CN108989092B - Wireless network prediction method, electronic equipment and storage medium - Google Patents
Wireless network prediction method, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN108989092B CN108989092B CN201810671779.8A CN201810671779A CN108989092B CN 108989092 B CN108989092 B CN 108989092B CN 201810671779 A CN201810671779 A CN 201810671779A CN 108989092 B CN108989092 B CN 108989092B
- Authority
- CN
- China
- Prior art keywords
- predicted value
- value
- variables
- unknown parameter
- data vector
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a wireless network prediction method, which comprises the following steps: selecting a plurality of network communication index variables and corresponding target index variables to establish a Gaussian process regression model, determining a covariance function in a Gaussian kernel function form, solving reasonable values of all parameters in the covariance function, setting a first to-be-predicted value at any time point and a first section data vector of a required independent variable to obtain a corresponding joint distribution function so as to calculate an estimated value of the first to-be-predicted value; and setting a second to-be-predicted value in any time period and a second section data vector of the required independent variable, and calculating to obtain an estimated value of the second to-be-predicted value. The method provided by the invention utilizes a Gaussian process regression method to regularly mine the historically accumulated live network data, construct a data model, effectively predict the change condition of the target network index in a certain period of time in the future, and provide more effective data reference for optimizing wireless network resource allocation and performance optimization.
Description
Technical Field
The present invention relates to wireless network communication technologies, and in particular, to a wireless network prediction method, an electronic device, and a storage medium.
Background
At present, with the rapid development of communication information technology, a wireless communication network gradually realizes large-scale popularization, and the dependence of users on wireless network requirements in daily life is gradually enhanced. How to accurately and effectively predict the vigorous use degree of users in different areas on a wireless network, avoid the situations of network blockage and the like, effectively improve the resource allocation efficiency of the wireless network, and is one of key contents of daily operation and maintenance optimization of communication operators so as to further improve the satisfaction degree of the users on the use of the wireless network. The current analysis of the wireless network using prosperity is usually found by analysis such as user's postnatal feedback or simple network index statistics, and the existing wireless network demand analysis method is mainly based on simple one-dimensional linear analysis among indexes, and cannot effectively synthesize multidimensional data to predict and feed back wireless network demands in a future period of time in advance
However, the existing analysis of the prosperity of the use of the wireless network has the following defects:
(1) The conventional methods such as a linear prediction method and a neural network have poor practicability, the prediction result is unstable under a small sample, and different network demand index predictions such as telephone traffic and GPRS service conditions cannot be well compatible;
(2) The change rule of the wireless network demand in a future period of time in a certain area cannot be effectively predicted;
(3) The prediction of the related indexes of the wireless network cannot be carried out by integrating multiple factors.
Disclosure of Invention
In order to overcome the disadvantages of the prior art, an object of the present invention is to provide a wireless network prediction method, which can more effectively predict the wireless network requirements in a future time period according to historical data.
It is another object of the present invention to provide an electronic device that more effectively predicts the wireless network demand in a time period in the future based on historical data.
It is a further object of the present invention to provide a computer storage medium that more effectively predicts wireless network demand over a period of time in the future based on historical data.
One of the purposes of the invention is realized by adopting the following technical scheme:
a wireless network prediction method comprises the following steps:
a data selection step: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
a function determination step: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
a parameter determination step: solving the reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
time point model prediction step: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, the dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
a time period model prediction step: and setting a second to-be-predicted value in any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function.
Further, the method also comprises the following steps:
and (3) error analysis: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
Further, in the data selecting step, the independent variables are expressed as: x n×s ={x 1 ,x 2 ,…x s Expressed as Y = { Y = dependent variable 1 ,y 2 ,…y n In which the vector X in the argument i ={x i1 ,x i2 ,…x is I =1,2, \ 8230;, m represents a cross-sectional dataset of independent variables, n in dependent variables is the dataset sequence length, i.e., the data volume, and the gaussian process regression model isWhereinIs a gaussian random noise process and is characterized by that,Z=(2π) n/2 |K| 1/2 μ is the mean vector of the independent variables; k is a covariance matrix, and
further, the covariance function is:wherein x i ,x j Respectively is the ith section data vector of the independent variable index and the ith section data vector of the independent variable index, s is the number of the input independent variables, sigma f Is a first unknown parameter, l is a second unknown parameter, σ n Is the third unknown parameter.
