CN106682385A - Health information interaction platform - Google Patents
Health information interaction platform Download PDFInfo
- Publication number
- CN106682385A CN106682385A CN201610881071.6A CN201610881071A CN106682385A CN 106682385 A CN106682385 A CN 106682385A CN 201610881071 A CN201610881071 A CN 201610881071A CN 106682385 A CN106682385 A CN 106682385A
- Authority
- CN
- China
- Prior art keywords
- video
- rehabilitation
- data
- module
- 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
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/20—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for electronic clinical trials or questionnaires
-
- 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
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a health information interaction platform. The health information interaction platform comprises a cloud server and a plurality of client sides in network connection with the cloud server. Every client side comprises a first processor, an input component, a display screen and a first storage, wherein the input component, the display screen and the first storage are connected with the first processor. The cloud server comprises a second processor, a second storage, a database and a verification component, wherein the second storage is used for storing multiple exercise rehabilitation videos, the database is used for updating user questionnaires and answers in real time, the verification component is used for verifying identity information and questionnaire answers inputted by users, and the second storage, the database and the verification component are connected with the second processor. Every first processor is provided with a judgment module for judging information inputted by the input component, an AHP model is arranged in the judgment module, and the judgment module is used for extracting assessment factors according to the questionnaire answers inputted by the users based on the questionnaires.
Description
Technical field
The present invention relates to health and fitness information interaction field, and in particular to a kind of health and fitness information interaction platform.
Background technology
China human mortality aging in recent years accelerates, and as the rhythm of life of the fast-developing people in epoch constantly adds
Hurry up, life stress is continuously increased, the thing followed is the growth of the various chronic disease incidences of disease, this physical and mental health to people,
The stable and development of society brings serious negative effect.
In order to quickly and easily obtaining the health and fitness information situation of itself and the health and fitness information situation being directed in day
Often nursed one's health in life, it is now newly developed go out some for health control health and fitness information interaction platform.Because will intelligence
Ground judges the health status of different people and also needs to recognize the New Type of Diseases for judging to continuously emerge, in these information exchange platforms
It is both provided with the machine learning system for making information exchange platform more intelligent.It is well known that machine learning system needs to make
With could constantly Optimization Learning based on certain data volume.And existing health and fitness information interaction platform is all too dependent on machine
Device learning system so that existing health and fitness information interaction platform be required for typing it is certain could start after rehabilitation data compare
Accurately run.In order that existing health and fitness information interaction platform there can be more accurate feedback effects, then coming into operation
It is front must first to this, platform input be a certain amount of treats that rehabilitation data carry out model training.This not only wastes health and fitness information interaction platform
Overall debug time, and troublesome poeration.More disadvantageously the precision of this platform extremely relies on the entirety for treating rehabilitation data
Quality, if early stage be used for training treat rehabilitation data and improper, health and fitness information interaction platform can be made to make in concrete input
Used time still precision is relatively low, in other words, the pole for using the sample for treating rehabilitation data by using as training of the platform
It is big to limit.
Therefore, be badly in need of developing now it is a kind of can immediately using and do not rely on the health and fitness information for treating rehabilitation data sample
Interaction platform.
The content of the invention
The invention is intended to provide one kind can directly use, do not rely on and treat that the health and fitness information interaction of rehabilitation data sample is flat
Platform.
Health and fitness information interaction platform in this programme, including cloud server and with cloud server by network connection
Multiple client；The client includes first processor and the input module, the display screen that are connected with first processor respectively
And first memory；The cloud server includes second processing device, and being stored with of being connected with second processing device respectively is more
The second memory of individual rehabilitation video combination, the database for real-time update user investigation questionnaire and its answer, for
The checking assembly that the identity information of family input and questionnaire answer are verified；Grade reference is provided with the first memory
Table；The relation that the grade reference table agreement is numbered with video；Rehabilitation of the video numbering with storage in the database
Video combination is corresponded.
Name Resolution：
Rehabilitation video：Including video plus audio frequency, picture and text it is various combine and based on video for carrying out health
The multiple video file for instructing.Rehabilitation video combination is multiple rehabilitation videos.
User investigation questionnaire and answer：For carrying out the questionnaire of data acquisition to user, and user fills in investigation
Answer after questionnaire.Wherein, rehabilitation data are treated in the answer of the questionnaire that user fills in, and treat that rehabilitation data can be understood as using
Family is obtaining all input datas before the video for pushing.
Grade reference table：Rehabilitation data are treated according to user input, various factors is extracted, by these factors according to degree not
It is classified together, every grade of factor and the video numbering of formation carry out one-to-one related table.
Video is numbered：Each rehabilitation video is used for carrying out uniquely identified numbering.
Operation principle and beneficial effect：
User is answered by client for the questionnaire in incoming client, the questionnaire that user answers
Answer be input in first processor by input module as treating rehabilitation data.First processor is deposited according to being arranged on first
Grade reference table in reservoir, will directly treat that rehabilitation data are carried out with each factor of grade reference table corresponding, obtain respectively with respectively
The corresponding video numbering of factor.Client sends the video numbering obtained from grade reference table to cloud server, and second
Processor is numbered according to video and will combine transmission extremely with the one-to-one rehabilitation diplopia frequency of these videos numberings in second memory
The rehabilitation video combination is passed to display screen and is shown by first processor, first processor.
Health and fitness information interaction platform in this programme can not collect it is a certain amount of whne rehabilitation data when, i.e., in database
Accumulation when rehabilitation data are zero, only just can directly produce rehabilitation video group with the grade reference table in first memory
It is closing as a result, it is possible to immediately using and do not rely on the health and fitness information interaction platform for treating rehabilitation data.
Further, the judge module for judging input module input information is provided with the first processor；It is described to sentence
AHP models are provided with disconnected module；AHP models employ the relative weighting that two-phase method solves each assessment factor of questionnaire；
First processor calculates the first association constant of user according to the weighted value of each assessment factor, and first processor will be arranged on the
Explanation data transfer corresponding with the first association constant is to display screen in one processor.
Judge module in first processor treats that rehabilitation data AHP models carry out weight calculation and draw one to input
First association constant, first processor is set in advance in the such as healthy or unsound theory in first processor by contrast
The explanation data transfer is shown to display screen after bright data are corresponding with the first association constant.User is shown by display screen
Explanation data out select to exit or carry out next step operation, convenient directly to treat rehabilitation data to judge user by input
It is whether healthy.
Further, the authentication module for video numbering is contacted with concrete video is provided with checking assembly；It is described
It is that the combination of rehabilitation video is correspondingly arranged that expert is provided with authentication module for each option in each factor in questionnaire
The second relevant parameter；With rehabilitation video combination sum as stochastic variable number, the rehabilitation video assembled scheme be given by expert
Standard combination number as model constraints, constrain the dominance relation or incidence relation between each rehabilitation video combination, build
Vertical plan model.
Checking assembly in cloud server is receiving the request of the requirement rehabilitation video combination transmission transmitted from client
When, questionnaire be directed in authentication module each rehabilitation video by the survey data that AHP models are obtained with being arranged
Second relevant parameter of combination is calculated and is shown that the multiple rehabilitation videos combination i.e. rehabilitation training scheme for being suitable to push is combined.
Authentication module is obtained second processing device the data transfer second memory of rehabilitation video assembled scheme combination, from the second storage
Corresponding rehabilitation video combination is extracted in device and passes to client.
Further, it is additionally provided with optimization module in the first processor；It is provided with the optimization module according to various common
The physician guidance opinion data of phenomenon establishment；Optimization module calculates corresponding video and numbers simultaneously by physician guidance opinion data
The video combination calculated as variable with the adjustment amount of video numbering and the absolute value of the difference of the maximum therapy degree of expert are as mesh
Scalar functions, carry out successive iteration and obtain multigroup comprehensive parameters adjustment amount, according to weight to adjustment amount using improved intelligent algorithm
The factor of carrying out option distributes and entropy is processed, and rehabilitation video is combined into corresponding video number table and is adjusted optimization.
