CN106650206A - Prediction method of high blood pressure based on incremental neural network model and prediction system - Google Patents
Prediction method of high blood pressure based on incremental neural network model and prediction system Download PDFInfo
- Publication number
- CN106650206A CN106650206A CN201610861827.0A CN201610861827A CN106650206A CN 106650206 A CN106650206 A CN 106650206A CN 201610861827 A CN201610861827 A CN 201610861827A CN 106650206 A CN106650206 A CN 106650206A
- Authority
- CN
- China
- Prior art keywords
- hypertension
- neuron
- network model
- neural network
- data
- 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.)
- Pending
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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a prediction method of high blood pressure based on incremental neural network model. The method comprises the steps that a database of daily data of high blood pressures is built; a neural network model is trained; daily living data are collected and sent to a server, and are saved to a daily data sheet of a user; daily data are extracted from the daily data sheet of the user, and form an N dimensional vector, after the normalization is done, the data are input in a hypertensive pathological neural network model to conduct probabilistic prediction of the hazard level of high blood pressure; the value W of the hazard level of high blood pressure is judged whether or not being larger than or equal to 3 using a smart home high blood pressure care equipment; when a user receives a warning from an annunciator, the user goes to the hospital to check, the check results are sent back to the server through the smart home high blood pressure care equipment, the servers judges whether or not the check results are correct; an incremental calculation method is carried out when the check results are incorrect, and the neural network model is dynamically corrected. The method and system are accurate in prediction, and the neural network model is tailor made for every user.
Description
Technical field
The invention belongs to field of medical technology, more particularly to a kind of hypertension based on increment type neural network model is pre-
Survey method and forecasting system.
Background technology
It is current domestic it is each it is health management system arranged be respectively provided with hypertension prediction and evaluation, the prediction mode that it is used is data
Match somebody with somebody.Its principle is that then personal lifestyle data entry system is shown into ill probability by system matches fixed data.But due to people
The complexity of body and disease, unpredictability, in the form of expression of bio signal and information, Changing Pattern (Self-variation with
Change after medical intervention) on, it being detected and signal representation, the data of acquisition and the analysis of information, decision-making etc. are all multi-party
All there is extremely complex non-linear relationship in face.So the use of traditional Data Matching can only be the data examination of blindness, it is impossible to
Judge the logic association and variable between data and data, the codomain deviation for obtaining is big, causes the specificity ten of system prediction
It is point poor, domestic health management system arranged effectively Accurate Prediction cannot be carried out to the hypertension of individual so current.
Most of before this is all to use BP neural network model to hypertension prediction, but when new detection data generation
When, it is necessary to neural network model is trained again, and operation efficiency is extremely low.And after system user scale increases, server
Will be unable to complete training mission in time.
The content of the invention
The purpose of the present invention is that and overcomes the deficiencies in the prior art, there is provided one kind is based on increment type neural network model
Hypertension Forecasting Methodology and forecasting system, the present invention predicts a large amount of patient in hospital pathology numbers by neural network model training
According to, hypertension pathology and hypertension earlier life variations in detail, clinical symptoms, examination criteria value, people at highest risk's feature are found, this
Logic association and variable between several causes of disease, ultimately forms the hypertension pathology nerve to Hypertension probability Accurate Prediction
Network model, the present invention by gathering user's daily life data, the periodicity of active analysis its data, it is regular eventually through
Hypertension pathology Neural Network model predictive user's suffers from hypertension probability, and user is reminded in the way of visual effect immediately
Doctor, constantly corrects neural network model, to set for each when Neural Network model predictive is inaccurate by increasable algorithm
Standby user sets up the neural network model trained for the user, with the increase of use time, to set up to the customer volume
Body neural network model customized, accuracy rate is greatly improved.
To achieve these goals, the invention provides a kind of hypertension prediction side based on increment type neural network model
Method, comprises the steps：
Step (1), acquisition hospital hypertension pathogeny pathological data source and the daily monitoring data of patient, so as to set up hypertension
Daily data database；
Step (2), according to step (1) set up the daily data database of hypertension off-line manner to neutral net mould
Type is trained, to obtain the hypertension pathology neural network model for training；
Step (3), the daily life data of user are acquired by intelligent monitoring device, and will collection daily life
Live data is sent to server, and server preserves the daily life data of user into the daily data logger of user；
Step (4), from the daily data logger of user same day data are extracted, form n-dimensional vector, and n-dimensional vector is done
It is general hypertension degree of danger to be carried out in the hypertension pathology neural network model trained in input step (2) after normalized
Rate predicts, obtains hypertension probability results array P, and server is by corresponding hypertension degree of danger value W of maximum probability in array P
Send wired home hypertension care appliances to；
After step (5), hypertension degree of danger value W of wired home hypertension care appliances the reception server transmission, sentence
Whether broken height blood pressure degree of danger value W is more than or equal to 3, and if greater than equal to 3, then attention device is warned to remind user, if little
In 3, then attention device is not warned；
Step (6), when user receive attention device warn when, user voluntarily removes examination in hospital, and inspection result is passed through
Wired home hypertension care appliances send back server, and server judges whether inspection result is correct, if inspection result is wrong
By mistake, then illustrate that hypertension pathology Neural Network model predictive is inaccurate, if inspection result is correct, illustrate hypertension pathology god
The prediction of Jing network models is accurate；
Step (7), when inspection result mistake, from the daily data logger of user extract m days in record preserve to
In incremental data table, when the record quantity in incremental data table is more than h bars, increasable algorithm is performed, to hypertension pathology god
Jing network models carry out dynamic corrections；
Step (8), repeat step (3)～(7).
