CN112700067A - Method and system for predicting service quality under unreliable mobile edge environment - Google Patents
Method and system for predicting service quality under unreliable mobile edge environment Download PDFInfo
- Publication number
- CN112700067A CN112700067A CN202110050515.2A CN202110050515A CN112700067A CN 112700067 A CN112700067 A CN 112700067A CN 202110050515 A CN202110050515 A CN 202110050515A CN 112700067 A CN112700067 A CN 112700067A
- Authority
- CN
- China
- Prior art keywords
- service
- local
- user
- matrix
- potential
- 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
- 238000000034 method Methods 0.000 title claims abstract description 62
- 239000011159 matrix material Substances 0.000 claims abstract description 132
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000009826 distribution Methods 0.000 claims description 8
- 230000007246 mechanism Effects 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 230000004044 response Effects 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013501 data transformation Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method and a system for predicting service quality under an unreliable mobile edge environment, which comprise the following steps: preprocessing the numerical values in the data set collected by each user; each user constructs a local service quality matrix according to the preprocessed data, and each user decomposes the local service quality matrix into a local user potential matrix and a local service potential matrix; the central server collects the local service potential matrixes of all the users and integrates the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the inaccessible service according to the local user potential matrix and the global service potential matrix. The invention has accurate prediction and higher reliability.
Description
Technical Field
The present invention relates to the technical field of service quality prediction, and in particular, to a method and a system for predicting service quality in an unreliable mobile edge environment.
Background
With the rapid deployment and wide adoption of the 5G cellular network, many applications of the internet of things are appearing in various fields such as intelligent medical treatment, real-time entertainment, virtual reality, intelligent transportation, intelligent manufacturing and the like, and the applications are generally operated on User Equipment (UE), generally a smart phone and a tablet computer, and are used for connecting with equipment of the intelligent internet of things and relying on an MEC to access various remote cloud services. The real-time interaction capacity [1] of the application of the Internet of things not only needs the timely response of intelligent Internet of things equipment, but also puts high-quality requirements on remote cloud services, such as low response time, high throughput, high availability and the like.
To ensure the performance of internet of things applications, Service Oriented Architecture (SOA) has become a major architectural paradigm for discovering high quality services from a large number of publicly accessible candidate services, and combining the required services in a loosely coupled manner at design time and runtime. In service selection, quality of service is often a key indicator for identifying high quality services. QoS characterizes the non-functional attributes of a service including response time, throughput, availability, etc. A user-aware QoS value for a service may effectively improve the performance of service selection due to server-side factors (e.g., bandwidth, workload, etc.), user-side factors (e.g., UE capacity, geographical location, etc.), and the MEC environment between the user and the service (e.g., capacity, bandwidth, network congestion at edge nodes, etc.). However, it is very difficult to acquire the QoS values of all candidate services because each user only observes the QoS value of the invoked service. Active evaluation of these QoS values is also not feasible due to the limited computational power of the UE and the high cost of invoking a large number of services.
In the prior art, collaborative filtering techniques are used in predictive models to predict unknown QoS values. Fig. 1 shows a conventional QoS prediction framework. The user provides the local service usage record to the central server to form a global QoS data set. The central server builds a prediction model on this data set and provides QoS prediction services to the user to reduce the actual service calls required for the measurements. Unfortunately, unreliable MEC environments make these approaches impractical.
In practice, these methods are severely limited in efficiency, privacy and reliability: for efficiency: the user needs to transmit local QoS data to the central server. Due to the capacity and energy limitations of the UE in MEC [5], the user cannot afford to transmit large amounts of data efficiently and with overhead. In addition, training the predictive models on the server side requires processing large amounts of data in a short time, which places extremely high demands on the storage and computing power of the central server. It is a significant challenge for a central server to provide timely predictive services for all users at runtime.
For privacy: the remote central server may not be trusted. In order to obtain predictive service, the user needs to provide local usage data to the central server, which carries the risk of data leakage. The central server may deduce personal information from these data and even sell their data. Thus, some users may be reluctant to participate in the collaborative construction of predictive models, with a fear that their data may be compromised. Therefore, limited historical training data has a large impact on the prediction accuracy of the model. Therefore, QoS prediction methods need to be able to protect user privacy from untrusted central servers while producing accurate prediction results using as much user data as possible.