After client receives the rehabilitation video combination for transmitting to come from cloud server, the optimization in first processor
Module by being set in advance in optimization module in physician guidance opinion data, using improved intelligent algorithm to rehabilitation video group
Closing related all kinds of parameter values carries out Automatic Optimal.
Further, it is provided with screening module in the second processing device；The screening module is by real-time update in database
Rehabilitation data are treated as the input data and categorical data of the decision-tree model being arranged in screening module, pair rehabilitation is treated with these
Problem row classification in the corresponding questionnaire of data, using CART algorithms decision-tree model is built；Screening module is directed to from defeated
Enter whether the user data being input into component needs to push the judged result of rehabilitation video combination to first processor transmission.
Screening module in second processing device, according to the user input data of real-time update in database to from input module
The survey answer of input carries out recurrence tree classification, makes screening module show that rehabilitation is regarded according to the input of user's symptom data
Frequency combination or the judged result of not rehabilitation video combination.
Further, the study module for arranging in the checking assembly is transferred into the institute in database by BP neural network
There are data to be trained and learn, obtain corresponding BP neural network parameter and be stored in database；Receive in checking assembly
Rehabilitation video combination judged result, study module according to input module be input into treat rehabilitation data and BP neural network parameter to
The corresponding video numbering for pushing the combination of rehabilitation video of second processing device output；At the same time background program by this secondary data more
BP neural network parameter before new.
Study module in checking assembly receives the result for carrying out rehabilitation video combination that screening module transmission comes, checking
Study module in component is entered by the user data that BP neural network will be stored in being obtained after real-time update in database
Row obtains BP neural network parameter after calculating.BP neural network parameter refers to that the institute of BP neural network topological structure can be determined
There is parameter.When user carries out rehabilitation video assembled scheme by input module to be customized, study module is by checking assembly in rehabilitation
BP neural network parameter is added in video anabolic process, changes the video numbering of the rehabilitation video combination for pushing, what optimization was pushed
Video is combined, while the model parameter of the BP neural network comprising current input is preserved to database.Second processing device
Video push is carried out to client by the video combination for obtaining.
Further, it is additionally provided with correcting module in the second processing device；SVMs is provided with the correcting module
Model, correcting module judges whether to re-start video push according to the feedback data of user, at the same second processing device according to
Relation between the feedback data real-time update rehabilitation video combination of user and video numbering；And pass revised associated data
It is delivered to database preservation.
After user is operated by the rehabilitation video combination for pushing, user is by input module input pin to push
The feedback result of rehabilitation video combination, wherein user is carried out instead by filling in regard to the survey after rehabilitation video combination operation
Feedback.Correcting module in second processing device is by feedback data motion supporting vector machine model, the correlation to the combination of rehabilitation video
Parameter is modified, and revised relevant parameter is updated in database into preservation.
Further, the running of health and fitness information interaction platform included at least two stages, and wherein first operating period exists
That collects 1000~2000 sample sizes enters into the mid-term operation phase after rehabilitation data.
The present invention has passed through video push period twice, and first time video push is mainly foundation and is arranged on first memory
In grade examine several tables to carry out video push, effectively have accumulated by this process treat rehabilitation data (as user input data,
User health data), it is that, followed by machine learning, the whole information exchange platform of constantly improve provides premise.
The present invention is acquired by the form of survey to the various inputs of user, is had to the input of user and is led
Tropism, allows users to provide the categorical data required for health and fitness information interaction platform, improves effective profit of user input information
With rate.
After the completion of first time video push, by being arranged on first processor in optimization module, can be automatically by
The improved intelligent algorithm pair all kinds of parameter values related to the combination of rehabilitation video are optimized, and have health and fitness information interaction platform
There is primary learning ability, effectively improve the accuracy of rehabilitation video combination of knowing clearly.
When second video push, database have collected certain treats rehabilitation data.In the user for collecting
Regression tree model, BP neural network and supporting vector machine model are passed sequentially through in data basis in health and fitness information interaction platform
The parameter related to the combination of rehabilitation video be adjusted it is perfect, make whole platform progressively run in have machine learning function.
Video push more hommization, intellectuality can be made according to the accumulation of user data, improve by rehabilitation video combine it is accurate
Degree.
Description of the drawings
Fig. 1 is the structural representation of the embodiment of the present invention.
Fig. 2 is the Initial operation schematic diagram of the embodiment of the present invention.
Fig. 3 is the mid-term operation schematic diagram of the embodiment of the present invention.
Specific embodiment
Below by specific embodiment, the present invention is further detailed explanation：
Reference in Figure of description includes：Checking assembly 1, second memory 2, database 3, outside port 4, nothing
Line communication component 5, second processing device 6, input module 7, video resolution component 8, display screen 9, the transmitting-receiving of first memory 10, network
Component 11, first processor 12.
As shown in figure 1, health and fitness information interaction platform includes cloud server and with cloud server by network connection
Multiple client；Client includes first processor 12, input module 7, video resolution component 8, display screen 9, first memory
10 and network transmitting-receiving subassembly 11, input module 7, video resolution component 8, display screen 9, first memory 10 and network transmitting-receiving subassembly
11 are connected by data/address bus with first processor 12, the operation of each component of the control of first processor 12 client；Network is received and dispatched
Component 11 sets up the data link of client and cloud server by least one procotol；Input module 7 is used for
The information such as user input authentication information and questionnaire answer；First memory 10 is stored via aforementioned network transmitting-receiving subassembly
11 receive from cloud server athletic rehabilitation video data and investigation treat rehabilitation data；Video resolution component 8 is used to solve
Analyse the athletic rehabilitation video data from cloud server received via aforementioned network transmitting-receiving subassembly 11；Display screen 9 is used to broadcast
Put the athletic rehabilitation video data Jing after parsing.
Cloud server is including the second memory 2 of multiple athletic rehabilitation videos that are stored with, for real-time update user tune
Interrogate volume answer database, for the identity information of user input and questionnaire answer are verified checking assembly 1,
Wireless communication components 5, outside port 4 and second processing device 6；Outside port 4, second memory 2, database, checking assembly 1,
Wireless communication components 5 are connected with second processing device 6；Outside port 4 is suitable to the second memory 2 and data of cloud server
Storehouse directs or through network and is indirectly coupled to other equipment；Checking assembly 1 is used for checking and receives via wireless communication components 5
Subscriber identity information and to second processing device 6 send the result signal；Second processing device 6 sends in response to checking assembly 1
Be verified signal, the questionnaire in controlling database is sent to aforementioned client via wireless communication components 5.User
By input module 7 will be directed to questionnaire in problem answers be input in first processor 12, first processor 12 by this
A little user data are sent to cloud server by network transmitting-receiving subassembly 11.The wireless communication components 5 of cloud server are received
User data and by these user data deliveries to second processing device 6.Second processing device 6 is by these user data deliveries to user
By these data transfers to checking assembly 1 while database 3, checking assembly 1 sends a signal to second after Validation Answer Key
Processor 6, makes second processing device 6 that corresponding video data is transferred from second memory 2, and these video datas are passed through into nothing
Line communication component 5 is transferred to client.
The later stage study module for carrying out later stage study is provided with the second processing device, in the later stage study module
The interstitial content of the BP neural network model of setting is more than the nodes in the BP neural network model in the study module.
By the calculating of the BP neural network of the multinode in later stage study module, can more optimize the combination of rehabilitation video
Related video parameter table, makes the video push of health and fitness information interaction platform more intelligent.