Further, the input layer of neural network model is n node, and hidden layer number is n*2+1, and output layer is 1
Node, k bars record is extracted from the daily data database table of hypertension and is trained, and per bar, record is a n-dimensional vector, is owned
Data are using front first Jing normalizeds so as to which numerical value is interval in [0,1], then perform following steps to neural network model
It is trained：
1) n-dimensional vector is input into neural network model, calculate all of weight vector in neural network model defeated to this
Enter the distance of n-dimensional vector, closest neuron is triumph neuron, and its computing formula is as follows：
Wherein：WkIt is the weight vector of triumph neuron, | | ... | | for Euclidean distance；
2) weight vector of triumph neuron and the neuron in triumph neuron field is adjusted, formula is as follows：
Wherein：WjT () is neuron；Wj(t+1) it is to adjust the weight vector after front and adjustment；J belongs to triumph neuron neck
Domain；α (t) is learning rate, and it is the function gradually successively decreased with the increase of iterations, and span is [0 1], through multiple
It is 0.71 that Optimal learning efficiency is chosen in experiment；DjIt is the distance of neuron j and triumph neuron；σ (t) is the letter for successively decreasing over time
Number；Each time all input n-dimensional vectors are all input in neural network model and are trained by iteration, when the iteration for reaching regulation
After number of times, neural network model training terminates.
Further, inspection result is sent back wired home hypertension care appliances the form of the object information of server
For：{ the hypertension degree of danger value of the actual judgement of doctor }, whether server judges inspection result after object information is received
Correctly.
Further, it is to the increasable algorithm that hypertension pathology neural network model carries out dynamic corrections：
In incremental data table per bar vector V { V1,V2,…,Vn, it is sent in neural network model learning function
Row study, learning procedure is as follows：
1) first little random number is assigned to each weight vector of output layer and does normalized, then using input mode vector V
Mean value Avg (V), be initialized as the weights of unique neuron in the 0th layer of neural network model, and be set to nerve of winning
Unit, calculates its quantization error QE；
2) 2 × 2 structures SOM are expanded out from the 0th layer of neuron, and its level identities Layer is set into 1；
3) for each 2 × 2 structure SOM subnet expanded out in Layer layers, the power of this 4 neurons is initialized
Value；The input vector set Ci of i-th neuron is set into sky, main label is set to NULL, the main label ratio r of neuron ii
It is set to 0；The abnormity early warning data vector V of new SOM inherits the triumph input vector set VX of his father's neuron；
4) a vector VX is selected from VXiDo following judgement：
If VXiFor the data of not tape label, then its Euclidean distance with each neuron is calculated, chosen distance is most short
Neuron is used as triumph neuron；
If VXiFor the data of tape label, then main label and VX are selectediLabel is identical and riThe maximum neuron of value is made
For triumph neuron, the triumph neuron main label is updated；
If can not find main label and VXiLabel identical neuron, then find and VXiClosest neuron i makees
For triumph neuron；
5) weights of neuron in triumph neuron and its neighborhood are adjusted, update the vectorial set W=W ∪ that win
{VXi, calculate main label, the main label ratio r of triumph neuroniWith comentropy EiIf. not up to predetermined frequency of training,
Go to step 4)；
6) quantization error QE of each neuron in the neural network model after calculating is adjustedi, neuronal messages entropy Ei
With the average quantization error MQE of subnet, formula is as follows：
Wherein：WiFor the weight vector of neuron i, CiFor the set that all input vectors for being mapped to neuron i are constituted；
Wherein：niRepresent fall on neuron label for i number of samples, m represents label data on neuron
Sum, T represents the sample label species set on neuron；
Then judge：
If MQE>QE × threshold value q of father node, wherein q=0.71 then inserts a line neuron in the SOM, turns step
It is rapid 4)；
If Ei>The E of father nodei× threshold value p, wherein p=0.42, then grow one layer of new subnet from the neuron, will
The subnet for newly growing increases in the subnet queue of Layer+1 layers；
If new neuron is not inserted into SOM does not grow new subnet yet, illustrate that the subnet training is completed；
7) for all 2 × 2 structures SOM of the Layer+1 layers newly expanded out, iteration operating procedure 3)～5) to it again
It is trained, until neural network model no longer produces new neuron and new layering, whole training terminates.
