CN106452934A - Analyzing method for network performance index change trend and device for realizing same - Google Patents
Analyzing method for network performance index change trend and device for realizing same Download PDFInfo
- Publication number
- CN106452934A CN106452934A CN201510486706.8A CN201510486706A CN106452934A CN 106452934 A CN106452934 A CN 106452934A CN 201510486706 A CN201510486706 A CN 201510486706A CN 106452934 A CN106452934 A CN 106452934A
- Authority
- CN
- China
- Prior art keywords
- network performance
- performance index
- index
- variation tendency
- difference
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/50—Testing arrangements
Abstract
The invention discloses an analyzing method for a network performance index change trend. The analyzing method comprises the steps of acquiring continuous T index values which correspond with each network performance index in a preset geographical range and a preset time range, wherein the sampling periods of the T index values are same, and N network performance indexes exist in the preset geographical range; calculating a difference between an m-th index value and an (m-1)th index value of each network performance index, and obtaining (T-1) differences which correspond with each network performance index; performing correlation calculation according to the difference which corresponds with each network performance index, and obtaining a correlation coefficient of each network performance index; classifying the N network performance indexes according to the correlation coefficients; performing normalization processing on each network performance index in one class after classification, and determining the sequence of change amplitudes of the network performance index in each class by means of a normalization result. The invention further discloses a device for realizing the analyzing method.
Description
Technical field
The present invention relates to mobile communication technology field, more particularly, to a kind of network performance index variation tendency point
Analysis method and apparatus.
Background technology
After communication network builds up, need normally to run by continuing to optimize guarantee network, so that network
Quality meets user's request.During the existing network optimization, network optimization personnel need according to substantial amounts of
Network performance index form, draws index change curve, and the variation tendency of manual observation indices.With
When, a lot of due to enjoying the index that network is related to, network optimization personnel simultaneously can not possibly draw each index successively
Change curve carry out case study, can only according to existing network optimization experience, choose that it thinks one
A little key indexs carry out observation analysis, form initial optimization scheme.In addition, prioritization scheme implement after network
The impossible impact to all indexs for the peep optimization scheme of optimization personnel, can only observe it according to previous experiences and recognize
For may affected index variation tendency, draw assessment result.
It can be seen that, existing optimal way needs to expend substantial amounts of manpower and time, have impact on the accuracy of analysis
With comprehensive;And, existing analysis result lacks extensibility, is not suitable for the network performance newly defining
Index, therefore analysis efficiency are relatively low.
Content of the invention
For solving existing technical problem, the embodiment of the present invention provides a kind of network performance index change to become
The analysis method of gesture and device.
Embodiments provide a kind of analysis method of network performance index variation tendency, the method includes:
The corresponding continuous T of each network performance index in collection default geographic range and preset time range
Desired value;Wherein, the sampling period of described T desired value is identical, has N in described default geographic range
Individual network performance index, described T and N is positive integer;
Calculate m-th desired value and the difference of the m-1 desired value of each network performance index, obtain each
The corresponding T-1 difference of network performance index, described m=2 ..., T;
Carry out correlation calculations according to the corresponding difference of each network performance index described, obtain each internetworking
Correlation coefficient between energy index;
According to described correlation coefficient, described N number of network performance index is classified;
Each network performance index in sorted same category is normalized, according to normalization
Result determines the sequence of the amplitude of variation of network performance index in each classification.
In the embodiment of the present invention, described in described foundation, the corresponding difference of each network performance index carries out dependency
Calculate, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
In the embodiment of the present invention, described according to described correlation coefficient, described N number of network performance index is carried out point
Class, including:
On the basis of choosing i-th network performance index according to analysis demand from described N number of network performance index
Index;
According to the r in described variation tendency correlation matrixi1,…rij,…riNThe value of element, determines described i-th
The pass of the variation tendency between any one index in network performance index and other N-1 network performance index
System;
According to the relation of the variation tendency between described i-th network performance index and other any one indexs,
Described N number of network performance index is classified.
In the embodiment of the present invention, between described i-th network performance index of foundation and other any one indexs
The relation of variation tendency, described N number of network performance index is classified, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
In the embodiment of the present invention, described each network performance index in sorted same category is returned
One change is processed, including:
To each network performance index in sorted same category, according to equation below by each internetworking
Index can normalize to [0,1] interval:
Wherein, described Pi is i-th network performance index, described PinormFor i-th network performance index
Normalization result.