The present invention has passed through video push period twice, and first time video push is mainly foundation and is arranged on first memory
In grade examine several tables to carry out video push, effectively have accumulated by this process treat rehabilitation data (as user input data,
User health data), it is that, followed by machine learning, the whole information exchange platform of constantly improve provides premise.
The present invention is acquired by the form of survey to the various inputs of user, is had to the input of user and is led
Tropism, allows users to provide the categorical data required for health and fitness information interaction platform, improves effective profit of user input information
With rate.
After the completion of first time video push, by being arranged on first processor in optimization module, can be automatically by
The improved intelligent algorithm pair all kinds of parameter values related to the combination of rehabilitation video are optimized, and have health and fitness information interaction platform
There is primary learning ability, effectively improve the accuracy of rehabilitation video combination of knowing clearly.
When second video push, database have collected certain treats rehabilitation data.In the user for collecting
Pass sequentially through in data basis regression tree model, BP neural network and supporting vector machine model in health and fitness information interaction platform with
The related parameter of rehabilitation video combination is adjusted perfect, makes whole platform have machine learning function in progressively running.Energy
Enough accumulation according to user data, make video push more hommization, intellectuality, improve the precision combined by rehabilitation video.
As shown in Fig. 2 at the initial stage of health and fitness information interaction platform operation, not having any use in health and fitness information interaction platform
User data, by user user's symptom data and corresponding rehabilitation are constantly collected during using health and fitness information interaction platform
Training data.
First, cloud server sends questionnaire to client, and user is by the reading investigation questionnaire of display screen 9 and leads to
Cross input module 7 to be input in first processor 12 for answer i.e. user's symptom data of questionnaire.
First processor 12 carries out will determine that result by showing after tentatively judging by judge module to the symptom description
Screen 9 presents to user, so that user carries out subsequent operation selection.
The AHP models used in first processor 12 have carried out AHP and have assigned power for conventional user's phenomenon input in advance.Adopt
The relative weighting of each factor is solved with two-phase method：Using (0,1,2) three scale method each reference factor is compared two-by-two
Such as the comparator matrix of table 1 is set up afterwards.
Table 1
And calculate the importance ranking index of each element；Second stage is adopted with down conversion：Bi=(Consideration 1, it is considered to
Factor 2, it is considered to factor 3, it is considered to factor 4, it is considered to factor 5, it is considered to factor 6, it is considered to factor 7, it is considered to factor 8, it is considered to factor 9)
With Bj=(Consideration 1, it is considered to factor 2, it is considered to factor 3, it is considered to factor 4, it is considered to factor 5, it is considered to factor 6, it is considered to factor 7,
Consideration 8, it is considered to factor 9) comparator matrix that constitutes is converted into judgment matrix, and final power is calculated using geometric average method
Weight vector.Wherein, it is considered to factor can be pain cause, pain duration and pain grade etc. it is related to symptom checking because
Element, and these factors all can be imbedded in the problem of questionnaire.User passes through to answer questionnaire, to treat the shape of rehabilitation data
Formula comes in the data input comprising these factors.
Based on known element BiWith BjImportance ranking index riAnd rjCarry out Judgement Matricies.C=cij.Judgment matrix
Coefficient cijRepresent element BiWith BjBetween significance level, it is desirable to if represent cij＞ 1 then represents BiCompare BjIt is important, Bi≥Bj。cijMore
Big significance level is higher.If cij＜ 1 then represents BiCompare BjIt is important, Bi≤Bj。cijWhen=1, Bi=Bj.In order to improve judgment matrix
Reliability and uniformity, it is assumed that the relative importance of all elements pair based on being given under same transformation criterion, if this
Transformation criterion is f (ri,rj)=cij。
By judgment matrix, the diagnostic data of 12 pairs of inputs of first processor is calculated and is drawn weight, and for power
The judged result being arranged in first memory 10 is delivered in display screen 9 again, makes user's understanding health status now.
The SPA models used in first memory 10 are previously provided with medical expert to pin after the analysis of each problematic factor option
Three following healthy phenomenon grade reference tables are established to essential body position.The material elements a total of nine of suggested design
It is individual, respectively to corresponding various Considerations.
Table 2-1
| Painful area (waist) | Phenomenon is serious | Phenomenon is general | Health | Explanation |
| Consideration 1 | 1,3 | 2 | 4 | Digitized representation option code name |
| Consideration 2 | 4,5,6 | 2,3 | 1 | Digitized representation option code name |
| Consideration 3 | 3 | 2 | 1 | Digitized representation option code name |
| Consideration 4 | 3,4 | 2 | 1 | Digitized representation option code name |
| Consideration 5 | 4 | 2,3 | 1 | Digitized representation option code name |
| Consideration 6 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 7 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 8 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 9 | 4,5,6 | 2,3,4 | 1 | Digitized representation option number |
Table 2-2
| Painful area (neck) | Phenomenon is serious | Phenomenon is general | Health | Explanation |
| Consideration 1 | 1,3 | 2 | 4 | Digitized representation option code name |
| Consideration 2 | 4,5,6 | 2,3 | 1 | Digitized representation option code name |
| Consideration 3 | 3 | 2 | 1 | Digitized representation option code name |
| Consideration 4 | 3,4 | 2 | 1 | Digitized representation option code name |
| Consideration 5 | 4 | 2,3 | 1 | Digitized representation option code name |
| Consideration 6 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 7 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 8 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 9 | 4,5 | 2,3 | 1 | Digitized representation option number |
Table 2-3
| Painful area (shoulder) | Phenomenon is serious | Phenomenon is general | Health | Explanation |
| Consideration 1 | 1,3 | 2 | 4 | Digitized representation option code name |
| Consideration 2 | 4,5,6 | 2,3 | 1 | Digitized representation option code name |
| Consideration 3 | 3 | 2 | 1 | Digitized representation option code name |
| Consideration 4 | 3,4 | 2 | 1 | Digitized representation option code name |
| Consideration 5 | 4 | 2,3 | 1 | Digitized representation option code name |
| Consideration 6 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 7 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 8 | 1 | 2 | 2 | Digitized representation option code name |
| Consideration 9 | 4,5,6 | 2,3,4 | 1 | Digitized representation option number |
Table 2-4
| Numeral | 1 | 2 | 3 | 4 | 5 | 6 |
| Option | A | B | C | D | E | F |
In table 2-1 to table 2-3, digitized representation option code name refers to that the numeral of choosing is exactly the code name i.e. rehabilitation video group of option
The numbering of conjunction.Digitized representation option number, refers to that numeral limits the number of option, such as have selected AB/AC/BD, is all 2,
ABC/BCD is then 3.The conversion relation of each answer choice in assessment table is write in table 2-4 exactly.
Three-level is divided into user health situation according to healthy symptom scales reference table, therefore the contact unit in SPA is ternary.Root
The symptom data submitted to according to user and healthy symptom scales reference table to the assessment system in it is same, different, counter be weighted,
Comprehensive contact number is finally given, corresponding healthy symptom is determined according to the magnitude relationship of contact number system number.Disease is corresponded to coefficient
The serious level of shape, different coefficient corresponds to the general level of symptom, the good type (health type) of reciprocal coefficient correspondence symptom, and the contact number for calculating is respectively
The maximum is the Health Category of user in number.First processor 12 transfers the Health Category data simultaneously from first memory 10
Pass it on display screen 9, enable users to be intuitive to see the health status of oneself.
Then, the symptom data of user input is transferred to cloud server by client.Second processing device 6 is received from wireless
User's symptom data that the transmission of communication component 5 comes, and the data are verified and customer data base 3 while being transferred to checking assembly 1
Storage.