Further, in step (4), hypertension probability results array P for obtaining is one 6 dimension variable, its 6 dimension variable
Respectively 6 hypertension degree of danger probability, the corresponding hypertension of maximum probability in 6 hypertension degree of danger probability is dangerous
Degree value W sends wired home hypertension care appliances to；In step (5), represent when hypertension degree of danger value W=1
Ideal, represents normal during W=2, inferior health is represented during W=3, represents dangerous during W=4, represents abnormally dangerous during W=5, during W=6
Represent ill.
Further, if user includes by other means health check-up and checks oneself, learn and oneself have suffered from hypertension, and intelligence
The attention device of family's hypertension care appliances is not warned, then it represents that wired home hypertension care appliances judge inaccurate, this
When execution step (6)～(7), wired home hypertension care appliances are sent to object information on server.
Present invention also offers a kind of forecasting system of the hypertension Forecasting Methodology, including intelligent monitoring device, intelligence
Device data acquisition device, server and wired home hypertension care appliances, the intelligent monitoring device and the smart machine
Data acquisition unit is connected, and the smart machine data acquisition unit is communicated by communication device one with the server network, institute
State wired home hypertension care appliances to communicate with the server network by communication device two.
Further, it is provided with attention device on the wired home hypertension care appliances.
Further, the intelligent monitoring device include Intelligent worn device, Intelligent water cup, Intelligent weight claim, intelligent horse
Bucket and Intelligent light sensing equipment.
Beneficial effects of the present invention：
1st, the present invention is by the neural network model training a large amount of patient in hospital pathological datas of prediction, find hypertension pathology and
Hypertension earlier life variations in detail, clinical symptoms, examination criteria value, people at highest risk's feature, the logic between this several causes of disease
Association and variable, ultimately form the hypertension pathology neural network model to Hypertension probability Accurate Prediction, and the present invention is logical
Cross collection user's daily life data, periodicity, the regularity of active analysis its data is eventually through hypertension pathology nerve net
Network model prediction user's suffers from hypertension degree of danger probability, and user's instant hospitalizing and prevention are reminded in the way of visual effect.
2nd, neural network model is constantly corrected by increasable algorithm when Neural Network model predictive is inaccurate, to be directed to
Each equipment user sets up the neural network model trained for the user, with the increase of use time, to set up to this
The neural network model that user makes to measure, accuracy rate is greatly improved.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is the flow chart of the embodiment of the present invention.
Specific embodiment
Invention is further illustrated below in conjunction with the accompanying drawings, but is not limited to the scope of the present invention.
Embodiment
As shown in figure 1, a kind of hypertension Forecasting Methodology based on increment type neural network model that the present invention is provided, including
Following steps：
Step (1), acquisition hospital hypertension pathogeny pathological data source and the daily monitoring data of patient, so as to set up hypertension
Daily data database；
Wherein daily monitoring data is 12 item datas, and its 12 item data is the age, and sex, heart rate, body fat, smoking capacity is (every
Day), the amount of drinking (daily), blood pressure, body weight, the length of one's sleep and quality, 12 item datas such as travel distance (daily) are of the invention with 12
Item data sets up 12 dimensional vectors；
Step (2), according to step (1) set up the daily data database of hypertension off-line manner to neutral net mould
Type is trained, to obtain the hypertension pathology neural network model for training；
Step (3), the daily life data of user are acquired by intelligent monitoring device, and will collection daily life
Live data is sent to server, and server preserves the daily life data of user into the daily data logger of user；
Step (4), from the daily data logger of user same day data are extracted, form n-dimensional vector, and n-dimensional vector is done
It is general hypertension degree of danger to be carried out in the hypertension pathology neural network model trained in input step (2) after normalized
Rate predicts, obtains hypertension probability results array P, and server is by corresponding hypertension degree of danger value W of maximum probability in array P
Send wired home hypertension care appliances to；
Wherein in step (4), hypertension probability results array P for obtaining is one 6 dimension variable, its 6 dimension variable difference
For 6 hypertension degree of danger probability, the corresponding hypertension degree of danger of maximum probability in 6 hypertension degree of danger probability
Value W sends wired home hypertension care appliances to.
After step (5), hypertension degree of danger value W of wired home hypertension care appliances the reception server transmission, sentence
Whether broken height blood pressure degree of danger value W is more than or equal to 3, and if greater than equal to 3, then attention device is warned to remind user, if little
In 3, then attention device is not warned；
Wherein in step (5), represent preferable when hypertension degree of danger value W=1, normal, W=3 is represented during W=2
When represent inferior health, represent dangerous during W=4, represent abnormally dangerous during W=5, represent ill during W=6.It is i.e. of the invention by high blood
Pressure degree of danger is divided into 6 grades, more accurate and visual.