In the embodiment of the present invention, the described change determining network performance index in each classification according to normalization result
The sequence of change amplitude, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
The embodiment of the present invention additionally provides a kind of analytical equipment of network performance index variation tendency, this device bag
Include:Collecting unit, computing unit, taxon and statistic unit;Wherein,
Described collecting unit, for gathering each network performance in default geographic range and preset time range
Index corresponding continuous T desired value;Wherein, the sampling period of described T desired value is identical, described pre-
If having N number of network performance index in geographic range, described T and N is positive integer;
Described computing unit, m-th desired value for calculating each network performance index is referred to the m-1
The difference of scale value, obtains the corresponding T-1 difference of each network performance index, described m=2 ..., T;According to institute
State the corresponding difference of each network performance index and carry out correlation calculations, obtain between each network performance index
Correlation coefficient;
Described taxon, for classifying to described N number of network performance index according to described correlation coefficient;
Described statistic unit, for carrying out normalizing to each network performance index in sorted same category
Change is processed, and determines the sequence of the amplitude of variation of network performance index in each classification according to normalization result.
In the embodiment of the present invention, described computing unit includes:Mathematic interpolation subelement and correlation calculations are single
Unit;Wherein,
Described mathematic interpolation subelement, for calculating m-th desired value of each network performance index and the
The difference of m-1 desired value, obtains the corresponding T-1 difference of each network performance index;
Described correlation calculations subelement, for carrying out according to the corresponding difference of each network performance index described
Correlation calculations, obtain the correlation coefficient between each network performance index;
In the embodiment of the present invention, described correlation calculations subelement corresponds to according to each network performance index described
Difference carry out correlation calculations, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
In the embodiment of the present invention, described taxon includes:Choose subelement, determination subelement and classification
Unit;Wherein,
Described selection subelement, for choosing i-th according to analysis demand from described N number of network performance index
Index on the basis of individual network performance index;
Described determination subelement, for according to the r in described variation tendency correlation matrixi1,…rij,…riNElement
Value, determines described i-th network performance index and any one index in other N-1 network performance index
Between variation tendency relation;
Described classification subelement, for according to described i-th network performance index and other any one indexs it
Between variation tendency relation, described N number of network performance index is classified.
In the embodiment of the present invention, described classification subelement according to i-th network performance index and other any one
The relation of the variation tendency between index, described N number of network performance index is classified, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
In the embodiment of the present invention, described statistic unit includes:Normalization subelement and determination subelement;Wherein,
Described normalization subelement, for carrying out to each network performance index in sorted same category
Normalized;
Described determination subelement, for determining the change of network performance index in each classification according to normalization result
The sequence of change amplitude.
In the embodiment of the present invention, described determination subelement determines internetworking in each classification according to normalization result
The sequence of the amplitude of variation of energy index, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
The analysis method of network performance index variation tendency provided in an embodiment of the present invention and device, collection is default
Each network performance index corresponding continuous T desired value in geographic range and preset time range;Wherein,
The sampling period of described T desired value is identical, has N number of network performance index in described default geographic range;
Calculate m-th desired value and the difference of the m-1 desired value of each network performance index, obtain each network
The corresponding T-1 difference of performance indications;Carry out correlation according to the corresponding difference of each network performance index described
Property calculate, obtain the correlation coefficient between each network performance index;According to described correlation coefficient to described N number of
Network performance index is classified;Normalizing is carried out to each network performance index in sorted same category
Change is processed, and determines the sequence of the amplitude of variation of network performance index in each classification according to normalization result.This
Inventive embodiments can grope and find unknown by the analysis to network performance index variation tendency dependency
Network problem is it is contemplated that solve potential network associate performance, fault etc.;And, network performance index point
Analysis process can be automatically performed, and need not move through picture trendgram and eye-observation, substantially increase analysis efficiency;This
Outward, for the sequence of the amplitude of variation of the network performance index having obtained, network optimization personnel can be to all nets
The ranking results of network performance indications are analyzed, and breach the conventional Optimization Experience of dependence and choose finite element network performance
The limitation that index is analyzed, accomplishes to analyze network index situation comprehensively and fully understands that prioritization scheme is implemented
Impact to existing network quality afterwards.
Brief description
In accompanying drawing (it is not necessarily drawn to scale), similar reference can be in different views
Described in similar part.The similar reference numerals with different letter suffix can represent the difference of similar component
Example.Accompanying drawing generally shows each embodiment discussed herein by way of example and not limitation.
