CN105933172A - Cloud computing based disease self-diagnosis service construction system - Google Patents

Abstract

The invention discloses a cloud computing based disease self-diagnosis service construction system, which comprises a data resource collection module, a task planning module, a credible combination assessment module and a disease diagnosis model generation module, wherein the data resource collection module comprises a modeling submodule, a resource replication submodule and a resource search submodule, and the credible combination assessment module comprises an assessment submodule and an assessment optimization submodule. The cloud computing based disease self-diagnosis service construction system is provided with the data resource collection module, thereby facilitating users to use data resources through matching service description information, further increasing the coverage rate of data resource information in the network, and improving the search efficiency of medical record data resources; and the cloud computing based disease self-diagnosis service construction system is provided with the credible combination assessment module in order to efficiently realize low-cost disease self-diagnosis services, the credibility of a big data service supporting cloud service combination scheme is improved, and storage and computing resources of the cloud are utilized in a maximum benefit manner. In addition, the assessment time is saved, and the assessment speed is improved.

Description

Disease self diagnosis service construction system based on cloud computing

Technical field

The present invention relates to medical field, be specifically related to disease self diagnosis service construction system based on cloud computing.

Background technology

Data resource big to medical treatment is managed and builds the big data, services of medical treatment need to solve both sides technical problem: on the one hand, The software of medical field typically should be played manufacturer's exploitation, based on different hardware and software platforms, in Floor layer Technology and business by different ingeniously There is isomery vertebra widely in flow process aspect.Although the in store big data resource of abundant medical treatment in these systems (as electronic health record, Medical image etc.), but the information island phenomenon formed due to the isomerism between system so that in different medical user colonies Between to realize data sharing extremely difficult.How by integrating the big data resource of the medical treatment in these systems, it is one and there is challenge The problem of vertebra；On the other hand, use these abundant the biggest data resources of doctor to build the big data, services of medical treatment, need on-demand extension Storage resource with calculate the support of resource, processing and analyzing and the structure of big data, services brings cost to put into big data And the problem in terms of maintenance.

Summary of the invention

For the problems referred to above, the present invention provides disease self diagnosis service construction system based on cloud computing.

The purpose of the present invention realizes by the following technical solutions:

Disease self diagnosis service construction system based on cloud computing, including:

(1) data resource collect module, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module, the processing procedure for data resource big to electronic health record is divided into data storage subtask, rope Draw calculating subtask and Data Management Analysis calculates subtask, and meet the cloud service resource of its demand for each subtask coupling Pond, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module: the mission planning of the big data, services for generating according to mission planning module, performs cloud The assessment of Services Composition scheme, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates resource, It includes assessing submodule and assessment optimizes submodule；

The operation that described assessment submodule specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule and is able to record that the utility function value of preferred cloud service assembled scheme and optimum combination cloud service Scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke it Functional value；

(4) diagnosing model generation module, for according to optimum cloud service assembled scheme, sets up electronic health record index, and Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule, resources duplication submodule and the resource lookup that module includes being sequentially connected with Submodule, the overlay network that the resource node under cloud environment is formed by described modeling submodule for using Unstructured Peer-to-Peer Network Being modeled, described resources duplication submodule carries out answering of resource information between each neighbor node in described overlay network System, described resource lookup submodule meets the electronic health record data resource of application demand for lookup coupling；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule use based on the data asset information master between neighbor node when carrying out the duplication of resource information Dynamic replication protocol:

Work as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1,x_j2,…x_jm},j∈[1,n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2,…x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1, x_j2... x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The invention have the benefit that

1, data resource is set and collects module, mission planning module, credible combined evaluation module and diagnosing model generation module, Achieve the structure of disease self diagnosis service system；

2, arranging modeling submodule, resources duplication submodule and the resource lookup submodule being sequentially connected with, it is non-structured right to use On network as the topological organization structure of data resource node under cloud environment, and service encapsulation of data resource, facilitate user and lead to Overmatching service description information uses data resource, further increases data asset information coverage rate in a network, improves The efficiency of medical record data resource lookup；

3, for efficiently realizing the disease self diagnosis service of low cost, it is provided with credible combined evaluation module, improves the big data of support The credibility of the cloud service assembled scheme of service, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and employing is commented Estimate optimization submodule, saved the evaluation time, improve estimating velocity.