Step (6), when user receive attention device warn when, user voluntarily removes examination in hospital, and inspection result is passed through
Wired home hypertension care appliances send back server, and server judges whether inspection result is correct, if inspection result is wrong
By mistake, then illustrate that hypertension pathology Neural Network model predictive is inaccurate, if inspection result is correct, illustrate hypertension pathology god
The prediction of Jing network models is accurate；
The form that wherein inspection result is sent back wired home hypertension care appliances the object information of server is：
{ the hypertension degree of danger value of the actual judgement of doctor }, just whether server judge inspection result after object information is received
Really.
Step (7), when inspection result mistake, from the daily data logger of user extract 7 days in record preserve to
In incremental data table, when the record quantity in incremental data table is more than 100, increasable algorithm is performed, to hypertension pathology
Neural network model carries out dynamic corrections；
Step (8), repeat step (3)～(7).
The input layer of the neural network model of the present invention is 12 nodes, and hidden layer number is 25, and output layer is 6 nodes
(i.e. 6 hypertension degree of danger probability), extracts 400000 records and is trained from the daily data database table of hypertension,
Per bar, record is 12 dimensional vectors, and all data are using front first Jing normalizeds so as to which numerical value is interval in [0,1], so
Perform following steps afterwards to be trained neural network model：1) 12 dimensional vectors are input into neural network model, god is calculated
All of weight vector is arrived in Jing network models
The distance of the dimensional vector of input 12, closest neuron is triumph neuron, and its computing formula is as follows：
Wherein：WkIt is the weight vector of triumph neuron, | | ... | | for Euclidean distance；
2) weight vector of triumph neuron and the neuron in triumph neuron field is adjusted, formula is as follows：
Wherein：WjT () is neuron；Wj(t+1) it is to adjust the weight vector after front and adjustment；J belongs to triumph neuron neck
Domain；α (t) is learning rate, and it is the function gradually successively decreased with the increase of iterations, and span is [0 1], through multiple
It is 0.71 that Optimal learning efficiency is chosen in experiment；DjIt is the distance of neuron j and triumph neuron；σ (t) is the letter for successively decreasing over time
Number；Each time all input n-dimensional vectors are all input in neural network model and are trained by iteration, when the iteration for reaching regulation
After number of times, neural network model training terminates.
The present invention be to the increasable algorithm that hypertension pathology neural network model carries out dynamic corrections：
In incremental data table per bar vector V { V1,V2,…,Vn, it is sent in neural network model learning function
Row study, learning procedure is as follows：
1) first little random number is assigned to each weight vector of output layer and does normalized, then using input mode vector V
Mean value Avg (V), be initialized as the weights of unique neuron in the 0th layer of neural network model, and be set to nerve of winning
Unit, calculates its quantization error QE；
2) 2 × 2 structures SOM are expanded out from the 0th layer of neuron, and its level identities Layer is set into 1；
3) for each 2 × 2 structure SOM subnet expanded out in Layer layers, the power of this 4 neurons is initialized
Value；The input vector set Ci of i-th neuron is set into sky, main label is set to NULL, the main label ratio r of neuron ii
It is set to 0；The abnormity early warning data vector V of new SOM inherits the triumph input vector set VX of his father's neuron；
4) a vector VX is selected from VXiDo following judgement：
If VXiFor the data of not tape label, then its Euclidean distance with each neuron is calculated, chosen distance is most short
Neuron is used as triumph neuron；
If VXiFor the data of tape label, then main label and VX are selectediLabel is identical and riThe maximum neuron of value is made
For triumph neuron, the triumph neuron main label is updated；
If can not find main label and VXiLabel identical neuron, then find and VXiClosest neuron i makees
For triumph neuron；
5) weights of neuron in triumph neuron and its neighborhood are adjusted, update the vectorial set W=W ∪ that win
{VXi, calculate main label, the main label ratio r of triumph neuroniWith comentropy EiIf. not up to predetermined frequency of training,
Go to step 4)；
6) quantization error QE of each neuron in the neural network model after calculating is adjustedi, neuronal messages entropy Ei
With the average quantization error MQE of subnet, formula is as follows：
Wherein：WiFor the weight vector of neuron i, CiFor the set that all input vectors for being mapped to neuron i are constituted；
Wherein：niRepresent fall on neuron label for i number of samples, m represents label data on neuron
Sum, T represents the sample label species set on neuron；
Then judge：
If MQE>QE × threshold value q of father node, wherein q=0.71 then inserts a line neuron in the SOM, turns step
It is rapid 4)；
If Ei>The E of father nodei× threshold value p, wherein p=0.42, then grow one layer of new subnet from the neuron, will
The subnet for newly growing increases in the subnet queue of Layer+1 layers；
If new neuron is not inserted into SOM does not grow new subnet yet, illustrate that the subnet training is completed；
7) for all 2 × 2 structures SOM of the Layer+1 layers newly expanded out, iteration operating procedure 3)～5) to it again
It is trained, until neural network model no longer produces new neuron and new layering, whole training terminates.