Fig. 1 is the analysis method flowchart of network performance index variation tendency described in the embodiment of the present invention;
Fig. 2 is the analytical equipment structural representation of network performance index variation tendency described in the embodiment of the present invention
One;
Fig. 3 is the analytical equipment structural representation of network performance index variation tendency described in the embodiment of the present invention
Two;
Fig. 4 is the analytical equipment structural representation of network performance index variation tendency described in the embodiment of the present invention
Three;
Fig. 5 is the analytical equipment structural representation of network performance index variation tendency described in the embodiment of the present invention
Four;
Fig. 6 is scene one methods described schematic flow sheet of the present invention;
Fig. 7 is scene two described device structural representation of the present invention.
Specific embodiment
Each network performance in embodiments of the invention, in collection default geographic range and preset time range
Index corresponding continuous T desired value;Wherein, the sampling period of described T desired value is identical, described pre-
If having N number of network performance index in geographic range, described T and N is positive integer;Calculate each internetworking
M-th desired value of energy index and the difference of the m-1 desired value, obtain each network performance index corresponding
T-1 difference;Described m=2 ..., T;Carry out correlation according to the corresponding difference of each network performance index described
Property calculate, obtain the correlation coefficient between each network performance index;According to described correlation coefficient to described N number of
Network performance index is classified;Normalizing is carried out to each network performance index in sorted same category
Change is processed, and determines the sequence of the amplitude of variation of network performance index in each classification according to normalization result.
Below in conjunction with the accompanying drawings and specific embodiment is described in further detail to the present invention.
Fig. 1 is the analysis method flowchart of network performance index variation tendency described in the embodiment of the present invention,
As shown in figure 1, the method includes:
Step 101:The default geographic range of collection is corresponding with each network performance index in preset time range
Continuous T desired value;Wherein, the sampling period of described T desired value is identical, described default geography model
Enclose interior total N number of network performance index, described T and N is positive integer;
Step 102:Calculate m-th desired value of each network performance index and the difference of the m-1 desired value,
Obtain the corresponding T-1 difference of each network performance index;Described m=2 ..., T;
Step 103:Carry out correlation calculations according to the corresponding difference of each network performance index described, obtain
Correlation coefficient between each network performance index;
Step 104:According to described correlation coefficient, described N number of network performance index is classified;
Step 105:Each network performance index in sorted same category is normalized,
Determine the sequence of the amplitude of variation of network performance index in each classification according to normalization result.
In the embodiment of the present invention, if network optimization personnel wish daily to analyze the network performance index of certain cell,
And the network performance index variation tendency of 30 days need to be analyzed, then gather all N number of network performance of this cell
Index is in the index sampled value of continuous 30 days, a daily desired value.Here, can be all by collection gained
Desired value is configured to a matrix, for example:30 row N row.
In the embodiment of the present invention, described in described foundation, the corresponding difference of each network performance index carries out dependency
Calculate, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
In the embodiment of the present invention, described according to described correlation coefficient, described N number of network performance index is carried out point
Class, including:
On the basis of choosing i-th network performance index according to analysis demand from described N number of network performance index
Index;
According to the r in described variation tendency correlation matrixi1,…rij,…riNThe value of element, determines described i-th
The pass of the variation tendency between any one index in network performance index and other N-1 network performance index
System;
According to the relation of the variation tendency between described i-th network performance index and other any one indexs,
Described N number of network performance index is classified.
In the embodiment of the present invention, between described i-th network performance index of foundation and other any one indexs
The relation of variation tendency, described N number of network performance index is classified, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
In the embodiment of the present invention, described each network performance index in sorted same category is returned
One change is processed, including:
To each network performance index in sorted same category, according to equation below by each internetworking
Index can normalize to [0,1] interval:
Wherein, described Pi is i-th network performance index, described PinormFor i-th network performance index
Normalization result.