For reliability: the users in the MEC may not be trusted. Since the central server builds the prediction model by collecting and analyzing historical QoS data observed by the user, the reliability of the QoS data determines to a large extent the accuracy and confidence of the prediction results. However, many untrusted users in the MEC have malicious behavior in the internet of things environment. Some users submit random QoS values, which are noise data for the model. Some users submit constant QoS values that provide limited useful information for model building. Other malicious users manipulate the predictions of the model by submitting false QoS values, overestimating the performance of some services and underestimating the performance of other competing services. Therefore, a trusted QoS prediction approach needs to be able to make accurate predictions based on partially unreliable data to tolerate untrusted users in the MEC.
Disclosure of Invention
Therefore, the technical problem to be solved by the invention is to overcome the problems of high cost, high error rate, environmental pollution and low reaction speed in the prior art, so that the method and the system for predicting the service quality in the environment with accurate prediction and considering the reliability and privacy protection are provided.
In order to solve the above technical problem, a method for predicting service quality in an unreliable mobile edge environment according to the present invention includes: step S1: preprocessing the numerical values in the data set collected by each user; step S2: each user constructs a local service quality matrix according to the preprocessed data, and each user decomposes the local service quality matrix into a local user potential matrix and a local service potential matrix; step S3: the central server collects the local service potential matrixes of all the users and integrates the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the inaccessible service according to the local user potential matrix and the global service potential matrix.
In one embodiment of the present invention, the method for preprocessing the values in the collected data set comprises: and carrying out data conversion on the numerical values in the data set to enable the numerical values to be close to normal distribution.
In an embodiment of the present invention, a method for each user to decompose the local quality of service matrix into a local user potential matrix and a local service potential matrix includes: decomposing the local quality of service matrix by a local loss function.
In one embodiment of the present invention, the method for each user to predict the value of the quality of the unaccessed service according to the local user potential matrix and the global service potential matrix is as follows: each user obtains a local copy of the global service potential matrix from the central server, initializes a local user potential factor, and the inner product of the local user potential factor and the local user potential factor is a locally predicted service quality matrix, and then iteratively updates the local copy of the global service potential matrix and the local user potential factor.
In one embodiment of the present invention, the method for initializing the potential factors of the local users comprises: the local user latent factor is initialized with a small random number.
In one embodiment of the invention, when the central server collects the local service potential matrixes of all users, a user reputation mechanism is introduced to distinguish the credibility of different users.
In an embodiment of the present invention, after each user predicts the value of the quality of service not accessed according to the local user potential matrix and the global service potential matrix, for the service quality data newly observed locally by each user, returning to step S1, then updating the local service quality matrix according to the preprocessed data, updating the local user potential matrix and the local service potential matrix based on the updated service quality matrix, each user uploads the incremental update value of the local service potential matrix to the central server, the central server collects the incremental update of all the local service potential matrices, updates the global service potential matrix after integration, and transmits the updated global service potential matrix to each user; and finally, predicting the numerical value of the quality of the non-access service according to the local user potential matrix and the global service potential matrix.
In one embodiment of the invention, when the central server collects the incremental updates of all the local service potential matrixes, a user reputation mechanism is introduced to distinguish the credibility of different users.
In one embodiment of the invention, when the central server collects the local service potential matrixes of all the users, the gradient parameters transmitted between the users and the central server are encrypted and decrypted by adopting a differential privacy technology.
The invention also provides a system for predicting the service quality in the unreliable mobile edge environment, which is characterized by comprising the following steps: the preprocessing module is used for preprocessing the numerical values in the data sets collected by each user; the building module is used for building a local service quality matrix according to the preprocessed data by each user, and decomposing the local service quality matrix into a local user potential matrix and a local service potential matrix by each user; the prediction module is used for collecting the local service potential matrixes of all the users by the central server and integrating the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the non-accessed service according to the local user potential matrix and the global service potential matrix.