If user includes by other means health check-up and checks oneself, learn and oneself have suffered from hypertension, and the high blood of wired home
The attention device of pressure care appliances is not warned, then it represents that wired home hypertension care appliances judge inaccurate, now perform step
Suddenly (6)～(7), wired home hypertension care appliances are sent to object information on server.
Present invention also offers a kind of forecasting system of the hypertension Forecasting Methodology, including intelligent monitoring device, intelligence
Device data acquisition device, server and wired home hypertension care appliances, the intelligent monitoring device and the smart machine
Data acquisition unit is connected, and the smart machine data acquisition unit is communicated by communication device one with the server network, institute
State wired home hypertension care appliances to communicate with the server network by communication device two.
Attention device is provided with the wired home hypertension care appliances of the present invention.
The intelligent monitoring device of the present invention claims including Intelligent worn device, Intelligent water cup, Intelligent weight, intelligent closestool
With Intelligent light sensing equipment etc..
The present invention finds hypertension pathology with height by a large amount of patient in hospital pathological datas of neural network model training prediction
Blood pressure earlier life variations in detail, clinical symptoms, examination criteria value, people at highest risk's feature, the logic between this several causes of disease is closed
Connection and variable, ultimately form the hypertension pathology neural network model to Hypertension probability Accurate Prediction, and the present invention passes through
Collection user's daily life data, periodicity, the regularity of active analysis its data is eventually through hypertension pathology neutral net
Model prediction user's suffers from hypertension probability, and user's instant hospitalizing and prevention are reminded in the way of visual effect.
All data of the present invention are preserved into server, can significantly save calculating cost, and hardware configuration is low, so as to sell
Valency is also low.
The present invention carries communication device one and communication device two, can automatically connect internet by wifi, and protects for a long time
Hold online.Various intelligent monitoring devices can easily access present device by modes such as network or bluetooths, set in acquisition
The daily life data of the monitoring of intelligent monitoring device, therefore the data that present device is obtained can be automatically uploaded after standby mandate
It is real-time, accurate, polynary.
Because everyone physical trait is different, the data characteristics shown during hypertension incidence also can be different.Cause
This is not conventional high by the method accuracy rate of neural network prediction hypertension.The present invention sets up training for each equipment user
Go out the neural network model for the user, in operation after a period of time, by producing to measure neutral net is made to the user
Forecast model, accuracy rate is greatly improved.
When neural network model is judged by accident, error message will be feedbacked to server by wired home hypertension care appliances,
For user's dynamic corrections neural network model, when similar characteristics data occurs in the next user, will not judge by accident again.Cause
This, with the increase of use time, the judgement of the wired home hypertension care appliances of the present invention will be more and more accurate.
General principle, principal character and the advantages of the present invention of the present invention has been shown and described above.The technology of the industry
Personnel it should be appreciated that the present invention is not restricted to the described embodiments, the simply explanation described in above-described embodiment and specification this
The principle of invention, of the invention without departing from the spirit and scope of the present invention also to have various changes and modifications, these changes
Change and improvement is both fallen within scope of the claimed invention.The claimed scope of the invention by appending claims and its
Equivalent is defined.
Claims (9)
1. a kind of hypertension Forecasting Methodology based on increment type neural network model, it is characterised in that comprise the steps：
Step (1), obtain hospital hypertension and cure the disease etiology and pathology data source and the daily monitoring data of patient, so as to set up hypertension
Daily data database；
Step (2), the daily data database of hypertension set up according to step (1) are entered off-line manner to neural network model
Row training, to obtain the hypertension pathology neural network model for training；
Step (3), the daily life data of user are acquired by intelligent monitoring device, and will collection daily life number
According to sending to server, server preserves the daily life data of user into the daily data logger of user；
Step (4), from the daily data logger of user same day data are extracted, form n-dimensional vector, and normalizing is done to n-dimensional vector
It is pre- hypertension degree of danger probability to be carried out in the hypertension pathology neural network model trained in input step (2) after change process
Survey, obtain hypertension probability results array P, server transmits corresponding hypertension degree of danger value W of maximum probability in array P
Give wired home hypertension care appliances；
After step (5), hypertension degree of danger value W of wired home hypertension care appliances the reception server transmission, judge high
Whether blood pressure degree of danger value W is more than or equal to 3, and if greater than equal to 3, then attention device is warned to remind user, if less than 3,
Then attention device is not warned；
Step (6), when user receives attention device and warns, user voluntarily removes examination in hospital, and by inspection result by intelligence
Family's hypertension care appliances send back server, and server judges whether inspection result is correct, if inspection result mistake,
Illustrate that hypertension pathology Neural Network model predictive is inaccurate, if inspection result is correct, illustrate hypertension pathology nerve net
Network model prediction is accurate；
Step (7), when inspection result mistake, from the daily data logger of user extract m days in record preserve to increment
In tables of data, when the record quantity in incremental data table is more than h bars, increasable algorithm is performed, to hypertension pathology nerve net
Network model carries out dynamic corrections；
Step (8), repeat step (3)～(7).