Further, in the parameter determination step, the maximum stomach value satisfying the condition of P (θ | X, Y) is selected, and the likelihood of P (θ | X, Y) is maximized by the conjugate gradient method to obtain a reasonable value of the stomach, where θ = { l, σ = f ,σ n }。
Further, in the time point model prediction step, the first value to be predicted is set asThe first section data vector of the argument required by the first value to be predicted isThen the joint distribution function isNamely, it isAccording to the formulaThe mean value of the medium probability distribution obtains the estimated value of the first value to be predicted asWherein the content of the first and second substances,
further, in the time period model prediction step, a second value to be predicted is set as The second cross-sectional data vector isMixing X * Step-by-step inputAnd calculating to obtain an estimation value of the second value to be predicted.
The second purpose of the invention is realized by adopting the following technical scheme:
an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
a data selection step: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
a function determination step: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
a parameter determination step: solving reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
a time point model prediction step: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, the dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
a time section model prediction step: and setting a second to-be-predicted value of any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function.
Preferably, after the time period model predicting step is executed, the method further includes the following steps: and (3) error analysis: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
The third purpose of the invention is realized by adopting the following technical scheme:
a computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out a wireless network prediction method according to one of the objects of the invention.
Compared with the prior art, the invention has the beneficial effects that:
the method provided by the invention utilizes a Gaussian process regression method to regularly mine the historically accumulated live network data, construct a data model, effectively predict the change condition of the target network index in a certain period of time in the future, and provide more effective data reference for optimizing wireless network resource allocation and performance optimization.
Drawings
Fig. 1 is a flowchart of a wireless network prediction method according to the present invention.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and the detailed description, and it should be noted that any combination of the embodiments or technical features described below can be used to form a new embodiment without conflict.
As shown in fig. 1, the present invention provides a wireless network prediction method based on a gaussian process regression model, which specifically includes the following steps:
s1: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
in the step, s network communication index variables are selected as influence factors (independent variables) of a prediction model for model training, target index variables Y (dependent variables) corresponding to the influence factors are selected, and a model independent variable data set is represented as X n×s ={x 1 ,x 2 ,…x s In which vector X i ={x i1 ,x i2 ,…x is I =1,2, \8230;, m represents a cross-sectional dataset of independent variables, n × s represents X with a first dimension of size n, a second dimension of size s, m is the largest dimension of the independent variables, and the model dependent variable vector is represented as Y = { Y = } 1 ,y 2 ,…y n Where n is the data set sequence length, i.e. the data volume.
The assumption of the Gaussian process regression model in the method isWhereinFor a Gaussian random noise process, f (x) is the joint probability distribution density function f (x) to GP (0, K) of the random variable of the Gaussian process, i.e.Z=(2π) n/2 |K| 1/2 μ is the mean vector of the independent variables; k is a covariance matrix, and
where f (χ) is the random variable joint probability distribution of the Gaussian processDensity function, X is input feature, T is matrix transposition operation, k (X) 1 ,X 1 ) Is a Gaussian kernel operation, and Z is an intermediate value.
Selecting time sequence historical data of a plurality of index variables collected by a base station of a certain city by taking the certain city as a unit, for example, if the change condition of the telephone traffic in a future week of a certain region needs to be predicted, taking the whole network, network elements or cells as region granularity and selecting independent variable indexes such as TCH/SDCCH congestion rate, TCH/SDCCH call completing rate, TCH/SDCCH call drop rate, cut-out success rate, cut-in success rate and the like with the granularity of 5 minutes as training data.
In order to avoid the influence of the index dimension of the independent variable or the dependent variable input by the model, the input data set of the model needs to be normalized, and the formula is as follows:in the formula: min (x) i ) And max (x) i ) Minimum and maximum of the sequence, x, respectively i Is the original sequence value.