Compared with the prior art, the technical scheme of the invention has the following advantages:
according to the method and the system for predicting the service quality under the unreliable mobile edge environment, in the aspect of prediction accuracy, privacy is protected, and meanwhile, the method has higher accuracy, and with the increase of data density, although all methods are higher and higher in prediction accuracy (because more training data contain more useful information), the data amount needing to be processed is increased, and the method provided by the invention is more suitable for high data density; from a reliability aspect, the accuracy of all methods decreases as the percentage of untrusted users increases, because less reliable data means less useful information and more noisy, and at different rates of untrusted users, the method proposed by the present application is able to retrieve reliable information from partially unreliable data, while at the same time taking into account privacy protection issues; in the aspect of prediction efficiency, the training tasks of the models are distributed to all users and the central server, each user only needs to update the local model according to a small amount of local data, and the central server only simply aggregates the updates of the users, so that the time for generating the prediction result by the method is short.
Drawings
In order that the present disclosure may be more readily and clearly understood, reference is now made to the following detailed description of the embodiments of the present disclosure taken in conjunction with the accompanying drawings, in which
FIG. 1 is a system framework diagram of existing predictions;
fig. 2 is a flow chart of a method for predicting the service quality in an unreliable mobile edge environment according to the present invention.
Detailed Description
Example one
As shown in fig. 2, the present embodiment provides a method for predicting quality of service in an unreliable mobile edge environment, including the following steps: step S1: preprocessing the numerical values in the data set collected by each user; step S2: each user constructs a local service quality matrix according to the preprocessed data, and each user decomposes the local service quality matrix into a local user potential matrix and a local service potential matrix; step S3: the central server collects the local service potential matrixes of all the users and integrates the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the inaccessible service according to the local user potential matrix and the global service potential matrix.
In the method for predicting the service quality in the unreliable mobile edge environment according to this embodiment, in step S1, a numerical value in a data set collected by each user is preprocessed, so that the prediction accuracy is favorably ensured; in the step S2, each user constructs a local quality of service matrix according to the preprocessed data, and each user decomposes the local quality of service matrix into a local user potential matrix and a local service potential matrix, so that the privacy problem of the user can be effectively protected through decomposition; in step S3, the central server collects the local service potential matrices of all users and integrates them into a global service potential matrix, the central server sends the global service potential matrix to all users, and each user predicts the numerical value of the quality of service not accessed according to the local user potential matrix and the global service potential matrix, so that not only the prediction result is accurate, but also the user stores the original quality of service data locally without being exposed to others, thereby protecting the user from privacy threats and improving reliability and privacy protection.
In step S1, the method for preprocessing the values in the collected data set includes: and carrying out data conversion on the numerical values in the data set to enable the numerical values to be close to normal distribution.
Since Response Time (RT) and Throughput (TP) are two typical quality of service attributes, these two distributions are characterized by high bias and large variance. However, the probability assumption of matrix decomposition requires an approximate normal distribution, and the value ranges of different qos attributes are very different, which greatly affects the prediction accuracy. All data were preprocessed using Box-Cox transforms to approximate a normal distribution by performing data transformations as follows:
where alpha controls the degree of transition. Let q bemaxAnd q isminThe upper and lower bounds of the QoS value are respectively specified by the user. Since b (x) is a monotonically non-decreasing function, b (q)max) And b (q)min) Is the upper and lower bounds after data conversion; the QoS value may then be mapped to [0, 1] by the following translation]In the interval:
predicted result qijCan be mapped to [0, 1] by a logistic function]In the interval:
in step S2, the method for each user to decompose the local quality of service matrix into a local user potential matrix and a local service potential matrix includes: decomposing the local quality of service matrix by a local loss function.
The local loss function is as follows:
wherein L ═ ΣiLi. Instead of minimizing the joint model, the local loss function L is loosely minimized at each customer sitei。UiAnd SiUpdating the local QoS matrix Q by an iterative processiUntil convergence:
Each user UiObtaining a local copy S of a global service potential matrix from a central serveri∈R1×nLocal user latent factor Ui∈R1×1Initialized with a small random number, the user can locally predict the quality of service value of an inaccessible service by calculating the inner product of the user potential vector and the corresponding service potential vector:
in step S3, the method for predicting the value of the quality of the unaccessed service by each user according to the local user potential matrix and the global service potential matrix includes: each user obtains a local copy of the global service potential matrix from the central server, initializes a local user potential factor, and the inner product of the local user potential factor and the local user potential factor is a locally predicted service quality matrix, and then iteratively updates the local copy of the global service potential matrix and the local user potential factor.