2. a kind of hypertension Forecasting Methodology based on increment type neural network model according to claim 1, its feature exists
In the input layer of neural network model is n node, and hidden layer number is n*2+1, and output layer is 1 node, from hypertension day
Extract k bars record in regular data database table to be trained, per bar, record is a n-dimensional vector, and all data are using front elder generation
Jing normalizeds so as to which numerical value is interval in [0,1], then perform following steps and neural network model are trained：
1) n-dimensional vector is input into neural network model, calculate all of weight vector in neural network model and tie up to input n
The distance of vector, closest neuron is triumph neuron, and its computing formula is as follows：
Wherein：WkIt is the weight vector of triumph neuron, | | ... | | for Euclidean distance；
2) weight vector of triumph neuron and the neuron in triumph neuron field is adjusted, formula is as follows：
Wherein：WjT () is neuron；Wj(t+1) it is to adjust the weight vector after front and adjustment；J belongs to triumph neuron field；α
T () is learning rate, it is the function gradually successively decreased with the increase of iterations, and span is [0 1], through many experiments
It is 0.71 to choose Optimal learning efficiency；DjIt is the distance of neuron j and triumph neuron；σ (t) is the function for successively decreasing over time；
Each time all input n-dimensional vectors are all input in neural network model and are trained by iteration, when the iteration time for reaching regulation
After number, neural network model training terminates.
3. a kind of hypertension Forecasting Methodology based on increment type neural network model according to claim 1, its feature exists
In wired home hypertension care appliances inspection result are sent back the form of the object information of server and are：{ doctor is actual to be sentenced
Disconnected hypertension degree of danger value }, server judges whether inspection result is correct after object information is received.
4. a kind of hypertension Forecasting Methodology based on increment type neural network model according to claim 1, its feature exists
In being to the increasable algorithm that hypertension pathology neural network model carries out dynamic corrections：
In incremental data table per bar vector V { V1,V2,…,Vn, it is sent in neural network model learning function and is learned
Practise, learning procedure is as follows：
1) first little random number is assigned to each weight vector of output layer and does normalized, then using the flat of input mode vector V
Average Avg (V), is initialized as the weights of unique neuron in the 0th layer of neural network model, and is set to triumph neuron, meter
Calculate its quantization error QE；
2) 2 × 2 structures SOM are expanded out from the 0th layer of neuron, and its level identities Layer is set into 1；
3) for each 2 × 2 structure SOM subnet expanded out in Layer layers, the weights of this 4 neurons are initialized；Will
The input vector set Ci of i-th neuron is set to sky, and main label is set to NULL, the main label ratio r of neuron iiIt is set to
0；The abnormity early warning data vector V of new SOM inherits the triumph input vector set VX of his father's neuron；
4) a vector VX is selected from VXiDo following judgement：
If VXiFor the data of not tape label, then its Euclidean distance with each neuron, the most short nerve of chosen distance are calculated
Unit is used as triumph neuron；
If VXiFor the data of tape label, then main label and VX are selectediLabel is identical and riThe maximum neuron of value is used as obtaining
Victory neuron, updates the triumph neuron main label；
If can not find main label and VXiLabel identical neuron, then find and VXiClosest neuron i is used as obtaining
Victory neuron；
5) weights of neuron in triumph neuron and its neighborhood are adjusted, update the vectorial set W=W ∪ { VX that wini,
Calculate main label, the main label ratio r of triumph neuroniWith comentropy EiIf. not up to predetermined frequency of training, go to step
4)；
6) quantization error QE of each neuron in the neural network model after calculating is adjustedi, neuronal messages entropy EiAnd son
The average quantization error MQE of net, formula is as follows：
Wherein：WiFor the weight vector of neuron i, CiFor the set that all input vectors for being mapped to neuron i are constituted；
Wherein：niRepresent fall on neuron label for i number of samples, m represents the total of label data on neuron
Number, T represents the sample label species set on neuron；
Then judge：
If MQE>QE × threshold value q of father node, wherein q=0.71 then inserts a line neuron in the SOM, goes to step 4)；
If Ei>The E of father nodei× threshold value p, wherein p=0.42, then grow one layer of new subnet from the neuron, will be new
The subnet for growing increases in the subnet queue of Layer+1 layers；
If new neuron is not inserted into SOM does not grow new subnet yet, illustrate that the subnet training is completed；
7) for all 2 × 2 structures SOM of the Layer+1 layers newly expanded out, iteration operating procedure 3)～5) it is re-started
Training, until neural network model no longer produces new neuron and new layering, whole training terminates.