In the embodiment of the present invention, the described change determining network performance index in each classification according to normalization result
The sequence of change amplitude, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
The embodiment of the present invention additionally provides a kind of analytical equipment of network performance index variation tendency, as Fig. 2 institute
Show, this device includes:Collecting unit 201, computing unit 202, taxon 203 and statistic unit 204;
Wherein,
Described collecting unit 201, for gathering each network in default geographic range and preset time range
Performance indications corresponding continuous T desired value;Wherein, the sampling period of described T desired value is identical, institute
State and in default geographic range, have N number of network performance index, described T and N is positive integer;
Described computing unit 202, for calculating m-th desired value and the m-1 of each network performance index
The difference of individual desired value, obtains the corresponding T-1 difference of each network performance index, described m=2 ..., T;According to
Carry out correlation calculations according to the corresponding difference of each network performance index described, obtain each network performance index
Between correlation coefficient;
Described taxon 203, for carrying out to described N number of network performance index according to described correlation coefficient
Classification;
Described statistic unit 204, for carrying out to each network performance index in sorted same category
Normalized, determines the sequence of the amplitude of variation of network performance index in each classification according to normalization result.
In one embodiment, as shown in figure 3, described computing unit 202 includes:Mathematic interpolation subelement 2021
With correlation calculations subelement 2022;Wherein,
Described mathematic interpolation subelement 2021, for calculate m-th desired value of each network performance index with
The difference of the m-1 desired value, obtains the corresponding T-1 difference of each network performance index;
Described correlation calculations subelement 2022, for according to the corresponding difference of each network performance index described
Carry out correlation calculations, obtain the correlation coefficient between each network performance index;
In the embodiment of the present invention, described correlation calculations subelement 2022 is according to each network performance index described
Corresponding difference carries out correlation calculations, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
In one embodiment, as shown in figure 4, described taxon 203 includes:Selection subelement 2031,
Determination subelement 2032 and classification subelement 2033;Wherein,
Described selection subelement 2031, for choosing from described N number of network performance index according to analysis demand
Index on the basis of i-th network performance index;
Described determination subelement 2032, for according to the r in described variation tendency correlation matrixi1,…rij,…riN
The value of element, determines arbitrary in N-1 network performance index of described i-th network performance index and other
The relation of the variation tendency between individual index;
Described classification subelement 2033, for according to described i-th network performance index and other, any one refers to
The relation of the variation tendency between mark, described N number of network performance index is classified.
In the embodiment of the present invention, described classification subelement 2033 according to i-th network performance index and other
The relation of the variation tendency between one index, described N number of network performance index is classified, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
In one embodiment, as shown in figure 5, described statistic unit 204 includes:Normalization subelement 2041
With determination subelement 2042;Wherein,
Described normalization subelement 2041, for each network performance index in sorted same category
It is normalized;
Described determination subelement 2042, for determining network performance index in each classification according to normalization result
Amplitude of variation sequence.
In the embodiment of the present invention, described normalization subelement 2041 is to each net in sorted same category
Network performance indications are normalized, including:
To each network performance index in sorted same category, according to equation below by each internetworking
Index can normalize to [0,1] interval:
Wherein, described Pi is i-th network performance index, described PinormFor i-th network performance index
Normalization result.
In the embodiment of the present invention, described determination subelement 2042 determines net in each classification according to normalization result
The sequence of the amplitude of variation of network performance indications, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
It can be seen that, the embodiment of the present invention can be touched by the analysis to network performance index variation tendency dependency
Rope and discovery unknown network problem are it is contemplated that solve potential network associate performance, fault etc.;And, network
The analysis process of performance indications can be automatically performed, and need not move through picture trendgram and eye-observation, substantially increase
Analysis efficiency;Additionally, the sequence of the amplitude of variation for the network performance index having obtained, network optimization people
Member can be analyzed to the ranking results of all-network performance indications, breach the conventional Optimization Experience of dependence and choose
The limitation that finite element network performance indications are analyzed, accomplishes to analyze network index situation and overall understanding comprehensively
Impact to existing network quality after prioritization scheme enforcement.
With reference to concrete application scene, present invention is described.
Scene one
It is assumed that a total of N number of network performance index (P1, P2 ... ..., PN) in this scene, as shown in fig. 6,
The method includes:
Step 601:Collection network performance index, geographic range and time range carry out desired value on demand
Process;
Gather continuous in certain geographic range (cell/region/the whole network etc.) and time range (hour/sky)
The T corresponding desired value of time point (hour/sky etc.), and all desired values are configured to original performance index
Matrix, as follows:
Wherein, described PijThe desired value of j-th time point of (i=1 ... N) individual performance indications for i-th.