The method for initializing the potential factors of the local users comprises the following steps: the local user latent factor is initialized with a small random number.
And when the central server collects the local service potential matrixes of all the users, a user reputation mechanism is introduced to distinguish the credibility of different users.
The method and the device introduce a user reputation mechanism to distinguish the credibility of different users. Using ri∈[0,1]To represent the user UiDefinition of user UiReliability of the submitted data. r isi11 represents UiComplete confidence, ri<1 denotes a moiety UiAnd (4) credibility.
R is calculated using the L1-AVG algorithmiThe following were used:
where d ∈ (0,1) is the decay constant, ajIs a service sjThe reputation weighted average derivative of. r isiAnd aiThe calculation is repeated, k representing the kth iteration. The present application initializes all usersBy using Representing r in two iterationsiδ is the error threshold. When max (Δ r)i)<Delta, the iterative process stops. Intuitively, the reputation of one user is evaluated by the difference between the submitted derivative value and the average derivative value of the other users.
After predicting the numerical value of the quality of service which is not accessed by each user according to the local user potential matrix and the global service potential matrix, returning to the step S1 for the service quality data which is newly observed locally by each user, updating the local service quality matrix according to the preprocessed data, updating the local user potential matrix and the local service potential matrix based on the updated service quality matrix, uploading the increment updating value of the local service potential matrix to the central server by each user, collecting the increment updating of all the local service potential matrices by the central server, updating the global service potential matrix after the increment updating and the integration of all the local service potential matrices are integrated, and transmitting the updated global service potential matrix to each user by the central server; and finally, predicting the numerical value of the quality of the non-access service according to the local user potential matrix and the global service potential matrix.
The method for updating the local user potential matrix comprises the following steps: updating local user latent factor Ui(ii) a The method for updating the global service potential matrix comprises the following steps: updating a local copy S of a global service potential matrixi: wherein
The method for updating the global service potential matrix comprises the following steps: updating a global service potential matrix Sj:
In each round, the user stores the original quality of service data locally and is not exposed to other people, so that the user is protected from privacy threats.
The central server also needs to introduce a user reputation mechanism to distinguish the trustworthiness of different users when collecting the incremental updates of all the local service potential matrices. As already discussed above, this will not be described in detail.
And when the central server collects the local service potential matrixes of all the users and the central server collects the incremental updates of all the local service potential matrixes, encrypting and decrypting the gradient parameters transmitted between the users and the central server by adopting a differential privacy technology.
For simplicity, the application will hereinafter denote the gradient with g, a key consisting of two large prime numbers being generated and distributed to each user. Let x and y be two prime numbers, N-xy be a public parameter, and epsilon be a privacy budget. For each user UiNoise sampled randomly from a Laplace distributionThe procedure for adding to the local gradient is as follows:
in the above equation, Δ f represents sensitivity and controls privacy budget. Smaller dots represent higher privacy levels. The calculation formula of the encryption gradient is as follows:
wherein y is-1And x-1The inverse of y and x, respectively. From each user UiReceive EiThereafter, the central server performs global aggregation in the following manner:
at the beginning of each round, each user UiA local copy of the newly encrypted global gradient E is received and decrypted by:
g≡E mod x,
g≡E mod y,
g≈∑igiis an unbiased estimate of the global gradient.
The following is an evaluation of the proposed credible and privacy preserving prediction method on a real Web service's service quality data set.