Plan model is provided with checking assembly 1.Set for video for the possibility option in model factor by medical expert
Put corresponding treatment parameter, i.e. direct the controlling for a certain symptom (such as pain degree 0-3 point) being specifically likely to occur by each video
Therapeutic effect has much, sets up associating between each factor option that video is recommended with impact scheme.With video as stochastic variable (depending on
Frequency sum is stochastic variable number, and stochastic variable can only take 0 or 1) measure, the dominance relation or incidence relation and expert between each video
Plan model is set up as model constraints to the standard combination number of video scheme, is obtained most by maximizing treatment degree
Whole value is 1 stochastic variable, and the corresponding video of these variables is final rehabilitation training scheme combination.
Checking assembly 1 obtains rehabilitation training scheme data splitting by plan model, and by the data transfer to second
Reason device 6, second processing device 6 transfers the rehabilitation video combination being stored in advance in second memory 2 according to the data, and by the health
The data is activation of diplopia frequency combination is to client.Client can be passed through by the rehabilitation video combination of access cloud server
Display screen 9 sees specific rehabilitation video.
The specific algorithm of plan model can be with as follows：
In the present embodiment, by client receive user based on questionnaire reply with regard to user's current physical condition
Description information；
Treat that rehabilitation data are the symptom description information for extracting, i.e. problem and the selected option of user；
Push scheme is rehabilitation training video；
Concrete questionnaire form, including multiple problems, each is that problem includes multiple options, and each option is specially a kind of disease
Shape describe, I for problem number, JiFor the option sum of i-th problem；
For each option, then the significance level of the option pair and affiliated problem is characterized with a weight parameter；
For each problem, then the problem pair is characterized with another weight parameter important with the selection that rehabilitation video is combined
Degree；
The comprehensive parameters of rehabilitation training video include value for multiple efficacy parameters of 0-100%, value be 0,1,2,
Priority parameters (M represents greatest priority number) of 3 ... M, and represent the stage parameter in video affiliated stage, value is 1,
2nd, 3 ... K (K represents greatest priority number)；
The restriction parameter of rehabilitation video combination is preferably single rehabilitation video and combines included rehabilitation training video
Quantity；
These parameters can by being preset in storage device in parameter correspondence table inquired about.
Then parameter setting, in the present embodiment is specifically as shown in table 3:
Table 3
Then, the select permeability for pushing result is converted into O-1 integer programming models；Specifically, that is, meeting certain
Under the conditions of maximization calculate
Max g (x)=Kernel (Kernel (R, Handamard (λ, U)), x)
Hadamard symbols represent the Hadamard products of matrix/vector in formula；Kernel represents vector dot product
The maximization is calculated to be needed to meet following constraints：
The quantity of video is N；
In the rehabilitation stage residing for user, the video for not meeting the stage-training must not be exported
Stage parameter only comprising the video in the stage that belongs to together, the rehabilitation stage according to corresponding to user and each video is selected
Take；
In the case of same satisfaction treatment, the video (being determined by the priority parameters of video) for selecting priority high；
Specific maximization is calculated as solving：
The X that the purpose for calculating is to solve for meet the constraint condition is maximized, in the X for solving, chooses corresponding according to its result
The combination of rehabilitation video is pushed to user, and now maximized numerical value is then the integrated effect parameter of the video combination, is below remembered
Forg。
Finally, by being arranged on first processor 12 in improved intelligent algorithm model, to whole health and fitness information interact
Model in platform is optimized.Physician guidance opinion data is provided with first processor 12, the physician guidance opinion data
It is that medical expert works out common tens class symptom datas, and holds medical expert's seminar to provide objective standard to each symptom
True therapeutic scheme combination.The corresponding comprehensive parameters of video are calculated according to this tens classes symptom data, with the adjustment of comprehensive parameters
Measure as variable, the video combination and the absolute value of the difference of the maximum therapy degree of expert that model is calculated is object function, is adopted
Improved intelligent algorithm carries out successive iteration and obtains multigroup comprehensive parameters adjustment amount, and factor choosing is carried out to adjustment amount according to weight
Item distribution and entropy are processed, and most at last the corresponding whole video parameter table of video is adjusted optimization.
Wherein, the push scheme being previously set is preferably rehabilitation training video, rehabilitation guide picture and text and rehabilitation training
Explain one or more in audio frequency；Each independent push scheme by professional person it is artificial impart comprehensive parameters.
Comprehensive parameters are included but is not limited to：Single push scheme for it is multinomial treat rehabilitation data efficacy parameter respectively,
Priority parameters of single push scheme etc.；
For example：It is r to push the corresponding effectively treatment ability in illness B of option AAB(rAB∈ [0,1]), the size of its value
The power of the push scheme effectively treatment ability can be weighed.
Priority parameters are that push scheme carries out precedence sequence, first filter out and treat the push that push scheme is associated
The push scheme of scheme nothing to do with connection, sets uncorrelated push alternative priority and is classified as 0, and associated push scheme is then by excellent
First sequence grade setting 1,2,3 ... ..., and ＜ ... the ＜ n of 1 ＜ 2 (there are multiple videos and be in same priority), for same
The information of dominance relation will carry out number and limit output according to comprehensive effectively treatment ability height.
Meanwhile, the combination of push scheme has also been previously set restriction parameter, and the restriction parameter is preferably single push scheme
The quantity of the single push scheme that combination is included；
According to the restriction parameter that rehabilitation data, each comprehensive parameters and push scheme for pushing scheme are combined is treated, calculate
Go out optimized push scheme combination and be sent to user.
If user is carried out after the i-th wheel rehabilitation training, the diagnosis of new round questionnaire will be participated in, when diagnostic result shows the user still
Next round training need to be received, system recommends rehabilitation training and corresponding training that the wheel carries out then according to next round diagnostic data
Number of days, the feedback in training process carries out i+1 wheel rehabilitation training push, until user can be fully achieved medical health
Till standard.This large amount of and unduplicated specific aims push scheme and push the monitor in real time that can be realized to user health situation,
Be conducive to the progressively rehabilitation of user's body.
Optimum push side is solved by the way of plan model is set up in the first stage of health and fitness information interaction platform
Case is combined.
Information will be pushed scheme as stochastic variable, combined for the effect of user is as mesh with maximizing push scheme
Mark, sets up constraints to establish plan model according to the restriction parameter of push scheme combination, finally tries to achieve the combination of push scheme
Result.
It is more highly preferred to, can also be by the push scheme for calculating combination and medical practitioner according to same user data system
Fixed push scheme combination is compared；Each comprehensive parameters for pushing scheme is updated according to comparison result.
In advance, medical expert works out common tens class illnesss, and hold medical expert's seminar to each symptom to
Go out objective accurate push scheme combination.After calculating the push scheme combination for pushing every time, with the renewal of comprehensive parameters
Measure as variable, for same symptom, the combination of push scheme and expert that model is calculated be given combination maximum therapy degree it
Poor absolute value is object function, carries out successive iteration using intelligent algorithm and obtains multigroup comprehensive parameters renewal amount, entropy assessment pair
Renewal amount carries out multifactor distribution, and most at last the corresponding whole video parameter table of video is updated optimization.
Traditional intelligent algorithm, by simulating a certain natural process some complicated engineering problems are solved；Such as heredity
The survival of the fittest in natural selection of algorithm simulation nature, with all individualities in a kind of groups as object, and instructs right using randomized technique
One is coded of parameter space and carries out effective search；Simulated annealing is then according to being solid matter annealing process and group
Close the similitude between optimization problem.Thermodynamic (al) theoretical set is used statistically, every bit in search space is imagined into
Molecule in air；The energy of molecule, is exactly the kinetic energy of itself；And the every bit in search space, also like air molecule one
Belt transect has " energy ", to represent the appropriate level to proposition.Algorithm first makees starting with an arbitrfary point in search space：Often
One step first selects one " neighbours ", then calculates again from existing position and reaches the probability of " neighbours ".