S2: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
the covariance function can be obtained according to the covariance matrix, and the covariance function is expressed in the form of a Gaussian kernel functionWherein x is 1 ,x j Respectively is the ith section data vector of the independent variable index and the ith section data vector of the independent variable index, s is the number of the input independent variables, sigma f Is a first unknown parameter, l is a second unknown parameter, σ n As a third unknown parameter, δ (X) i ,X j ) Is X i And X j Of the measured data.
S3: solving the reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
in this step, since Gauss is used in the above processThe equation assumes that the mean of the normal distribution is zero, so the final model result has a large relationship with the selection of covariance, and the parameters involved in the above steps are θ = { l, σ = f ,σ n And b, selecting the maximum stomach value under the condition of meeting the P (theta | X, Y) by using a commonly used parameter estimation method for the parameter needing to be learned as maximum posterior estimation, but since no prior knowledge is given to the stomach, the P (theta | X, Y) is degenerated to the maximum likelihood P (Y | X, theta) at the moment, and the following formula holds: by maximizing the likelihood by a conjugate gradient method, a reasonable value corresponding to the stomach can be obtained.
The conjugate gradient method is calculated as follows:
a, initializing iteration steps: t =0, initial values are given: theta.theta. (0) Giving a positive definite matrix Q; b, calculating the derivative of f (theta) of formula 3, andif g is (0) =0, when the function maximum condition has been satisfied, the iterative computation is stopped. Otherwise, the optimal value searching direction adopts the steepest descent method, namely the function f (theta) is in theta (0) In the negative direction of the gradient, i.e. d (0) =-g (0) (ii) a c, calculating an iteration step size:the calculation yields the next iteration point: theta (t+1) =θ (t) +α t d (t) (ii) a e, calculatingIf g is (t+1) =0, the iterative calculation is stopped, and in t +1 iterationsSo that d (t+1) And d (0) ,d (1) …d (t) Forming a Q conjugate direction; f, then in t +1 iterations, d is calculated (t+1) =-g (t+1) +β t d (t) (ii) a g, letting t = t +1, returning to the step c), and iteratively repeating the calculation; h until a convergence condition threshold (ε =1 e-6) or g is reached (t+1) =0, obtained θ (t+1) The value is a current reasonable estimate.
S4: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, a dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
setting the first value to be predicted asThe first section data vector of the argument required by the first to-be-predicted value isDefining a plurality of network communication index variables (independent variables) and target index variables (dependent variables) corresponding to the network communication index variables as training data, wherein the first to-be-predicted value in the step is actually also a dependent variable, the first to-be-predicted value and the first section data vector are prediction data, and according to the characteristic that the training data and the prediction data are from the same data sample according to the nature of a Gaussian process, obtaining the Gaussian distribution with a high-dimensional joint distribution function of the training data and the prediction data, wherein the joint distribution function isNamely thatAccording to the formulaThe mean value of the medium probability distribution obtains the estimated value of the first value to be predicted asWherein, the first and the second end of the pipe are connected with each other, is a mean value of 0 and a variance ofA gaussian distribution of (a).
S5: and setting a second to-be-predicted value of any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function.
Setting the second value to be predicted asThe second cross-sectional data vector isX is to be * Step-by-step inputAnd calculating to obtain an estimation value of a second value to be predicted.
S6: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
Aiming at the model prediction result, the estimation value and the actual value of the predicted value are subjected to error analysis by respectively adopting two measurement indexes of standard deviation and relative error, wherein the standard deviationRelative errorThe RMRE, smaller MAPE values indicate more optimal prediction results.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:
a data selection step: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
a function determination step: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
a parameter determination step: solving reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
time point model prediction step: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, the dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
a time section model prediction step: setting a second to-be-predicted value in any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function;
and (3) error analysis: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
The present invention may also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a wireless network prediction method of the present invention.
