | Medical data sharing system and method |
CN111312409B (en) * | 2018-12-11 | 2023-11-10 | 康泰医学系统(秦皇岛)股份有限公司 | Medical data sharing system and method |
CN110164552A (en) * | 2019-05-27 | 2019-08-23 | 苏州嘉展科技有限公司 | A kind of system and method for extensive screening diabetic and tracking treatment effect |
CN110336706A (en) * | 2019-07-23 | 2019-10-15 | 中国工商银行股份有限公司 | Network message method for transmission processing and device |
CN110336706B (en) * | 2019-07-23 | 2022-09-13 | 中国工商银行股份有限公司 | Network message transmission processing method and device |
CN112309564A (en) * | 2019-07-26 | 2021-02-02 | 深圳百诺明医说科技有限公司 | Artificial intelligence diagnostic system and intelligent robot |
CN111785372A (en) * | 2020-05-14 | 2020-10-16 | 浙江知盛科技集团有限公司 | Collaborative filtering disease prediction system based on association rule and electronic equipment thereof |
CN117349030A (en) * | 2023-12-04 | 2024-01-05 | 深圳本贸科技股份有限公司 | Medical digital system, method and equipment based on cloud computing cluster |
CN117349030B (en) * | 2023-12-04 | 2024-06-14 | 深圳本贸科技股份有限公司 | Medical digital system, method and equipment based on cloud computing cluster |
Also Published As
Publication number | Publication date |
---|---|
CN108573758B (en) | 2021-10-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108573758A (en) | A kind of intelligent medical big data service system and application process | |
Kumar et al. | Big data analytics for healthcare industry: impact, applications, and tools | |
Eswari et al. | Predictive methodology for diabetic data analysis in big data | |
Chen et al. | A parallel patient treatment time prediction algorithm and its applications in hospital queuing-recommendation in a big data environment | |
US10373712B2 (en) | Aggregation, partitioning, and management of healthcare data for efficient storage and processing | |
Funkner et al. | Data-driven modeling of clinical pathways using electronic health records | |
CN108597605A (en) | A kind of life big data acquisition of personal health and analysis system | |
CN105335804A (en) | Community health service system | |
US20210035693A1 (en) | Methods, systems, and apparatuses for predicting the risk of hospitalization | |
Enriko et al. | Heart disease diagnosis system with k-nearest neighbors method using real clinical medical records | |
Khan et al. | Development of national health data warehouse for data mining. | |
Baskaran et al. | Predicting breast screening attendance using machine learning techniques | |
Vanathi et al. | A robust architectural framework for big data stream computing in personal healthcare real time analytics | |
CN110569372A (en) | construction method of heart disease big data knowledge graph system | |
Sampath et al. | Diabetic data analysis in healthcare using Hadoop architecture over big data | |
Ashoori et al. | A model to predict the sequential behavior of healthy blood donors using data mining | |
Tayebati et al. | Process mining case study approach: Extraction of unconventional event logs to improve performance in Hospital Information Systems (HIS) | |
Antonelli et al. | MeTA: Characterization of medical treatments at different abstraction levels | |
Carnevale et al. | How to enable clinical workflows to integrate big healthcare data | |
CN108091398B (en) | Patient grouping method and device | |
Egho et al. | Healthcare trajectory mining by combining multidimensional component and itemsets | |
CN114120177A (en) | User behavior guiding method and system based on intelligent health monitoring | |
Hsiao et al. | Data analysis of medical records in veterinary hospital using clustering method and association rule | |
CN111986815A (en) | Project combination mining method based on co-occurrence relation and related equipment | |
Bokhari et al. | A framework for clustering dental patients' records using unsupervised learning techniques |
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 |