Nowadays, more improved intelligent algorithms are occurred in that, with more Computationally efficient and search precision；These algorithms are all
In can be used for embodiments of the invention, these algorithms include but is not limited to genetic annealing algorithms/improvement particle algorithm/self adaptation
The differential evolution algorithm of adjustment.
In order to further optimize renewal, physician guidance suggestion work-in parameters table is further employed in this enforcement to be carried out
Renewal amount is calculated.
The source of doctor's suggestion data is that in advance, expert works out tens common class illnesss, and each class illness is comprising multiple
Symptom, is equal to and multiple symptom options is have selected in questionnaire；If P illness, the symptom option of each illness can be used
One answer vector UpTo represent；
Then, hold charrette and objective and accurate push scheme combination is given to each illness.
Further parameter setting is as shown in table 4：
Table 4
Then, for the renewal of comprehensive parameters is converted into following minimization problem：
This enforcement is solved using improved intelligent algorithm to the minimization problem；It is alternative herein to employ,
The differential evolution algorithm of genetic annealing algorithms/improvement particle algorithm/self-adaptative adjustment, three kinds of algorithms.
In genetic annealing algorithms, specifically using binary coding, just for population generation after, according to the survival of the fittest and winning
The bad principle eliminated, develops by generation and produces best xpApproximate solution；And in order to accelerate search speed and strengthen the local of algorithm
Search capability, by defining single individual energy assessment function and initial temperature, is scanned for using simulated annealing.
In improved particle algorithm, the calculation of new colony's extreme value and individual extreme value has been used.
For each individual extreme value Pbi, from remaining with for the extreme value that an individual is randomly selected in other individual extreme values
Pbj is denoted as, the new individual extreme value for producing is denoted as PbiThen：
Pbi(t)=r*Pbi(t)+(1-r)*Pbj(t)
Wherein t is represented and is run to t generations, and i represents i-th particle, and j represents j-th particle, and r is [0,1] for randomly generating
Between number.
Improvement to individual colony's extreme value Gb is as follows:
All of individual extreme value is ranked up, then from wherein picking out K best individuality Pb1′,Pb2′,,,
Pbk', Gb ' is represented with this K individual weighted average, then
Wherein t is represented and is run to t generations, and i represents i-th particle picked out, aiMeet
In the differential evolution algorithm of self-adaptative adjustment, adaptive operator λ is added in conventional differential evolution algorithm, repeatedly
Change mutation operator during generation, by way of gradually reducing, make population keep diversity of individuals in the early stage, it is to avoid be precocious,
Retain excellent information in the later stage, it is to avoid optimal solution is destroyed, increase searches the probability of globally optimal solution：
In formula：F0For mutation operator；GmFor maximum evolutionary generation：G is current evolutionary generation.
By the one kind in said method, x is finally solvedpBest fit approximation solution, and obtain final micro- according to weight
Adjust matrix Yp, i.e.,：
Yp=Kernel (xp,πp)；
Y is individually calculated to all illnesssp；
Finally, rehabilitation video is updated with the incidence matrix R of symptom description information using the fine setting matrix, i.e.,：
R ' is then the incidence matrix after updating.
This is arrived, the first operating period of health and fitness information interaction platform is completed.
As shown in figure 3, Initial operation for a period of time after, after the user data that customer data base 3 collects enough, be good for
Health information exchange platform initially enters the mid-term operation phase.
First, the answer of questionnaire is that symptom data is input to first by input module 7 by client by user
In processor 12, first processor 12 is delivered to these symptom datas in cloud server by network transmitting-receiving subassembly 11.Cloud
Wireless communication components 5 in the server of end receive symptom data and symptom data are transferred into second processing device 6.
Decision-tree model, the user's symptom collected during customer data base 3 is run in the early stage are used in second processing device 6
Data and user health status data respectively as decision-tree model input data and categorical data, to questionnaire in ask
Topic row classification, is decimal number by the data processing of these questionnaire problems, and total Options that will be in each problem regard 0 as
Or 1 bi-values, when option is selected, then take 1, otherwise take 0；Then these values are linked to be into one group of binary string, adopt
It (is discrete type { 2 for the feature value of single choice problem that binary system is converted into decimal number by conversion method processed1,22,…,2n-1, n
Number of options corresponding to the problem }；For multiple-choice question, then feature value is continuous type { 12n-1- 1, n are right for the problem
The number of options answered }), decision-tree model is built using CART algorithms.
By the problem data in survey through pretreatment, regard the total Options in each problem the two of 0 or 1 as
Unit's value, when option is selected, then takes 1, otherwise takes 0；Then these values are linked to be into a binary string, using system conversion side
Binary system is converted into decimal number by method；Feature set is formed, feature set specifically can be expressed as：Feature1, Feature2,
Feature3, Feature4, Feature5, Feature6, Feature7, Feature8, Feature9 }
Calculate through conversion it is known that the concrete value condition of each feature set is as shown in the table：
Table 5
Arranged according to decision-tree model, classification collection is { Level1, Level2, Level3 }, corresponding value may be configured as
{ 1,2,3 }.
User is classified from the information that client is input into by decision-tree model, and is judged whether by classification results
Need to carry out scheme push to user.If Level1 or Level2 then prove to need to carry out Rehabilitation Training in Treating, otherwise without
Must recommend (to refer to assessment table, specify that the one or two grade of explanation symptom substantially, third level explanation body is healthier, nothing
Rehabilitation training must be carried out).
If it is determined that result be to push, then can from cloud server to client send be unsatisfactory for condition letter
Breath：You are very healthy need not to push Regimen Chemotherapy.If it is judged that to need push scheme, then the output of decision-tree model is tied
Fruit can be transmitted directly in the BP neural network model being arranged in checking assembly 1.
User accesses cloud server by client, and can see on display screen 9 judge through decision-tree model
The health status of oneself.And then following operation is carried out according to the prompting on display screen 9.
Data collected by first operating period are trained and are learned by the BP neural network in checking assembly 1 on backstage
Practise, obtain corresponding BP neural network parameter and preserve beyond the clouds.When there is user to be customized scheme, model beyond the clouds is preserved
The corresponding rehabilitation video combination of rehabilitation data output is treated by what is submitted to according to user；At the same time background program by this secondary data more
BP neural network parameter before new.During whole model training, network inputs are user's symptom data (binary coding
Character string), desired output is doctor to the resulting treatment video data after video parameter table carries out one-zero programming model calculating
Video output data after being modified.Data (binary decision after being modified to the video scheme obtained by initial stage
Vector form), network settings are two hidden-layer structure, and are learnt using gradient descent algorithm.
BP neural network model is specifically calculated as follows：
Step one：Initialization network:By the user data passed over by decision-tree model and supporting vector machine model
These input datas are processed into binary data by the input of composition neutral net, and the input and output of neutral net constitute one
Individual sequence, is expressed as (x, y), and according to this sequence, we are capable of determining that input layer, hidden layer and the output node layer of network
Number, respectively n node, l node, m node.Then, connection weight and threshold value are initialized, if input layer, hidden layer are refreshing
Connection weight between Jing is first is wij, the connection weight between hidden layer and output layer neuron is wjk, hidden layer and output layer
Threshold value be respectively a and b.Finally, neuron excitation function f (X) and learning rate η of neutral net are given.
Step 2：Calculate the output of hidden layer：According to input variable X of network, the connection between input layer and hidden layer
Weight wij, and threshold value a of hidden layer, the output H of hidden layer can be drawn：
F is the excitation function of hidden layer in above formula；L is the node number of hidden layer.Excitation function is：
Step 3：Calculate the output of output layer：The output of hidden layer, contact hidden layer and output are calculated according to step 2
Connection weight w of layerjkAnd output layer threshold value b, BP neural network prediction output 0 can be drawn:
Step 4：Calculation error:O is exported according to the neural network forecast that the desired output Y and step 3 of network draw, can
Predicated error e of calculating network:
ek=Yk-OkK=1,2 ..., m
Step 5：Update weights:Connection weight w according to predicated error e of neutral net to networkijAnd wjkCarry out more
Newly：
wjk=wjk+ηHjekJ=1,2 ..., l；K=1,2 ..., m
Wherein, η is the learning rate of neutral net.