Accompanying drawing explanation

The invention will be further described to utilize accompanying drawing, but the embodiment in accompanying drawing does not constitute any limitation of the invention, for Those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to obtains the attached of other according to the following drawings Figure.

Fig. 1 is the connection diagram of each module of the present invention；

Fig. 2 is the workflow diagram that the present invention assesses submodule.

Reference:

Data resource collect module 1, mission planning module 2, credible combined evaluation module 3, diagnosing model generation module 4, Modeling submodule 11, resources duplication submodule 12, resource lookup submodule 13, assessment submodule 31, assessment optimize submodule 32。

Detailed description of the invention

The invention will be further described with the following Examples.

Embodiment 1

See Fig. 1, Fig. 2, the disease self diagnosis service construction system based on cloud computing of the present embodiment, including:

(1) data resource collect module 1, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module 2, for the processing procedure of data resource big to electronic health record be divided into data storage subtask, Index calculates subtask and Data Management Analysis calculates subtask, and meets the cloud service money of its demand for each subtask coupling Pond, source, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module 3: the mission planning of the big data, services for generating according to mission planning module 2, holds The assessment of Services Composition scheme of racking, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates Resource, it includes assessing submodule 31 and assessment optimizes submodule 32；

The operation that described assessment submodule 31 specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module 2 G={G₁, G₂, G₃, correspondingIt is constrained to C={C₁, C₂, C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule 32 and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business

Scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke it Functional value；

(4) diagnosing model generation module 4, for according to optimum cloud service assembled scheme, sets up electronic health record index,

And

Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule 11, resources duplication submodule 12 and that module 1 includes being sequentially connected with Resource lookup submodule 13, described modeling submodule 11 is for using Unstructured Peer-to-Peer Network to the resource node under cloud environment The overlay network formed is modeled, and described resources duplication submodule 12 is between each neighbor node in described overlay network Carrying out the duplication of resource information, described resource lookup submodule 13 meets the electronic health record data of application demand for lookup coupling Resource；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1, x_i2... x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule 12 use when carrying out the duplication of resource information to be believed based on the data resource between neighbor node Breath Active Replication agreement:

Work as x_iWhen adding overlay network, by x_iWith { x_l1, x_l2... x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule 13 specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1, x_j2... x_jm, j ∈ [1, n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1, x_j2... x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1, x_j2... x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1, x_j2... x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1, x_j2... x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The present embodiment arranges data resource and collects module 1, mission planning module 2, credible combined evaluation module 3 and medical diagnosis on disease mould Type generation module 4, it is achieved that the structure of disease self diagnosis service system；Modeling submodule 11, resource that setting is sequentially connected with are multiple System module 12 and resource lookup submodule 13, use non-structured peer-to-peer network as data resource node under cloud environment Topological organization structure, and service encapsulation of data resource, facilitate user and use data resource by coupling service description information, Further increase data asset information coverage rate in a network, improve the efficiency of medical record data resource lookup；For efficient real The disease self diagnosis service of existing low cost, is provided with credible combined evaluation module 3, improves the cloud service supporting big data, services The credibility of assembled scheme, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and uses assessment to optimize submodule 32, save the evaluation time, improve estimating velocity；The present embodiment value α=0.25, β=0.45, medical record data resource is looked into Efficiency is looked for improve 3.5%.

Embodiment 2

See Fig. 1, Fig. 2, the disease self diagnosis service construction system based on cloud computing of the present embodiment, including:

(1) data resource collect module 1, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module 2, for the processing procedure of data resource big to electronic health record be divided into data storage subtask, Index calculates subtask and Data Management Analysis calculates subtask, and meets the cloud service money of its demand for each subtask coupling Pond, source, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module 3: the mission planning of the big data, services for generating according to mission planning module 2, holds The assessment of Services Composition scheme of racking, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates Resource, it includes assessing submodule 31 and assessment optimizes submodule 32；

The operation that described assessment submodule 31 specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module 2 G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule 32 and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business

Scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke it Functional value；

(4) diagnosing model generation module 4, for according to optimum cloud service assembled scheme, sets up electronic health record index,

And

Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule 11, resources duplication submodule 12 and that module 1 includes being sequentially connected with Resource lookup submodule 13, described modeling submodule 11 is for using Unstructured Peer-to-Peer Network to the resource node under cloud environment The overlay network formed is modeled, and described resources duplication submodule 12 is between each neighbor node in described overlay network Carrying out the duplication of resource information, described resource lookup submodule 13 meets the electronic health record data of application demand for lookup coupling Resource；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule 12 use when carrying out the duplication of resource information to be believed based on the data resource between neighbor node Breath Active Replication agreement:

Work as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule 13 specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1,x_j2,…x_jm, j ∈ [1, n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2... x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1,x_j2,…x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The present embodiment arranges data resource and collects module 1, mission planning module 2, credible combined evaluation module 3 and medical diagnosis on disease mould Type generation module 4, it is achieved that the structure of disease self diagnosis service system；Modeling submodule 11, resource that setting is sequentially connected with are multiple System module 12 and resource lookup submodule 13, use non-structured peer-to-peer network as data resource node under cloud environment Topological organization structure, and service encapsulation of data resource, facilitate user and use data resource by coupling service description information, Further increase data asset information coverage rate in a network, improve the efficiency of medical record data resource lookup；For efficient real The disease self diagnosis service of existing low cost, is provided with credible combined evaluation module 3, improves the cloud service supporting big data, services The credibility of assembled scheme, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and uses assessment to optimize submodule 32, save the evaluation time, improve estimating velocity；The present embodiment value α=0.26, β=0.46, medical record data resource is looked into Efficiency is looked for improve 3%.

Embodiment 3

See Fig. 1, Fig. 2, the disease self diagnosis service construction system based on cloud computing of the present embodiment, including:

(1) data resource collect module 1, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module 2, for the processing procedure of data resource big to electronic health record be divided into data storage subtask, Index calculates subtask and Data Management Analysis calculates subtask, and meets the cloud service money of its demand for each subtask coupling Pond, source, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module 3: the mission planning of the big data, services for generating according to mission planning module 2, holds The assessment of Services Composition scheme of racking, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates Resource, it includes assessing submodule 31 and assessment optimizes submodule 32；

The operation that described assessment submodule 31 specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module 2 G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-h represents the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule 32 and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business

Scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke it Functional value；

(4) diagnosing model generation module 4, for according to optimum cloud service assembled scheme, sets up electronic health record index,

And

Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule 11, resources duplication submodule 12 and that module 1 includes being sequentially connected with Resource lookup submodule 13, described modeling submodule 11 is for using Unstructured Peer-to-Peer Network to the resource node under cloud environment The overlay network formed is modeled, and described resources duplication submodule 12 is between each neighbor node in described overlay network Carrying out the duplication of resource information, described resource lookup submodule 13 meets the electronic health record data of application demand for lookup coupling Resource；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule 12 use when carrying out the duplication of resource information to be believed based on the data resource between neighbor node Breath Active Replication agreement:

Work as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule 13 specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1,x_j2,…x_jm},j∈[1,n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2,…x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1,x_j2,…x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The present embodiment arranges data resource and collects module 1, mission planning module 2, credible combined evaluation module 3 and medical diagnosis on disease mould Type generation module 4, it is achieved that the structure of disease self diagnosis service system；Modeling submodule 11, resource that setting is sequentially connected with are multiple System module 12 and resource lookup submodule 13, use non-structured peer-to-peer network as data resource node under cloud environment Topological organization structure, and service encapsulation of data resource, facilitate user and use data resource by coupling service description information, Further increase data asset information coverage rate in a network, improve the efficiency of medical record data resource lookup；For efficient real The disease self diagnosis service of existing low cost, is provided with credible combined evaluation module 3, improves the cloud service supporting big data, services The credibility of assembled scheme, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and uses assessment to optimize submodule 32, save the evaluation time, improve estimating velocity；The present embodiment value α=0.27, β=0.47, medical record data resource is looked into Efficiency is looked for improve 3.2%.