Using j-th desired value and the difference of -1 desired value of jth of each network performance index, can be characterized this
Performance indications are in the change of j-th time point.Accordingly, to all time points of all-network performance indications
Desired value all carries out difference operation, can get original performance index variation tendency matrix:
Wherein, described Δ Pi includes all (T-1) difference of i-th network performance index;
Step 602:Calculate and analyze the dependency of baseline network performance indications and the variation tendency of other indexs;
(each network performance refers to each column vector in calculating original performance index variation tendency matrix D elta
Mark) between correlation coefficient, variation tendency correlation matrix can be obtained:
Wherein, described rijFor the correlation coefficient of Δ Pi in Delta and Δ Pj, value is between [- 1,1]:
Step 603:It is dependent on the dependency of variation tendency between reference index, network performance index is carried out point
Class;
Network optimization personnel are in problem analysis or assessment effect of optimization it is assumed that have selected i-th network performance
As reference index, then i-th network performance index is related to the variation tendency of other remaining indexs for index
Property is i.e. by RdeltaIn i-th arrange characterizing, i.e. Ri_based=[ri1… rij… riN], wherein rijFor i-th
The correlation coefficient of the variation tendency of individual network performance index and j-th network performance index, closer to 1 sign
I-th network performance index is more identical with the variation tendency of j-th network performance index, closer to -1 sign
I-th network performance index is more contrary with the variation tendency of j-th network performance index, characterizes the closer to 0
The variation tendency of i network performance index and j-th network performance index does not have significant relation.
Assume that i-th network performance index is to decline in observation cycle, network optimization personnel were analyzing
It should be understood which index is also decline in journey, and which index can rise.Can be according to Ri_basedTo institute
There is N-1 network performance index to be classified, be such as divided into:With i-th network performance index variation tendency phase
With, variation tendency is contrary or no significant relation etc. three class, sorting technique can using clustering algorithm (such as
K-means algorithm) it is also possible to delimit interval range as needed to classify.
Step 604:The sequence of the amplitude of variation of network performance index in each classification of analysis.
Identical and contrary two classes for variation tendency, network optimization personnel are during analysis it should also be understood that referring to
The amplitude of mark change, that is,:Assume that index i is to decline in observation cycle, then network optimization personnel need
Understand which index fall is more, which index ascensional range is more.Adopt the variance can be with measurement index
Amplitude of variation.Because different network performance indexes has different dimensions, it is each index of lateral comparison
Change is it is necessary first to be normalized operation to sorted each index.With the following method can respectively as adopted
It is interval that individual index all normalizes to [0,1]:
Calculate all normalization indexs PinormVariance, and sorted from big to small according to variance, you can obtain each
The sequence of index amplitude of variation in individual classification.Network optimization personnel can obtain accordingly, with i-th net
In the index that network performance indications decline together, which index is reduced by up to;With i-th network performance
The decline of index and in the index that rises, which index rises at most.
Scene two
Based on the method described in scene one, this scene proposes a kind of device corresponding with methods described, such as schemes
Shown in 7, this device includes:
Data preprocessing module 701:For existing network all-network performance indications being pressed the reality of network optimization personnel
Border needs, and is processed into necessarily geographical granularity (geographic range) and the original performance in cycle (time range) and refers to
Mark matrix.The input of this module for network optimization personnel need analysis geographical granularity (for example cell/region/
The whole network etc.), cycle (when for example little/sky/week etc.) and all-network performance indications sampled value.It is output as institute
Need geographical granularity and the network performance index original matrix in the cycle.For example, network optimization personnel wish daily
Analyze the network performance index of certain cell, then need to input all N number of network performance indexes continuous 30 of certain cell
It index sampled value, such as:Every 15 minutes sampled points in one day, then, data preprocessing module
Substantial amounts of for each network performance index index sampled value is daily processed, such as by 701:Will be multiple in the middle of one day
Sampled point (every 15 minutes corresponding one sampled points) averaged obtains the desired value of a day, and output should
The original performance index matrix that the cell all-network performance indications desired value of continuous 30 days is constituted, matrix size
Arrange for 30 row N.
Trend correlation computing module 702:For calculating variation tendency between all-network performance indications
Dependency.The input of this module is the all-network performance index value structure after data preprocessing module 701 process
The original performance index matrix becoming, by calculating each column vector of variation tendency matrix of original performance index matrix
Between correlation coefficient, exportable variation tendency correlation matrix.