In the actual operation process of the model framework constructed by the method, a network optimization engineer can define and input model data indexes with different region granularities and different time spans according to the prediction requirement to carry out model training and prediction analysis, and the model effectively predicts a certain period of time in the future, thereby being beneficial to the advance planning in the network operation and maintenance process to the great extent and formulating an effective resource allocation strategy.
The above embodiments are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby, and any insubstantial changes and substitutions made by those skilled in the art based on the present invention are within the protection scope of the present invention.
Claims (10)
1. A wireless network prediction method is characterized by comprising the following steps:
a data selection step: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
a function determining step: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
a parameter determination step: solving reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
time point model prediction step: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, a dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
a time section model prediction step: and setting a second to-be-predicted value in any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function.
2. The wireless network prediction method of claim 1, further comprising the steps of:
and (3) error analysis: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
3. The wireless network prediction method of claim 1, wherein in the data selection step, the independent variables are expressed as: x n×s ={x 1 ,x 2 ,...x s Denoted by Y = { Y }, dependent variable 1 ,y 2 ,...y n Where the vector X in the argument i ={x i1 ,x i2 ,...x is Where m denotes a cross-sectional dataset of independent variables, n × s denotes X with a first dimension of n, a second dimension of s, m being the largest dimension of the independent variables, n in dependent variables being the dataset sequence length, i.e., the data volume, and the gaussian process regression model isWhereinIs a gaussian random noise process and is characterized by that,Z=(2π) n/2 |K| 1/2 mu is the mean vector of the independent variables; k is a covariance matrix, and
4. As claimed in claim 3The wireless network prediction method is characterized in that the covariance function is as follows:wherein x is i ,x j Respectively is the ith section data vector of the independent variable index and the ith section data vector of the independent variable index, s is the number of the input independent variables, sigma f Is a first unknown parameter, l is a second unknown parameter, σ n As a third unknown parameter, δ (X) i ,X j ) Is X i And X j The covariance of (a).
5. The wireless network prediction method of claim 4, wherein in the parameter determination step, the maximum θ value satisfying the condition of P (θ | X, Y) is selected, and the likelihood is maximized for P (θ | X, Y) by the conjugate gradient method to obtain the value of θ, wherein θ = { l, σ = f ,σ n And theta is a parameter needing to be learned.
6. The method of claim 5, wherein the time model predicting step sets the first value to be predicted asThe first section data vector of the argument required by the first to-be-predicted value isThen the joint distribution function isNamely thatAccording to the formulaObtaining the first waiting probability distribution mean valueEstimate of the predicted value isWherein, the first and the second end of the pipe are connected with each other,is a mean value of 0 and a variance ofA gaussian distribution of (a).
8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:
a data selection step: selecting a plurality of network communication index variables as independent variables, selecting target index variables corresponding to the independent variables respectively as dependent variables, and establishing a Gaussian process regression model of the independent variables according to the independent variables;
a function determination step: determining a covariance function in a Gaussian kernel function form according to the Gaussian process regression model, wherein the covariance function comprises a first unknown parameter, a second unknown parameter and a third unknown parameter;
a parameter determination step: solving reasonable values of the first unknown parameter, the second unknown parameter and the third unknown parameter;
time point model prediction step: setting a first to-be-predicted value at any time point and a first section data vector of an independent variable required by the first to-be-predicted value, and acquiring a corresponding joint distribution function according to the independent variable, a dependent variable, the first to-be-predicted value and the first section data vector to calculate an estimated value of the first to-be-predicted value;
a time section model prediction step: and setting a second to-be-predicted value in any time period and a second section data vector of an independent variable required by the second to-be-predicted value, and calculating to obtain an estimated value of the second to-be-predicted value according to the second section data vector and the joint distribution function.