The data set contains the quality of service values of 4532 services invoked by 142 users over 64 consecutive time slices, each service lasting 15 minutes. The data conversion is carried out by adjusting alpha to-0.007 for RT and alpha to-0.005 for TP respectively, so that the data conversion is closer to normal distribution. Under each density, the method repeats experiments for 20 times by using different random seeds, and calculates the average prediction precision of each method. In order to understand the influence of epsilon-difference privacy on the prediction precision, experiments are carried out under different privacy levels. Experiments show that when the epsilon value is small, the prediction precision is greatly influenced by noise. As epsilon increases, the prediction becomes more and more accurate because there is less and less noise affecting the model. When epsilon is larger than 0.8, the method provided by the application can realize continuous high prediction precision. In order to evaluate the reliability of the service quality prediction method in the unreliable MEC environment, the percentage of the unreliable users is set from 5% to 30%, and the step length is 5%.
Example two
Based on the same inventive concept, the present embodiment provides a system for predicting service quality in an unreliable mobile edge environment, and the principle of solving the problem is similar to the method for predicting service quality in an unreliable mobile edge environment, and repeated parts are not repeated.
The embodiment provides a system for predicting service quality in an unreliable mobile edge environment, which includes:
the preprocessing module is used for preprocessing the numerical values in the data sets collected by each user;
the building module is used for building a local service quality matrix according to the preprocessed data by each user, and decomposing the local service quality matrix into a local user potential matrix and a local service potential matrix by each user;
the prediction module is used for collecting the local service potential matrixes of all the users by the central server and integrating the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the non-accessed service according to the local user potential matrix and the global service potential matrix.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should be understood that the above examples are only for clarity of illustration and are not intended to limit the embodiments. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. And obvious variations or modifications therefrom are within the scope of the invention.
Claims (10)
1. A method for predicting service quality in an unreliable mobile edge environment is characterized by comprising the following steps:
step S1: preprocessing the numerical values in the data set collected by each user;
step S2: each user constructs a local service quality matrix according to the preprocessed data, and each user decomposes the local service quality matrix into a local user potential matrix and a local service potential matrix;
step S3: the central server collects the local service potential matrixes of all the users and integrates the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the inaccessible service according to the local user potential matrix and the global service potential matrix.
2. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: the method for preprocessing the numerical values in the collected data set comprises the following steps: and carrying out data conversion on the numerical values in the data set to enable the numerical values to be close to normal distribution.
3. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: the method for decomposing the local service quality matrix into a local user potential matrix and a local service potential matrix by each user comprises the following steps: decomposing the local quality of service matrix by a local loss function.
4. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: the method for predicting the numerical value of the quality of the non-access service by each user according to the local user potential matrix and the global service potential matrix comprises the following steps: each user obtains a local copy of the global service potential matrix from the central server, initializes a local user potential factor, and the inner product of the local user potential factor and the local user potential factor is a locally predicted service quality matrix, and then iteratively updates the local copy of the global service potential matrix and the local user potential factor.
5. The method of claim 4, wherein the method comprises: the method for initializing the potential factors of the local users comprises the following steps: the local user latent factor is initialized with a small random number.
6. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: and when the central server collects the local service potential matrixes of all the users, a user reputation mechanism is introduced to distinguish the credibility of different users.
7. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: after predicting the numerical value of the quality of service which is not accessed by each user according to the local user potential matrix and the global service potential matrix, returning to the step S1 for the service quality data which is newly observed locally by each user, updating the local service quality matrix according to the preprocessed data, updating the local user potential matrix and the local service potential matrix based on the updated service quality matrix, uploading the increment updating value of the local service potential matrix to the central server by each user, collecting the increment updating of all the local service potential matrices by the central server, updating the global service potential matrix after the increment updating and the integration of all the local service potential matrices are integrated, and transmitting the updated global service potential matrix to each user by the central server; and finally, predicting the numerical value of the quality of the non-access service according to the local user potential matrix and the global service potential matrix.
8. The method of claim 7, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: and when the central server collects the incremental updates of all the local service potential matrixes, a user reputation mechanism is introduced to distinguish the credibility of different users.
9. The method of claim 1, wherein the method for predicting the quality of service in the unreliable mobile edge environment comprises: and when the central server collects the local service potential matrixes of all the users, the gradient parameters transmitted between the users and the central server are encrypted and decrypted by adopting a differential privacy technology.