Sort module 703:For classifying to network performance index variation tendency.The input of this module is
The analysis reference index of network optimization personnel selection, the variation tendency of trend correlation computing module 702 output
Correlation matrix, the classification number needing, are output as required all kinds of index sets.Sorting technique can be using poly-
Class algorithm or artificially defined.For example:Network optimization personnel wish to be divided into and reference index network performance index
Variation tendency is identical, trend is contrary and no significant relation three class, and sort module is exportable three class index sets.
Evaluation module 704:Input as mould of classifying:All kinds of indexs of 703 outputs, by relatively each index normalizing
Variance after change, obtains the amplitude of variation sequence of all kinds of indexs.For example, network optimization personnel wish analysis with
The maximum index of amplitude of variation in reference index variation tendency identical index set, you can take corresponding index set to return
The maximum index of variance after one change.
The embodiment of the present invention can be groped and sent out by the analysis to network performance index variation tendency dependency
Existing unknown network problem is it is contemplated that solve potential network associate performance, fault etc.;And, network performance refers to
Target analysis process can be automatically performed, and needs not move through picture trendgram and eye-observation, substantially increases analysis effect
Rate;Additionally, the sequence of the amplitude of variation for the network performance index having obtained, network optimization personnel can be right
The ranking results of all-network performance indications are analyzed, and breach the conventional Optimization Experience of dependence and choose finite net
The limitation that network performance indications are analyzed, accomplishes to analyze network index situation comprehensively and fully understands optimization side
Impact to existing network quality after case enforcement.
Those skilled in the art are it should be appreciated that embodiments of the invention can be provided as method, system or meter
Calculation machine program product.Therefore, the present invention can using hardware embodiment, software implementation or combine software and
The form of the embodiment of hardware aspect.And, the present invention can adopt and wherein include calculating one or more
Computer-usable storage medium (including but not limited to disk memory and the optical storage of machine usable program code
Device etc.) the upper computer program implemented form.
The present invention is with reference to method according to embodiments of the present invention, equipment (system) and computer program
Flow chart and/or block diagram describing.It should be understood that can be by computer program instructions flowchart and/or side
Each flow process in block diagram and/or the knot of the flow process in square frame and flow chart and/or block diagram and/or square frame
Close.Can provide these computer program instructions to general purpose computer, special-purpose computer, Embedded Processor or
The processor of other programmable data processing device with produce a machine so that by computer or other can
The instruction of the computing device of programming data processing equipment produces for realizing in one flow process or multiple of flow chart
The device of the function of specifying in flow process and/or one square frame of block diagram or multiple square frame.
These computer program instructions may be alternatively stored in and can guide computer or other programmable data processing device
So that being stored in this computer-readable memory in the computer-readable memory working in a specific way
Instruction produces the manufacture including command device, and this command device is realized in one flow process of flow chart or multiple stream
The function of specifying in journey and/or one square frame of block diagram or multiple square frame.
These computer program instructions also can be loaded in computer or other programmable data processing device, makes
Obtain and series of operation steps is executed on computer or other programmable devices to produce computer implemented place
Reason, thus the instruction of execution is provided for realizing in flow chart one on computer or other programmable devices
The step of the function of specifying in flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.
The above, only presently preferred embodiments of the present invention, it is not intended to limit the protection model of the present invention
Enclose.
Claims (13)
1. a kind of analysis method of network performance index variation tendency is it is characterised in that the method includes:
The corresponding continuous T of each network performance index in collection default geographic range and preset time range
Desired value;Wherein, the sampling period of described T desired value is identical, has N in described default geographic range
Individual network performance index, described T and N is positive integer;
Calculate m-th desired value and the difference of the m-1 desired value of each network performance index, obtain each
The corresponding T-1 difference of network performance index, described m=2 ..., T;
Carry out correlation calculations according to the corresponding difference of each network performance index described, obtain each internetworking
Correlation coefficient between energy index;
According to described correlation coefficient, described N number of network performance index is classified;
Each network performance index in sorted same category is normalized, according to normalization
Result determines the sequence of the amplitude of variation of network performance index in each classification.
2. method according to claim 1 is it is characterised in that each network performance described in described foundation
The corresponding difference of index carries out correlation calculations, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
3. method according to claim 2 it is characterised in that described according to described correlation coefficient to institute
State N number of network performance index to be classified, including:
On the basis of choosing i-th network performance index according to analysis demand from described N number of network performance index
Index;
According to the r in described variation tendency correlation matrixi1,…rij,…riNThe value of element, determines described i-th
The pass of the variation tendency between any one index in network performance index and other N-1 network performance index
System;
According to the relation of the variation tendency between described i-th network performance index and other any one indexs,
Described N number of network performance index is classified.