Embodiment 4

See Fig. 1, Fig. 2, the disease self diagnosis service construction system based on cloud computing of the present embodiment, including:

(1) data resource collect module 1, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module 2, for the processing procedure of data resource big to electronic health record be divided into data storage subtask, Index calculates subtask and Data Management Analysis calculates subtask, and meets the cloud service money of its demand for each subtask coupling Pond, source, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module 3: the mission planning of the big data, services for generating according to mission planning module 2, holds The assessment of Services Composition scheme of racking, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates Resource, it includes assessing submodule 31 and assessment optimizes submodule 32；

The operation that described assessment submodule 31 specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module 2 G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule 32 and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business

Scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke it Functional value；

(4) diagnosing model generation module 4, for according to optimum cloud service assembled scheme, sets up electronic health record index,

And

Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule 11, resources duplication submodule 12 and that module 1 includes being sequentially connected with Resource lookup submodule 13, described modeling submodule 11 is for using Unstructured Peer-to-Peer Network to the resource node under cloud environment The overlay network formed is modeled, and described resources duplication submodule 12 is between each neighbor node in described overlay network Carrying out the duplication of resource information, described resource lookup submodule 13 meets the electronic health record data of application demand for lookup coupling Resource；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule 12 use when carrying out the duplication of resource information to be believed based on the data resource between neighbor node Breath Active Replication agreement:

Work as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule 13 specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1,x_j2,…x_jm},j∈[1,n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2,…x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1,x_j2,…x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The present embodiment arranges data resource and collects module 1, mission planning module 2, credible combined evaluation module 3 and medical diagnosis on disease mould Type generation module 4, it is achieved that the structure of disease self diagnosis service system；Modeling submodule 11, resource that setting is sequentially connected with are multiple System module 12 and resource lookup submodule 13, use non-structured peer-to-peer network as data resource node under cloud environment Topological organization structure, and service encapsulation of data resource, facilitate user and use data resource by coupling service description information, Further increase data asset information coverage rate in a network, improve the efficiency of medical record data resource lookup；For efficient real The disease self diagnosis service of existing low cost, is provided with credible combined evaluation module 3, improves the cloud service supporting big data, services The credibility of assembled scheme, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and uses assessment to optimize submodule 32, save the evaluation time, improve estimating velocity；The present embodiment value α=0.28, β=0.49, medical record data resource is looked into Efficiency is looked for improve 3.6%.

Embodiment 5

See Fig. 1, Fig. 2, the disease self diagnosis service construction system based on cloud computing of the present embodiment, including:

(1) data resource collect module 1, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module 2, for the processing procedure of data resource big to electronic health record be divided into data storage subtask, Index calculates subtask and Data Management Analysis calculates subtask, and meets the cloud service money of its demand for each subtask coupling Pond, source, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module 3: the mission planning of the big data, services for generating according to mission planning module 2, holds The assessment of Services Composition scheme of racking, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates Resource, it includes assessing submodule 31 and assessment optimizes submodule 32；

The operation that described assessment submodule 31 specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module 2 G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\γ}}\right)={\mathrm{\Σ}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\Σ}}_{v=1}^3{\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }\left(k\right)}}\×w_k\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\ω}}{R}_h\right)\×x_{v\mathrm{\ω}-h}\≤C_k,1\≤k\≤3\\ {\mathrm{\Σ}}_{\mathrm{\ω}=1}^{m_v}{\mathrm{\Σ}}_{h=1}^{L_{v\mathrm{\ω}}}{x}_{v\mathrm{\ω}-h}=1,x_{v\mathrm{\ω}-h}\∈\left\{0,1\right\}\\ {\mathrm{\Σ}}_k^d{w}_k=1,w_k\∈\[0,1\]\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωRh is It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule 32 and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke Its functional value；

(4) diagnosing model generation module 4, for according to optimum cloud service assembled scheme, sets up electronic health record and indexes also Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Wherein, described electronic health record index includes that case history inverted index, case history filter index and case history details index, described case history Inverted index is used for retrieving the case history identical with user's disease symptoms from the big data resource of electronic health record according to user's disease symptoms, Described case history filters index and filters the case history inconsistent with user's Sex, Age and age, institute for the sex according to user and age State case history details to index for the detailed content retrieving case history from the big data resource of electronic health record.