10. A system for predicting quality of service in an unreliable mobile edge environment, comprising:
the preprocessing module is used for preprocessing the numerical values in the data sets collected by each user;
the building module is used for building a local service quality matrix according to the preprocessed data by each user, and decomposing the local service quality matrix into a local user potential matrix and a local service potential matrix by each user;
the prediction module is used for collecting the local service potential matrixes of all the users by the central server and integrating the local service potential matrixes into a global service potential matrix, the central server sends the global service potential matrix to all the users, and each user predicts the numerical value of the quality of the non-accessed service according to the local user potential matrix and the global service potential matrix.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110050515.2A CN112700067B (en) | 2021-01-14 | 2021-01-14 | Method and system for predicting service quality in unreliable mobile edge environment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110050515.2A CN112700067B (en) | 2021-01-14 | 2021-01-14 | Method and system for predicting service quality in unreliable mobile edge environment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112700067A true CN112700067A (en) | 2021-04-23 |
CN112700067B CN112700067B (en) | 2024-05-24 |
Family
ID=75515101
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110050515.2A Active CN112700067B (en) | 2021-01-14 | 2021-01-14 | Method and system for predicting service quality in unreliable mobile edge environment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112700067B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114489938A (en) * | 2022-01-05 | 2022-05-13 | 汕头大学 | Method for constructing user side QoS prediction model based on cloud edge collaborative mode |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050226193A1 (en) * | 2004-03-31 | 2005-10-13 | Nokia Corporation | Monitoring quality of service in a wireless communications network |
CN101252451A (en) * | 2008-03-22 | 2008-08-27 | 桂林电子科技大学 | Method and middleware for Web server system requesting to distinguish service ensuring service quality |
CN103746957A (en) * | 2013-10-10 | 2014-04-23 | 安徽师范大学 | Trust evaluation system based on privacy protection and construction method thereof |
CN107679415A (en) * | 2017-09-25 | 2018-02-09 | 深圳大学 | Secret protection cooperates with the collaborative filtering method based on model of Web service prediction of quality |
CN108563660A (en) * | 2017-12-29 | 2018-09-21 | 温州大学 | service recommendation method, system and server |
CN109257217A (en) * | 2018-09-19 | 2019-01-22 | 河海大学 | Web service QoS prediction technique based on secret protection under mobile peripheral surroundings |
CN109376901A (en) * | 2018-09-07 | 2019-02-22 | 苏州大学 | A kind of service quality prediction technique based on decentralization matrix decomposition |
WO2019056573A1 (en) * | 2017-09-25 | 2019-03-28 | 深圳大学 | Differential privacy-based system and method for collaborative web quality-of-service prediction for privacy protection |
US20190158606A1 (en) * | 2018-12-28 | 2019-05-23 | Francesc Guim Bernat | QUALITY OF SERVICE (QoS) MANAGEMENT IN EDGE COMPUTING ENVIRONMENTS |
CN110825955A (en) * | 2019-06-27 | 2020-02-21 | 安徽师范大学 | Distributed differential privacy recommendation method based on location based service |
CN110868326A (en) * | 2019-11-27 | 2020-03-06 | 武汉虹信通信技术有限责任公司 | Network service quality analysis method, edge device and central server |
-
2021
- 2021-01-14 CN CN202110050515.2A patent/CN112700067B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050226193A1 (en) * | 2004-03-31 | 2005-10-13 | Nokia Corporation | Monitoring quality of service in a wireless communications network |
CN101252451A (en) * | 2008-03-22 | 2008-08-27 | 桂林电子科技大学 | Method and middleware for Web server system requesting to distinguish service ensuring service quality |
CN103746957A (en) * | 2013-10-10 | 2014-04-23 | 安徽师范大学 | Trust evaluation system based on privacy protection and construction method thereof |
CN107679415A (en) * | 2017-09-25 | 2018-02-09 | 深圳大学 | Secret protection cooperates with the collaborative filtering method based on model of Web service prediction of quality |