5. a kind of hypertension Forecasting Methodology based on increment type neural network model according to claim 1, its feature exists
In in step (4), hypertension probability results array P for obtaining is one 6 dimension variable, and its 6 dimension variable is respectively 6 high blood
Pressure degree of danger probability, sends corresponding hypertension degree of danger value W of maximum probability in 6 hypertension degree of danger probability to
Wired home hypertension care appliances；In step (5), represent preferable when hypertension degree of danger value W=1, table during W=2
Show normal, inferior health is represented during W=3, represent dangerous during W=4, represent abnormally dangerous during W=5, represent ill during W=6.
6. a kind of hypertension Forecasting Methodology based on increment type neural network model according to claim 1, its feature exists
In, if user includes by other means health check-up and checks oneself, learn and oneself have suffered from hypertension, and wired home hypertension is nursed
The attention device of equipment is not warned, then it represents that wired home hypertension care appliances judge inaccurate, now execution step (6)～
(7), wired home hypertension care appliances are sent to object information on server.
7. the forecasting system of hypertension Forecasting Methodology described in a kind of employing claim 1～7, it is characterised in that including intelligent prison
Control equipment, smart machine data acquisition unit, server and wired home hypertension care appliances, the intelligent monitoring device and institute
State smart machine data acquisition unit to be connected, the smart machine data acquisition unit is by communication device one and the server net
Network is communicated, and the wired home hypertension care appliances are communicated by communication device two with the server network.
8. the forecasting system of hypertension Forecasting Methodology according to claim 8, it is characterised in that the wired home hypertension
Attention device is provided with care appliances.
9. the forecasting system of hypertension Forecasting Methodology according to claim 8, it is characterised in that the intelligent monitoring device bag
Include Intelligent worn device, Intelligent water cup, Intelligent weight claim, intelligent closestool and Intelligent light sensing equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610861827.0A CN106650206A (en) | 2016-09-28 | 2016-09-28 | Prediction method of high blood pressure based on incremental neural network model and prediction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610861827.0A CN106650206A (en) | 2016-09-28 | 2016-09-28 | Prediction method of high blood pressure based on incremental neural network model and prediction system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106650206A true CN106650206A (en) | 2017-05-10 |
Family
ID=58854757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610861827.0A Pending CN106650206A (en) | 2016-09-28 | 2016-09-28 | Prediction method of high blood pressure based on incremental neural network model and prediction system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106650206A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109009061A (en) * | 2018-06-20 | 2018-12-18 | 博动医学影像科技（上海）有限公司 | The calculation method and device for obtaining flow characteristic value are corrected based on blood pressure |
| CN109935327A (en) * | 2019-03-15 | 2019-06-25 | 南方医科大学顺德医院(佛山市顺德区第一人民医院) | Hypertensive patient's cardiovascular risk grading appraisal procedure based on intelligence decision support system |
| CN110932274A (en) * | 2019-12-18 | 2020-03-27 | 辽宁工业大学 | Power system measurement and load parameter analysis and identification method |
| CN113712513A (en) * | 2021-09-08 | 2021-11-30 | 北京百度网讯科技有限公司 | Physiological detection method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1477581A (en) * | 2003-07-01 | 2004-02-25 | �Ϻ���ͨ��ѧ | Predictive modelling method application to computer-aided medical diagnosis |
| CN102647292A (en) * | 2012-03-20 | 2012-08-22 | 北京大学 | Intrusion detecting method based on semi-supervised neural network |
| US20140351188A1 (en) * | 2013-05-22 | 2014-11-27 | Exablade Corporation | Prefetch system and method |
| CN105023220A (en) * | 2015-08-14 | 2015-11-04 | 北京大学第三医院 | Health guidance system for polycystic ovarian syndrome |
| CN105118010A (en) * | 2015-09-30 | 2015-12-02 | 成都信汇聚源科技有限公司 | Chronic disease management method with functions of real-time data processing and real-time information sharing and life style intervention information |
| CN105912829A (en) * | 2015-11-12 | 2016-08-31 | 上海童舒房科技股份有限公司 | Bad habit correcting system based on cloud |
-
2016
- 2016-09-28 CN CN201610861827.0A patent/CN106650206A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1477581A (en) * | 2003-07-01 | 2004-02-25 | �Ϻ���ͨ��ѧ | Predictive modelling method application to computer-aided medical diagnosis |