Wherein, described data resource collects modeling submodule 11, resources duplication submodule 12 and that module 1 includes being sequentially connected with Resource lookup submodule 13, described modeling submodule 11 is for using Unstructured Peer-to-Peer Network to the resource node under cloud environment The overlay network formed is modeled, and described resources duplication submodule 12 is between each neighbor node in described overlay network Carrying out the duplication of resource information, described resource lookup submodule 13 meets the electronic health record data of application demand for lookup coupling Resource；

If x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For this Ground resource pool,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents that neighbours save The number of point, m ＜ n；

A, described resources duplication submodule 12 use when carrying out the duplication of resource information to be believed based on the data resource between neighbor node Breath Active Replication agreement:

Work as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn information on services, Create the duplication message of a resource information, and described duplication message is transmitted to all neighbor node x_lmReplicate, if reciprocity Any node in network receives one when replicating message, judges whether to receive described multiple according to the number information replicating message Message processed, if receiving, abandons described duplication message, if receiving first, then according to resource information and the node replicating message Positional information, updatesIn content, and according to replicate message vital values, determine forward or abandon described duplication message, its In, resource information needs periodically to synchronize between neighbor node；

The operation that B, described resource lookup submodule 13 specifically perform is:

If initiating inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jRandom choose go out to constant pitch Point set is p_j×{x_j1,x_j2,…x_jm},j∈[1,n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2,…x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\αQ}}_{learn}+\mathrm{\β}\×I\[\frac{N_{x_{j\mathrm{\μ}}}\left(t\right)\left(T_{x_{j\mathrm{\μ}}}-{T^{\′}}_{x_{j\mathrm{\μ}}}\right)}{{T^{\′}}_{x_{j\mathrm{\μ}}}\×T_{x_{j\mathrm{\μ}}}}\]\×\frac{1+N_{x_{j\mathrm{\μ}}}\left(t\right)}{T_{x_{j\mathrm{\μ}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1,x_j2,…x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

The present embodiment arranges data resource and collects module 1, mission planning module 2, credible combined evaluation module 3 and medical diagnosis on disease mould Type generation module 4, it is achieved that the structure of disease self diagnosis service system；Modeling submodule 11, resource that setting is sequentially connected with are multiple System module 12 and resource lookup submodule 13, use non-structured peer-to-peer network as data resource node under cloud environment Topological organization structure, and service encapsulation of data resource, facilitate user and use data resource by coupling service description information, Further increase data asset information coverage rate in a network, improve the efficiency of medical record data resource lookup；For efficient real The disease self diagnosis service of existing low cost, is provided with credible combined evaluation module 3, improves the cloud service supporting big data, services The credibility of assembled scheme, it is achieved maximum benefitization ground uses the storage in high in the clouds and calculates resource, and uses assessment to optimize submodule 32, save the evaluation time, improve estimating velocity；The present embodiment value α=0.3, β=0.5, medical record data resource lookup Efficiency improves 4%.

Last it should be noted that, above example is only in order to illustrate technical scheme, rather than to scope Restriction, although having made to explain to the present invention with reference to preferred embodiment, it will be understood by those within the art that, Technical scheme can be modified or equivalent, without deviating from the spirit and scope of technical solution of the present invention.

Claims (5)

1. disease self diagnosis service construction system based on cloud computing, is characterized in that, including:

(1) data resource collect module, for according to disease self diagnosis service demand, collection be distributed in cloud each hospital, In clinic and the application of each medical software, the electronic health record data of patient, form the big data resource of electronic health record；

(2) mission planning module, the processing procedure for data resource big to electronic health record is divided into data storage subtask, rope Draw calculating subtask and Data Management Analysis calculates subtask, and meet the cloud service resource of its demand for each subtask coupling Pond, forms cloud service assembled scheme, to obtain storage resource required in big data handling procedure or to calculate resource；

(3) credible combined evaluation module, the mission planning of the big data, services for generating according to mission planning module, perform cloud The assessment of Services Composition scheme, selects optimum cloud service assembled scheme, provides storage for disease self diagnosis service and calculates resource,

(4) diagnosing model generation module, for according to optimum cloud service assembled scheme, sets up electronic health record index, and Use big data analysing method to calculate and obtain disease self diagnosis model diagnosing model.