9. The electronic device of claim 8, further comprising, after performing the time period model prediction step, the steps of: an error analysis step: and carrying out error analysis on the estimated value of the second predicted value by taking the standard deviation and the relative error as measurement indexes, and judging whether the standard deviation and the relative error are within a preset range.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the wireless network prediction method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810671779.8A CN108989092B (en) | 2018-06-26 | 2018-06-26 | Wireless network prediction method, electronic equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810671779.8A CN108989092B (en) | 2018-06-26 | 2018-06-26 | Wireless network prediction method, electronic equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108989092A CN108989092A (en) | 2018-12-11 |
CN108989092B true CN108989092B (en) | 2023-01-17 |
Family
ID=64538916
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810671779.8A Active CN108989092B (en) | 2018-06-26 | 2018-06-26 | Wireless network prediction method, electronic equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108989092B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111385128B (en) * | 2018-12-29 | 2023-04-07 | 中兴通讯股份有限公司 | Method and device for predicting burst load, storage medium, and electronic device |
CN109996247B (en) * | 2019-03-27 | 2022-03-11 | 中国电子科技集团公司信息科学研究院 | Networked resource allocation method, device, equipment and storage medium |
CN110139288B (en) * | 2019-04-08 | 2022-07-01 | 简链科技(广东)有限公司 | Network communication method, device, system and recording medium |
CN111126624A (en) * | 2019-12-20 | 2020-05-08 | 北京首汽智行科技有限公司 | Method for judging validity of model prediction result |
CN112380273A (en) * | 2020-11-11 | 2021-02-19 | 北京达佳互联信息技术有限公司 | Data estimation method and device |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102542126B (en) * | 2011-10-10 | 2015-04-22 | 上海交通大学 | Soft measurement method based on half supervision learning |
CN102621953B (en) * | 2012-03-20 | 2014-04-09 | 天津大学 | Automatic online quality monitoring and prediction model updating method for rubber hardness |
JP6313730B2 (en) * | 2015-04-10 | 2018-04-18 | タタ コンサルタンシー サービシズ リミテッドTATA Consultancy Services Limited | Anomaly detection system and method |
-
2018
- 2018-06-26 CN CN201810671779.8A patent/CN108989092B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN108989092A (en) | 2018-12-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108989092B (en) | Wireless network prediction method, electronic equipment and storage medium | |
WO2021169577A1 (en) | Wireless service traffic prediction method based on weighted federated learning | |
Nie et al. | Network traffic prediction based on deep belief network in wireless mesh backbone networks | |
CN109587713B (en) | Network index prediction method and device based on ARIMA model and storage medium | |
CN110647900B (en) | Intelligent safety situation prediction method, device and system based on deep neural network | |
CN110991905B (en) | Risk model training method and device | |
CN107547154B (en) | Method and device for establishing video traffic prediction model | |
US20170185922A1 (en) | Hierarchical Capital Allocation Using Clustered Machine Learning | |
Lam et al. | Tail analysis without parametric models: A worst-case perspective | |
CN113037877B (en) | Optimization method for time-space data and resource scheduling under cloud edge architecture | |
CN110621026B (en) | Multi-moment prediction method for base station flow | |
CN113282409B (en) | Edge calculation task processing method and device and computer equipment | |
CN112911626A (en) | Wireless network flow prediction method based on multi-graph convolution | |
Rao et al. | Time Series Forecasting methods suitable for prediction of CPU usage | |
CN117241295A (en) | Wireless communication network performance optimization method, device and storage medium | |
Zhang et al. | Graph-based traffic forecasting via communication-efficient federated learning | |
CN112990721B (en) | Power user value analysis method and system based on payment behaviors | |
CN104217296A (en) | Listed company performance comprehensive evaluation method | |
WO2017131696A1 (en) | Database server to predict sales | |
US11775852B2 (en) | Network optimization | |
CN117094535A (en) | Artificial intelligence-based energy supply management method and system | |
CN116302481B (en) | Resource allocation method and system based on sparse knowledge graph link prediction | |
CN114971092B (en) | Method and device for predicting concrete consumption of main control building of transformer substation | |
CN117117833A (en) | Photovoltaic output power prediction method and device, electronic equipment and storage medium | |
Cai et al. | Demand-driven task scheduling using 2d chromosome genetic algorithm in mobile cloud |
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 |