Step 6：Update threshold value:According to predicated error e of network threshold value a, b of network hidden layer and output layer is entered again
Row updates:
bk=bk+ekK=1,2 ..., m
Step 7：Judge whether to meet algorithm iteration termination condition, if being unsatisfactory for, return to step two.
The BP result of calculations that BP neural network model is obtained are the second propelling data, the second propelling data and the second storage
Video scheme is sent to client or the position by storage video scheme by the video scheme correspondence in device 2, second processing device 6
Data is activation is obtained to client by client.
In order to improve the convergence rate of BP neural network, second order gradient method is adopted after step 5：
W (t+1)=w (t)-η [▽2E(t)]-1▽E(t)
Wherein：
Although second order gradient method has reasonable convergence, need to calculate second dervatives of the E to w, amount of calculation is very big.
It is general directly not adopt second order gradient method, and variable-metric method or conjugate gradient method are adopted, they have such as the convergence of second order gradient method
Fast advantage, and without the need for directly calculating second order gradient.
Wherein, variable metric algorithm：
W (t+1)=w (t)+μ H (t) D (t)
Δ w (t)=w (t)-w (t-1)
Δ D (t)=D (t)-D (t-1)
The supporting vector machine model in second processing device 6 is arranged on using two common category support vector machines models, will
Client uses the feedback data set after video：{ receiving the pain after treatment, function improves degree } is used as known training
Collection.
Step one：If known training set:
T={ (x1,y1),…,(xl,yl)}∈(X×Y)l
Wherein, xi∈X∈Rn, yi∈ Y={ 1, -1 } (i=1,2 ..., l)；xiIt is characterized vector.
Step 2：The appropriate outstanding K of kernel function (x, x ') and appropriate parameter C is selected, optimization problem is constructed and solve:
Obtain optimal solution:
Step 3：Choose α*A positive componentAnd calculate threshold value accordingly:
Step 4：Construction decision function:
It is 1 or -1 according to f (x), it is resolved that its classification ownership.
F (x) represents that user's each several part function is recovered normal when being 1, through neutral net, model to user pushes
Video is accurately useful, and the rehabilitation training for still needing to participate in next stage is represented when f (x) is -1, now by the { pain of user feedback
Pain：Average, variance, function improves degree：Average, variance } supporting vector machine model is input to, then again by BP nerve nets
Network model is optimized again to client output video data, the i.e. calculating to BP neural network model.
In the continuous feedback of supporting vector machine model, can be according to the feedback data of user is by the symptom of user input and pushes away
Constantly corrected between the video for sending, make neural network model below more accurate when combining for user's rehabilitation video.
Meanwhile, amendment data are tracked for user when video rehabilitation training is carried out, and user can be made to realize more rapid rehabilitation
Effect (can receive the adjustment later stage push scheme of the symptomatic reaction after rehabilitation training, reach the customization of hommization) according to user.
Deep learning model is mainly set up in the later stage operation of health and fitness information interaction platform, increases mid-term machine learning
The neutral net hidden node number in stage, and using user feedback data as the interim foundation treated, direct construction is bigger
All of model before the neutral net replacement of type.
In this programme, health and fitness information interaction platform, including the operational process at least two stages, wherein Initial operation rank
Section enters into the mid-term operation phase collect 1000~2000 sample sizes after rehabilitation data.The mid-term operation phase is in operation
The operation phase in later stage is entered into after rehabilitation data to 5000~7000 sample sizes.One complete sample size includes treating rehabilitation
Data obtain the satisfied i.e. stopping nature of rehabilitation video combination from input health and fitness information interaction platform to user and stop using health
Produced all data in information exchange platform this complete procedure.
First operating period generally comprises AHP models, SPV models and plan model；The mid-term operation phase generally comprises certainly
Plan tree-model, BP neural network model and supporting vector machine model.Operation phase in later stage is introduced into convolutional neural networks and sets up deep
Degree learning system.
Improved intelligent algorithm used in the present embodiment is differential evolution algorithm.It is comprised the following steps that：
Step one：Initialization.Adopt differential evolution algorithm by the use of NP dimension for D real number value parameter vector as each
The population in generation, each individuality is expressed as：Xi,G(i=1,2 ..., NP)
In formula：I represents the-individual sequence in population；G represents evolutionary generation；NP represents population scale, is minimizing
During NP keep it is constant.
It is assumed that all random initializtion populations are met with non-uniform probability distribution.The boundary of setting parameter variable isThen：
In formula, the uniform random number that rand [0,1] is produced between [0,1].
Step 2：Variation.To each object vector Xi,G(i=1,2 ..., NP), the variation of basic differential evolution algorithm
Vectorial following generation：
Wherein, randomly selected sequence number r1r2And r3It is different, and r1r2And r3Also should be different from object vector sequence number i,
So must being fulfilled for NP >=4.Mutation operator F ∈ [0,2] is a real constant factor, controls the amplification of variation variable.Cause
Real constant is taken for mutation operator, the more difficult determination of mutation operator in enforcement, aberration rate is too big, makes algorithm search inefficiency, tries to achieve
Globally optimal solution precision is low；Aberration rate is too little, and population diversity is reduced, and " precocity " phenomenon easily occurs.Therefore add it is following from
Adequate variation operator λ, according to algorithm search progress, adaptive mutation rate design is as follows：
In formula：F0For mutation operator；GmFor maximum evolutionary generation：G is current evolutionary generation.
When algorithm starts, adaptive mutation rate is F0~2F0, with higher value, diversity of individuals is kept in the early stage, keep away
Exempt from precocity；As algorithm progress mutation operator is gradually reduced, to the close F of later stage aberration rate0, retain excellent information, it is to avoid optimum
Solution is destroyed, and increase searches the probability of globally optimal solution.
Step 3：Intersect.In order to increase the diversity of interference parameter vector, crossover operation is introduced.Then trial vector is changed into：
ui,G+1=(u1i,G+1,u2i,G+1,...,uDi,G+1)
(i=1,2 ... NP；J=1,3 ..., D)
In formula：Randb (j) produces j-th estimate of randomizer between [0,1]；Rnb (i) ∈ 1,2 ..., D
It is a randomly selected sequence, with it u is guaranteedi,G+1At least from Vi,G+1Obtain a parameter；CR is crossover operator, value
Scope is [0,1].
In order to keep the diversity of colony, the crossover operator of following random scope is devised：
p0=rand (0,1)；
Step 4：Select.To determine trial vector ui,G+1Whether member in the next generation can be become, will according to greedy criterion
Object vector X in trial vector and current populationi,GIt is compared.If object function will be minimized, then with less
The vector of target function value will win a position on the ground in population of future generation.All individualities in the next generation are all than current population
Correspondence is individual more preferably or at least equally good.
Step 5：The process of boundary condition.In having the problem of boundary constraint, it is ensured that the parameter value for producing new individual is located at
It is necessary in the feasible zone of problem, a straightforward procedure is the new individual for not meeting boundary constraint to be used in feasible zone at random
The parameter vector of generation replaces.
I.e.：IfOrSo：
Test example：
Compare for convenience, health and fitness information interaction platform (this example) and the existing health and fitness information interaction from embodiment is flat
Platform (comparative example) being tested, the newness degree and making material all same of two health and fitness information interaction platforms.