WO2019056573A1 (en) * | 2017-09-25 | 2019-03-28 | 深圳大学 | Differential privacy-based system and method for collaborative web quality-of-service prediction for privacy protection |
WO2019056572A1 (en) * | 2017-09-25 | 2019-03-28 | 深圳大学 | Model-based collaborative filtering method for collaborative web quality-of-service prediction for privacy protection |
CN108563660A (en) * | 2017-12-29 | 2018-09-21 | 温州大学 | service recommendation method, system and server |
CN109376901A (en) * | 2018-09-07 | 2019-02-22 | 苏州大学 | A kind of service quality prediction technique based on decentralization matrix decomposition |
CN109257217A (en) * | 2018-09-19 | 2019-01-22 | 河海大学 | Web service QoS prediction technique based on secret protection under mobile peripheral surroundings |
US20190158606A1 (en) * | 2018-12-28 | 2019-05-23 | Francesc Guim Bernat | QUALITY OF SERVICE (QoS) MANAGEMENT IN EDGE COMPUTING ENVIRONMENTS |
CN110825955A (en) * | 2019-06-27 | 2020-02-21 | 安徽师范大学 | Distributed differential privacy recommendation method based on location based service |
CN110868326A (en) * | 2019-11-27 | 2020-03-06 | 武汉虹信通信技术有限责任公司 | Network service quality analysis method, edge device and central server |
Non-Patent Citations (3)
Title |
---|
冯博昊: "移动边缘计算场景下网络服务灵活适配的解决方案", 《电信科学》, 18 March 2019 (2019-03-18) * |
宁雪莉: "基于地理社交网络的频繁位置隐私保护算法", 《计算机应用》, vol. 38, no. 3, 10 March 2018 (2018-03-10) * |
张鹏程;金惠颖;: "一种移动边缘环境下面向隐私保护QoS预测方法", 计算机学报, no. 08 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114489938A (en) * | 2022-01-05 | 2022-05-13 | 汕头大学 | Method for constructing user side QoS prediction model based on cloud edge collaborative mode |
Also Published As
Publication number | Publication date |
---|---|
CN112700067B (en) | 2024-05-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Nguyen et al. | Privacy-preserved task offloading in mobile blockchain with deep reinforcement learning | |
Zheng et al. | Web service QoS prediction via collaborative filtering: A survey | |
Lin et al. | Using gans for sharing networked time series data: Challenges, initial promise, and open questions | |
KR102208188B1 (en) | A distributed multi-party security model training framework for privacy protection | |
KR20200088767A (en) | Logistic regression modeling method using secret sharing | |
Karuppusamy et al. | Chronological salp swarm algorithm based deep belief network for intrusion detection in cloud using fuzzy entropy | |
CN113505882B (en) | Data processing method based on federal neural network model, related equipment and medium | |
EP3195560A1 (en) | Lateral movement detection | |
US10067746B1 (en) | Approximate random number generator by empirical cumulative distribution function | |
CN112910710B (en) | Network flow space-time prediction method and device, computer equipment and storage medium | |
CN114696990B (en) | Multi-party computing method, system and related equipment based on fully homomorphic encryption | |
US20230034384A1 (en) | Privacy preserving machine learning via gradient boosting | |
CN102272728B (en) | Method, apparatus, and computer program product for polynomial-based data transformation and utilization | |
Akter et al. | Edge intelligence-based privacy protection framework for iot-based smart healthcare systems | |
Zhang et al. | Towards efficient, credible and privacy-preserving service QoS prediction in unreliable mobile edge environments | |
CN117540106B (en) | Social activity recommendation method and device for protecting multi-mode data privacy | |
CN112700067B (en) | Method and system for predicting service quality in unreliable mobile edge environment | |
Bahrami et al. | CloudPDB: A light-weight data privacy schema for cloud-based databases | |
Liu et al. | Privacy-preserving collaborative web services QoS prediction via differential privacy | |
Lecuyer et al. | Pyramid: Enhancing selectivity in big data protection with count featurization | |
CN111865899A (en) | Threat-driven cooperative acquisition method and device | |
Nafi et al. | A new trusted and secured E-commerce architeture for cloud computing | |
Miehling et al. | Control-theoretic approaches to cyber-security | |
Takeshita et al. | Slap: simpler, improved private stream aggregation from ring learning with errors | |
Aishwarya et al. | AI Strategy for Stake Cloud Computing and Edge Computing: A State of the art survey |
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 |