Disease self diagnosis service construction system based on cloud computing the most according to claim 1, is characterized in that, described data Resources gathering module includes modeling submodule, resources duplication submodule and the resource lookup submodule being sequentially connected with, described modeling The overlay network that resource node under cloud environment is formed by module for using Unstructured Peer-to-Peer Network is modeled, described resource Replicon module carries out the duplication of resource information, described resource lookup between each neighbor node in described overlay network Module meets the electronic health record data resource of application demand for lookup coupling.

Disease self diagnosis service construction system based on cloud computing the most according to claim 1, is characterized in that, described electronics Case history index include case history inverted index, case history filter index and case history details index, described case history inverted index for according to Family disease symptoms retrieves the case history identical with user's disease symptoms from the big data resource of electronic health record, and described case history filters index and uses Filtering the case history inconsistent with user's Sex, Age and age in the sex according to user and age, described case history details index is used for The detailed content of case history is retrieved from the big data resource of electronic health record.

Disease self diagnosis service construction system based on cloud computing the most according to claim 2, is characterized in that, described resource Replicon module uses when carrying out the duplication of resource information based on the data asset information Active Replication agreement between neighbor node, tool Body is: set x_iFor a peer node in Unstructured Peer-to-Peer Network, { x_i1,x_i2,…x_imIt is x_iNeighbor node collection,For Local resource pond,For neighbor node resource information pond, i ∈ [1, n], n are the sum that peer-to-peer network comprises node, and m represents neighbours The number of node, m < n, works as x_iWhen adding overlay network, by x_iWith { x_l1,x_l2,…x_lmSet up connection, x_iBasis furtherIn Information on services, create the duplication message of a resource information, and described duplication message be transmitted to all neighbor node x_lmCarry out Replicating, if any node in peer-to-peer network receives a duplication message, the number information according to replicating message judges whether Receiving described duplication message, if receiving, abandoning described duplication message, if receiving first, then according to the money replicating message Source information and node location information, updateIn content, and according to replicating the vital values of message, determine to forward or abandon described Replicating message, wherein, resource information needs periodically to synchronize between neighbor node；Described lookup also mates satisfied application need The electronic health record data resource asked, including: set initiation inquiry request M_jNode be x_j, at x_jNeighbor node set according to Probability p_jThe peer node that random choose goes out integrates as p_j×{x_j1,x_j2,…x_jm},j∈[1,n]；

When peer node x_iReceive x_jInquiry request M sent_jTime, checkWithIn whether contain and meet inquiry request M_jElectricity Sub-medical record data information, if so, according to described electronic health record data message and the position of electronic health record data message place peer node Confidence ceases, and creates the response message of inquiryAnd according to x_jPositional information, by described response messageReturn to x_j, so After by x_jVital values subtract 1, if x_jVital values be 0, abandon inquiry request M_jIf not, 0, use Q learning algorithm to calculate p_j×{x_j1,x_j2,…x_jmThe Q-value of each peer node in }, by inquiry request M_jIt is transmitted to p_j×{x_j1,x_j2,…x_jmIn }, Q-value is maximum Node, Probability p_jSpan when network is leisurely and carefree be (5,8], the span when network congestion be [0,3)；

Set the computing formula of Q-value as:

$Q_{new}=Q_{old}+{\mathrm{\&alpha;Q}}_{learn}+\mathrm{\&beta;}\&times;I\&lsqb;\frac{N_{x_{j\mathrm{\&mu;}}}\left(t\right)(T_{x_{j\mathrm{\&mu;}}}-{T^{\&prime;}}_{x_{j\mathrm{\&mu;}}})}{{T^{\&prime;}}_{x_{j\mathrm{\&mu;}}}\&times;T_{x_{j\mathrm{\&mu;}}}}\&rsqb;\&times;\frac{1+N_{x_{j\mathrm{\&mu;}}}\left(t\right)}{T_{x_{j\mathrm{\&mu;}}}}$