First, two groups of group members of two groups of ages, sex and conditions all sames are chosen, respectively simultaneously with the present embodiment
Health and fitness information interaction platform (this example) every other day enter in the identical period with existing health and fitness information interaction platform (comparative example)
Row first use.Contrast for convenience, each age bracket and conditions identical user take respectively two kinds of masculinity and femininity
Property others carry out contrast and use.Have recorded in table 6 each group member input after rehabilitation data are finished to receiving push health
The response time of diplopia frequency combination.
Table 6
Unit in table 5 is the second, wherein it is health when not having rehabilitation data completely to test the primary response time
Information exchange platform.As can be seen from the above table, response time 20 years old and 30 years old of the people of each age bracket in the same time period
The response time of people is most short, and this skilled operation with the people at the two ages is relevant.And with same age bracket in a small group and
The response time difference of the masculinity and femininity of identical conditions is little.
In for every a line, in two groups of group members of the same same age bracket of conditions and same sex, using installation
There is the group member of the health and fitness information interaction platform in the present embodiment, (without itself band rehabilitation data storage when using first
Situation) just there is the shorter response time, and using the present embodiment health and fitness information interaction platform can with using time
Several increase and more shorten the response time, and use which group group member of existing health and fitness information interaction platform, although response
Time is gradually shortened also with increasing for access times, but the response time of correspondence time period is longer than another group of group member.
Absolutely prove that health and fitness information interaction platform can remain able to carry out in the case where not having rehabilitation data in the present embodiment
Normally use, i.e., the health and fitness information interaction platform in the present embodiment is not relied on treats rehabilitation data.Also illustrate that simultaneously with
What is continually entered during the use of family treats rehabilitation data, and machine learning system can constantly carry out study optimization, improve and calculate speed
Degree, makes whole health and fitness information interaction platform rapidly the scheme of suitable user are pushed into user.
Then, the rehabilitation training for making every group of group member be carried out according to the push scheme that the 4th time obtains in table 6 six months by a definite date,
And at March, April, May and June, the rehabilitation efficacy of each group member is evaluated respectively.Evaluation criterion
To go the health check-up for completing identical standard of identical hospital, table 7 is that the result carried out according to physical examination result is counted, wherein 100 represent
Return to one's perfect health, 0 represents do not have rehabilitation completely, that is, push scheme and do not play a role completely.
Table 7
As can be seen from the above table, the people of each age bracket is different in the rehabilitation speed of same time period, wherein 10 years old, 20 years old
Very fast with the rehabilitation speed of the people of 30 years old, this is relevant with the fitness of the people at these three ages.And with same year in a small group
Age section is little with the rehabilitation speed difference of the masculinity and femininity of identical conditions.
In for every a line, in two groups of group members of the same same age bracket of conditions and same sex, using this reality
The physical examination result that the health and fitness information interaction platform in example is better than another group of group member in each health check-up middle of the month physical examination result is applied,
Absolutely prove that the scheme pushed using the health and fitness information interaction platform of the present embodiment is more suitable for user, pushed rehabilitation video group
The precision of conjunction is higher, and more preferably, the health and fitness information interaction platform in the present embodiment treats rehabilitation number to the effect of generation by identical
It is more preferable according to the rehabilitation efficacy for obtaining.
Above-described is only embodiments of the invention, and the general knowledge here such as known concrete structure and characteristic is not made in scheme
Excessive description, one skilled in the art know that technical field that the present invention belongs to is all of before the applying date or priority date
Ordinary technical knowledge, can know all of prior art in the field, and with using normal experiment hand before the date
The ability of section, one skilled in the art can improve and implement under the enlightenment that the application is given with reference to self-ability
This programme, some typical known features or known method should not become one skilled in the art and implement the application
Obstacle.It should be pointed out that for a person skilled in the art, on the premise of without departing from present configuration, can also make
Go out some deformations and improvement, these should also be considered as protection scope of the present invention, these are all without the effect for affecting the present invention to implement
Fruit and practical applicability.This application claims protection domain should be defined by the content of its claim, the tool in specification
Body embodiment etc. records the content that can be used for explaining claim.
Claims (8)
1. health and fitness information interaction platform, including cloud server and the multiple client for passing through network connection with cloud server；
It is characterized in that：The client includes first processor and the input module, the display screen that are connected with first processor respectively
And first memory；The cloud server includes second processing device, and being stored with of being connected with second processing device respectively is more
The second memory of individual rehabilitation video combination, the database for real-time update user investigation questionnaire and its answer, for
The checking assembly that the identity information of family input and questionnaire answer are verified；Grade reference is provided with the first memory
Table；The relation that the grade reference table agreement is numbered with video；Rehabilitation of the video numbering with storage in the database
Video combination is corresponded.
2. health and fitness information interaction platform according to claim 1, it is characterised in that：Be provided with the first processor for
Judge the judge module of input module input information；AHP models are provided with the judge module；AHP models employ two ranks
Section method solves the relative weighting of each assessment factor of questionnaire；First processor calculates use according to the weighted value of each assessment factor
The first association constant at family, first processor will be arranged in first processor explanation data corresponding with the first association constant and pass
It is delivered to display screen.
3. health and fitness information interaction platform according to claim 1, it is characterised in that：It is provided with checking assembly for by video
The authentication module that numbering is contacted with concrete video；Be provided with the authentication module expert in questionnaire it is each because
Each option in element is the second relevant parameter that the combination of rehabilitation video is correspondingly arranged；With rehabilitation video combination sum as random change
Amount number, the standard combination number of the rehabilitation video assembled scheme be given by expert is used as model constraints, constrains each rehabilitation
Dominance relation or incidence relation between video combination, sets up plan model.
4. health and fitness information interaction platform according to claim 1, it is characterised in that：It is additionally provided with the first processor excellent
Change module；The physician guidance opinion data according to the establishment of various common phenomena is provided with the optimization module；Optimization module passes through
The video group that physician guidance opinion data is calculated corresponding video numbering and calculated as variable with the adjustment amount of video numbering
The absolute value of the difference of the maximum therapy degree of conjunction and expert is object function, carries out successive iteration using improved intelligent algorithm and obtains
To multigroup comprehensive parameters adjustment amount, the distribution of factor option is carried out to adjustment amount according to weight and entropy is processed, by the combination of rehabilitation video
Corresponding video number table is adjusted optimization.
5. health and fitness information interaction platform according to claim 1, it is characterised in that：Screening is provided with the second processing device
Module；Real-time update in database is treated the rehabilitation data as the decision tree mould being arranged in screening module by the screening module
The input data and categorical data of type, pair treats the problem row classification in the corresponding questionnaire of rehabilitation data with these, adopts
CART algorithms build decision-tree model；Screening module is transmitted for the user data being input into from input module to first processor
Whether the judged result that push rehabilitation video combination is needed.
6. health and fitness information interaction platform according to claim 1, it is characterised in that：The study arranged in the checking assembly
All data that module is transferred into database by BP neural network are trained and learn, and obtain corresponding BP nerve nets
Network parameter is stored in database；The judged result of rehabilitation video combination is received in checking assembly, study module is according to input
What component was input into treats that rehabilitation data and BP neural network parameter push what rehabilitation video was combined to the output of second processing device is corresponding
Video is numbered；At the same time the BP neural network parameter before background program updates this online data.
7. health and fitness information interaction platform according to claim 1, it is characterised in that：It is additionally provided with the second processing device and repaiies
Positive module；Be provided with supporting vector machine model in the correcting module, correcting module judged according to the feedback data of user be
It is no to re-start video push, while second processing device is combined and video according to the feedback data real-time update rehabilitation video of user
Relation between numbering；And revised associated data is delivered into database preservation.