| CN102647292A (en) * | 2012-03-20 | 2012-08-22 | 北京大学 | Intrusion detecting method based on semi-supervised neural network |
| US20140351188A1 (en) * | 2013-05-22 | 2014-11-27 | Exablade Corporation | Prefetch system and method |
| CN105023220A (en) * | 2015-08-14 | 2015-11-04 | 北京大学第三医院 | Health guidance system for polycystic ovarian syndrome |
| CN105118010A (en) * | 2015-09-30 | 2015-12-02 | 成都信汇聚源科技有限公司 | Chronic disease management method with functions of real-time data processing and real-time information sharing and life style intervention information |
| CN105912829A (en) * | 2015-11-12 | 2016-08-31 | 上海童舒房科技股份有限公司 | Bad habit correcting system based on cloud |
Non-Patent Citations (1)
| Title |
|---|
| MEHMED KANTARDZIC: "《数据挖掘：概念、模型、方法和算法》", 31 January 2013, 清华大学出版社 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109009061A (en) * | 2018-06-20 | 2018-12-18 | 博动医学影像科技（上海）有限公司 | The calculation method and device for obtaining flow characteristic value are corrected based on blood pressure |
| CN109009061B (en) * | 2018-06-20 | 2021-11-19 | 博动医学影像科技（上海）有限公司 | Calculation method and device for obtaining blood flow characteristic value based on blood pressure correction |
| CN109935327A (en) * | 2019-03-15 | 2019-06-25 | 南方医科大学顺德医院(佛山市顺德区第一人民医院) | Hypertensive patient's cardiovascular risk grading appraisal procedure based on intelligence decision support system |
| CN109935327B (en) * | 2019-03-15 | 2023-08-08 | 南方医科大学顺德医院(佛山市顺德区第一人民医院) | Cardiovascular risk layering evaluation method for hypertension patient based on intelligent decision support |
| CN110932274A (en) * | 2019-12-18 | 2020-03-27 | 辽宁工业大学 | Power system measurement and load parameter analysis and identification method |
| CN113712513A (en) * | 2021-09-08 | 2021-11-30 | 北京百度网讯科技有限公司 | Physiological detection method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106407699A (en) | Coronary heart disease prediction method and prediction system based on incremental neural network model | |
| CN108366763A (en) | Method and system for assessing the state of mind | |
| CN106446560A (en) | Hyperlipidemia prediction method and prediction system based on incremental neural network model | |
| CN106446552A (en) | Prediction method and prediction system for sleep disorder based on incremental neural network model | |
| CN106355035A (en) | Pneumonia prediction method and prediction system based on incremental neural network model | |
| CN106650206A (en) | Prediction method of high blood pressure based on incremental neural network model and prediction system | |
| CN113241196B (en) | Remote medical treatment and grading monitoring system based on cloud-terminal cooperation | |
| CN106384013A (en) | Incremental neural network model-based type-II diabetes prediction method and prediction system | |
| CN106250712A (en) | A kind of ureteral calculus Forecasting Methodology based on increment type neural network model and prognoses system | |
| CN106355034A (en) | Sub-health prediction method and prediction system based on incremental neural network model | |
| CN106295238A (en) | A kind of hypertensive nephropathy Forecasting Methodology based on increment type neural network model and prognoses system | |
| CN106384008A (en) | Incremental neural network model-based allergic rhinitis prediction method and prediction system | |
| CN106384012A (en) | Incremental neural network model-based allergic dermatitis prediction method and prediction system | |
| CN106407693A (en) | Hepatitis B prediction method and prediction system based on incremental neural network model | |
| CN106384005A (en) | Incremental neural network model-based depression prediction method and prediction system | |
| CN106250715A (en) | A kind of chronic pharyngolaryngitis Forecasting Methodology based on increment type neural network model and prognoses system | |
| CN106407694A (en) | Neurasthenia prediction method and prediction system based on incremental neural network model | |
| CN106202986A (en) | A kind of tonsillitis Forecasting Methodology based on increment type neural network model and prognoses system | |
| CN106384009A (en) | Incremental neural network model-based HIV prediction method and prediction system | |
| CN106446561A (en) | Incremental neural network model based urticaria prediction method and system | |
| CN106295239A (en) | A kind of fatty liver Forecasting Methodology based on increment type neural network model and prognoses system | |
| CN106407697A (en) | Chronic fatigue syndrome prediction method and prediction system based on incremental neural network model | |
| CN106407695A (en) | Anxiety disorder prediction method and prediction system based on incremental neural network model | |
| CN106339605A (en) | Colonitis prediction method and colonitis prediction system based on incremental nerve network model | |
| CN106446563A (en) | Incremental neural network model based constipation prediction method and system |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170510 |
|
| WD01 | Invention patent application deemed withdrawn after publication |