Wherein, Q_newRepresent the new value of Q, Q_oldRepresent the old value of Q, Q_learnRepresenting the value learnt, α represents learning rate, β represents congested factor,Represent moment t node x_jμBuffer queue in pending inquiry request message number,Represent p_j×{x_j1,x_j2,…x_jmNode x in }_jμProcess the time of an inquiry request message defined,Represent p_j× {x_j1,x_j2,…x_jmNode x in }_jμProcess the time that an inquiry request message is actually required；Function I [x] is at x > 0 time value Being 1, during x≤0, value is 0, and the span of α is [0.25,0.3], and the span of β is [0.45,0.5].

Disease self diagnosis service construction system based on cloud computing the most according to claim 1, is characterized in that, described credible Combined evaluation module includes assessing submodule and assessment optimizes submodule；The operation that described assessment submodule specifically performs is:

A, according to cloud service resource pool SP_vWith corresponding service qualityHistorical record, carries out the effectiveness of cloud service assembled scheme Each parameter of utility function in the modeling of function X initialization model, if the mission planning obtained by mission planning module G={G₁,G₂,G₃, correspondingIt is constrained to C={C₁,C₂,C₃, each subtask G_vCorresponding cloud service resource pool SP_vAltogether There is m_vIndividual service, for cloud service resource pool SP_vIn each service SP_vω, it comprisesHistorical record number is L_vω, By SP_vγ the feasible cloud service assembled scheme formed is CS_γ, v ∈ [1,3], ω ∈ [1, m_v], Definition Model is:

$\left\{\begin{array}{c}X\left({\mathrm{CS}}_{\mathrm{\&gamma;}}\right)={\mathrm{\&Sigma;}}_k^3\frac{Q_{O_{S\max }}\left(k\right)-{\mathrm{\&Sigma;}}_{v=1}^3{\mathrm{\&Sigma;}}_{\mathrm{\&omega;}=1}^{m_v}{\mathrm{\&Sigma;}}_{h=1}^{L_{v\mathrm{\&omega;}}}{q}_d\left({\mathrm{SP}}_{v\mathrm{\&omega;}}{R}_h\right)\&times;x_{v\mathrm{\&omega;}-h}}{Q_{O_{S\max }}\left(k\right)-Q_{O_{S\min }}\left(k\right)}\&times;w_k\\ {\mathrm{\&Sigma;}}_{\mathrm{\&omega;}=1}^{m_v}{\mathrm{\&Sigma;}}_{h=1}^{L_{v\mathrm{\&omega;}}}{q}_k\left({\mathrm{SP}}_{v\mathrm{\&omega;}}{R}_h\right)\&times;x_{v\mathrm{\&omega;}-h}\&le;C_k,1\&le;k\&le;3\\ {\mathrm{\&Sigma;}}_{\mathrm{\&omega;}=1}^{m_v}{\mathrm{\&Sigma;}}_{h=1}^{L_{v\mathrm{\&omega;}}}{x}_{v\mathrm{\&omega;}-h}\mathrm{=1,}{x}_{v\mathrm{\&omega;}-h}\&Element;\{0,1\}\\ {\mathrm{\&Sigma;}}_k^d{w}_k=1,w_k\&Element;\&lsqb;0,1\&rsqb;\end{array}\right.$

Wherein,For kth dimensionMaximum,For kth dimensionMinima, SP_vωR_hFor It is under the jurisdiction of SP_vωOneHistorical record, x_vω-hRepresent the parameter of utility function in model；

B, each feasible cloud service assembled scheme is ranked up by order from small to large according to utility function value, before selecting Z can Services Composition scheme of racking is set according to application example as preferred cloud service assembled scheme, the value of Z；

C, each group of preferred cloud service assembled scheme is calculated the meansigma methods of its utility function value；

D, the meansigma methods of selection utility function value are that maximum preferred cloud service assembled scheme is as optimum cloud service assembled scheme；

Described assessment optimizes submodule and is able to record that utility function value and the optimum combination cloud clothes of preferred cloud service assembled scheme Business scheme, and learn as sample, if new preferred cloud service assembled scheme had occurred, then directly invoke Its functional value.