8. health and fitness information interaction platform according to claim 1, it is characterised in that the operation of health and fitness information interaction platform
Journey included at least two stages, and wherein first operating period is collecting entering after rehabilitation data for 1000~2000 sample sizes
To the mid-term operation phase.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610881071.6A CN106682385B (en) | 2016-09-30 | 2016-09-30 | Health information interaction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610881071.6A CN106682385B (en) | 2016-09-30 | 2016-09-30 | Health information interaction system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106682385A true CN106682385A (en) | 2017-05-17 |
| CN106682385B CN106682385B (en) | 2020-02-11 |
Family
ID=58840055
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610881071.6A Active CN106682385B (en) | 2016-09-30 | 2016-09-30 | Health information interaction system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106682385B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108766320A (en) * | 2018-06-01 | 2018-11-06 | 深圳安麦思科技有限公司 | A kind of liquid crystal displaying windows system |
| CN110391020A (en) * | 2018-04-18 | 2019-10-29 | 智信生医科技股份有限公司 | Anoxic factor expert system and its method and computer program product |
| CN111695090A (en) * | 2020-06-15 | 2020-09-22 | 甘肃中医药大学 | Health education investigation evaluation system based on Android platform |
| CN111834005A (en) * | 2020-07-02 | 2020-10-27 | 医渡云（北京）技术有限公司 | Method, device, medium and equipment for screening medical data based on infectious diseases |
| CN113485983A (en) * | 2021-06-07 | 2021-10-08 | 上海众言网络科技有限公司 | Questionnaire data processing method and device |
| CN113593697A (en) * | 2021-07-31 | 2021-11-02 | 吾征智能技术(北京)有限公司 | System, equipment and medium for judging diseases based on breath odor |
| TWI790795B (en) * | 2021-10-29 | 2023-01-21 | 財團法人資訊工業策進會 | Model adjustment method, model adjustment system and non-transitory computer readable medium |
| CN115831371A (en) * | 2023-02-16 | 2023-03-21 | 成都体育学院 | Child physique health assessment system and equipment |
| WO2023241280A1 (en) * | 2022-06-15 | 2023-12-21 | 北京京东拓先科技有限公司 | Health state assessment method and apparatus, and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001025970A1 (en) * | 1999-10-06 | 2001-04-12 | Accenture Llp | Method and estimator for providing operations maturity model assessment |
| CN103632328A (en) * | 2013-12-05 | 2014-03-12 | 青岛海尔软件有限公司 | Household chronic disease management system with health guidance video playing function |
| CN104794355A (en) * | 2015-04-29 | 2015-07-22 | 孙卫唯 | Method for executing online rehabilitation guidance in computer system and rehabilitation guidance system |
| CN105321135A (en) * | 2014-07-04 | 2016-02-10 | 北京大学第三医院 | Personalized exercise prescription design method and system |
-
2016
- 2016-09-30 CN CN201610881071.6A patent/CN106682385B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001025970A1 (en) * | 1999-10-06 | 2001-04-12 | Accenture Llp | Method and estimator for providing operations maturity model assessment |
| CN103632328A (en) * | 2013-12-05 | 2014-03-12 | 青岛海尔软件有限公司 | Household chronic disease management system with health guidance video playing function |
| CN105321135A (en) * | 2014-07-04 | 2016-02-10 | 北京大学第三医院 | Personalized exercise prescription design method and system |
| CN104794355A (en) * | 2015-04-29 | 2015-07-22 | 孙卫唯 | Method for executing online rehabilitation guidance in computer system and rehabilitation guidance system |
Non-Patent Citations (2)
| Title |
|---|
| 刘小菊 等: "基于ＢＰ神经网络的学生运动处方管理系统", 《浙江体育科学》 * |
| 张维秋: "五自由度上肢康复机器人康复评价系统的研究和实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110391020A (en) * | 2018-04-18 | 2019-10-29 | 智信生医科技股份有限公司 | Anoxic factor expert system and its method and computer program product |
| CN108766320A (en) * | 2018-06-01 | 2018-11-06 | 深圳安麦思科技有限公司 | A kind of liquid crystal displaying windows system |
| CN111695090A (en) * | 2020-06-15 | 2020-09-22 | 甘肃中医药大学 | Health education investigation evaluation system based on Android platform |
| CN111695090B (en) * | 2020-06-15 | 2023-05-05 | 甘肃中医药大学 | Health education investigation evaluation system based on Android platform |
| CN111834005A (en) * | 2020-07-02 | 2020-10-27 | 医渡云（北京）技术有限公司 | Method, device, medium and equipment for screening medical data based on infectious diseases |
| CN113485983A (en) * | 2021-06-07 | 2021-10-08 | 上海众言网络科技有限公司 | Questionnaire data processing method and device |
| CN113485983B (en) * | 2021-06-07 | 2023-12-26 | 上海众言网络科技有限公司 | Questionnaire data processing method and device |
| CN113593697A (en) * | 2021-07-31 | 2021-11-02 | 吾征智能技术(北京)有限公司 | System, equipment and medium for judging diseases based on breath odor |
| CN113593697B (en) * | 2021-07-31 | 2023-10-13 | 吾征智能技术(北京)有限公司 | System, equipment and medium for judging diseases based on breath odor |
| TWI790795B (en) * | 2021-10-29 | 2023-01-21 | 財團法人資訊工業策進會 | Model adjustment method, model adjustment system and non-transitory computer readable medium |
| WO2023241280A1 (en) * | 2022-06-15 | 2023-12-21 | 北京京东拓先科技有限公司 | Health state assessment method and apparatus, and storage medium |
| CN115831371A (en) * | 2023-02-16 | 2023-03-21 | 成都体育学院 | Child physique health assessment system and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106682385B (en) | 2020-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106779084B (en) | Machine learning system and method | |
| CN106682385A (en) | Health information interaction platform | |
| Rotshtein et al. | Football predictions based on a fuzzy model with genetic and neural tuning | |
| CN106778014A (en) | A kind of risk Forecasting Methodology based on Recognition with Recurrent Neural Network | |
| CN106485069A (en) | The method and system of rehabilitation information pushing | |
| CN103218510B (en) | A kind of combined type task time drives chronic disease prevention and control and health control method | |
| CN106407711A (en) | Recommendation method and recommendation system of pet feeding based on cloud data | |
| CN108573303A (en) | It is a kind of that recovery policy is improved based on the complex network local failure for improving intensified learning certainly | |
| CN108805268A (en) | Deeply learning strategy network training method based on evolution algorithm | |
| CN106472332A (en) | Pet feeding method and system based on dynamic intelligent algorithm | |
| Keshtkar Langaroudi et al. | Sports result prediction based on machine learning and computational intelligence approaches: A survey | |
| CN106447117B (en) | Pet feeding method and system based on the daily data analysis of pet | |
| CN106651001A (en) | Needle mushroom yield prediction method based on improved neural network and implementation system | |
| CN107103419A (en) | One kind of groups software development process analogue system and method | |
| CN108364131A (en) | The automatic identification of personality type is carried out using neural network and divides the method for group | |
| CN112711621A (en) | Universal object interconnection training platform and control method and device | |
| Lv | [Retracted] Martial Arts Competitive Decision‐Making Algorithm Based on Improved BP Neural Network | |
| CN107657453A (en) | Cheat recognition methods and the device of data | |
| JP2022541648A (en) | Physical training system with machine learning based training program | |
| CN116883007A (en) | Method, system, electronic equipment and storage medium for recommending collection-promoting action | |
| Ji et al. | March madness prediction: a matrix completion approach | |
| Guo et al. | [Retracted] Signal Recognition Based on APSO‐RBF Neural Network to Assist Athlete’s Competitive Ability Evaluation | |
| CN110154042A (en) | A kind of actual combat humanoid robot iteration cloud record system | |
| CN106407693A (en) | Hepatitis B prediction method and prediction system based on incremental neural network model | |
| Choudhari et al. | Dream11 IPL Team Recommendation using Machine Learning and Skill-Based Ranking of Players |
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 |