4. method according to claim 3 is it is characterised in that i-th network performance of described foundation refers to
The relation of the variation tendency between mark and other any one indexs, described N number of network performance index is carried out point
Class, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
5. method according to claim 1 it is characterised in that described in sorted same category
Each network performance index be normalized, including:
To each network performance index in sorted same category, according to equation below by each internetworking
Index can normalize to [0,1] interval:
Wherein, described Pi is i-th network performance index, described PinormFor i-th network performance index
Normalization result.
6. method according to claim 1 is it is characterised in that described determine often according to normalization result
The sequence of the amplitude of variation of network performance index in individual classification, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
7. a kind of analytical equipment of network performance index variation tendency is it is characterised in that this device includes:Adopt
Collection unit, computing unit, taxon and statistic unit;Wherein,
Described collecting unit, for gathering each network performance in default geographic range and preset time range
Index corresponding continuous T desired value;Wherein, the sampling period of described T desired value is identical, described pre-
If having N number of network performance index in geographic range, described T and N is positive integer;
Described computing unit, m-th desired value for calculating each network performance index is referred to the m-1
The difference of scale value, obtains the corresponding T-1 difference of each network performance index, described m=2 ..., T;According to institute
State the corresponding difference of each network performance index and carry out correlation calculations, obtain between each network performance index
Correlation coefficient;
Described taxon, for classifying to described N number of network performance index according to described correlation coefficient;
Described statistic unit, for carrying out normalizing to each network performance index in sorted same category
Change is processed, and determines the sequence of the amplitude of variation of network performance index in each classification according to normalization result.
8. device according to claim 7 is it is characterised in that described computing unit includes:Difference meter
Operator unit and correlation calculations subelement;Wherein,
Described mathematic interpolation subelement, for calculating m-th desired value of each network performance index and the
The difference of m-1 desired value, obtains the corresponding T-1 difference of each network performance index;
Described correlation calculations subelement, for carrying out according to the corresponding difference of each network performance index described
Correlation calculations, obtain the correlation coefficient between each network performance index.
9. device according to claim 8 is it is characterised in that described correlation calculations subelement foundation
The corresponding difference of described each network performance index carries out correlation calculations, including:
One T-1 row N row is constructed for element with the corresponding T-1 difference of each network performance index described
Matrix, each column vector correspondence of described matrix includes the corresponding T-1 difference of network performance index,
Described T-1 difference is arranged by increasing or decreasing mode, and the corresponding T-1 of described N number of network performance index
The arrangement mode all same of individual difference;
Calculate the correlation coefficient between each column vector of described matrix, obtain variation tendency correlation matrix;
Wherein, each of described variation tendency correlation matrix element rijRepresent i-th network performance index
With the correlation coefficient of the variation tendency of j-th network performance index, rijValue between [- 1,1], described i,
J=1 ..., N.
10. device according to claim 8 is it is characterised in that described taxon includes:Choose
Subelement, determination subelement and classification subelement;Wherein,
Described selection subelement, for choosing i-th according to analysis demand from described N number of network performance index
Index on the basis of individual network performance index;
Described determination subelement, for according to the r in described variation tendency correlation matrixi1,…rij,…riNElement
Value, determines described i-th network performance index and any one index in other N-1 network performance index
Between variation tendency relation;
Described classification subelement, for according to described i-th network performance index and other any one indexs it
Between variation tendency relation, described N number of network performance index is classified.
11. devices according to claim 10 are it is characterised in that described classification subelement is according to i-th
The relation of the variation tendency between individual network performance index and other any one indexs, by described N number of internetworking
Can index be classified, including:
If rijValue close to 1, then characterize i-th network performance index and j-th network performance index
Variation tendency is identical;If rijValue close -1, then characterize i-th network performance index and j-th network
The variation tendency of performance indications is contrary;If rijValue close to 0, then characterize i-th network performance index
There is no significant relation with the variation tendency of j-th network performance index;
A class will be divided into described i-th network performance index variation tendency identical network performance index, will
Be divided into a class with described i-th contrary network performance index of network performance index variation tendency, will with described
I-th network performance index variation tendency does not have the network performance index of significant relation to be divided into a class.
12. devices according to claim 7 are it is characterised in that described statistic unit includes:Normalizing
Beggar's unit and determination subelement;Wherein,
Described normalization subelement, for carrying out to each network performance index in sorted same category
Normalized;
Described determination subelement, for determining the change of network performance index in each classification according to normalization result
The sequence of change amplitude.
13. devices according to claim 12 are it is characterised in that described determination subelement is according to normalizing
Change the sequence that result determines the amplitude of variation of network performance index in each classification, including:
Calculate the variance of the normalization result of each network performance index in described same category, and by gained
Variance be ranked up from big to small or from small to large, obtain each network performance in described same category
The sequence of the amplitude of variation of index.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510486706.8A CN106452934B (en) | 2015-08-10 | 2015-08-10 | Method and device for analyzing network performance index change trend |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510486706.8A CN106452934B (en) | 2015-08-10 | 2015-08-10 | Method and device for analyzing network performance index change trend |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106452934A true CN106452934A (en) | 2017-02-22 |
CN106452934B CN106452934B (en) | 2020-02-11 |
Family
ID=58093079
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510486706.8A Active CN106452934B (en) | 2015-08-10 | 2015-08-10 | Method and device for analyzing network performance index change trend |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106452934B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108923952A (en) * | 2018-05-31 | 2018-11-30 | 北京百度网讯科技有限公司 | Method for diagnosing faults, equipment and storage medium based on service monitoring index |
CN109587001A (en) * | 2018-11-15 | 2019-04-05 | 新华三信息安全技术有限公司 | A kind of performance indicator method for detecting abnormality and device |
CN109840185A (en) * | 2019-01-11 | 2019-06-04 | 新华三大数据技术有限公司 | A kind of index correlating method and device |
CN111585789A (en) * | 2020-04-13 | 2020-08-25 | 中国联合网络通信集团有限公司 | Data prediction method and device |
CN115396929A (en) * | 2022-08-15 | 2022-11-25 | 中国联合网络通信集团有限公司 | Performance data prediction method, device and storage medium |
CN115618076A (en) * | 2022-11-07 | 2023-01-17 | 山东省地质矿产勘查开发局第一地质大队(山东省第一地质矿产勘查院) | Hydrogeological data management method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101159586A (en) * | 2007-04-09 | 2008-04-09 | 中国移动通信集团设计院有限公司 | Communication network performance optimization method and device |
CN101572623A (en) * | 2009-04-30 | 2009-11-04 | 上海大学 | Method for comprehensively evaluating network performance based on subjective and objective combination evaluation |
CN101609416A (en) * | 2009-07-13 | 2009-12-23 | 清华大学 | Improve the method for performance tuning speed of distributed system |
CN101751448A (en) * | 2009-07-22 | 2010-06-23 | 中国科学院自动化研究所 | Commendation method of personalized resource information based on scene information |
CN102722577A (en) * | 2012-06-05 | 2012-10-10 | 中兴通讯股份有限公司 | Method and device for determining dynamic weights of indexes |
CN103731851A (en) * | 2014-01-15 | 2014-04-16 | 中国移动通信集团陕西有限公司 | Network performance evaluation method and device |
US8832262B2 (en) * | 2011-12-15 | 2014-09-09 | Cisco Technology, Inc. | Normalizing network performance indexes |
-
2015
- 2015-08-10 CN CN201510486706.8A patent/CN106452934B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101159586A (en) * | 2007-04-09 | 2008-04-09 | 中国移动通信集团设计院有限公司 | Communication network performance optimization method and device |
CN101572623A (en) * | 2009-04-30 | 2009-11-04 | 上海大学 | Method for comprehensively evaluating network performance based on subjective and objective combination evaluation |
CN101609416A (en) * | 2009-07-13 | 2009-12-23 | 清华大学 | Improve the method for performance tuning speed of distributed system |
CN101751448A (en) * | 2009-07-22 | 2010-06-23 | 中国科学院自动化研究所 | Commendation method of personalized resource information based on scene information |
US8832262B2 (en) * | 2011-12-15 | 2014-09-09 | Cisco Technology, Inc. | Normalizing network performance indexes |
CN102722577A (en) * | 2012-06-05 | 2012-10-10 | 中兴通讯股份有限公司 | Method and device for determining dynamic weights of indexes |
CN103731851A (en) * | 2014-01-15 | 2014-04-16 | 中国移动通信集团陕西有限公司 | Network performance evaluation method and device |
Non-Patent Citations (1)
Title |
---|
罗赟骞等: "基于云模型和熵权的网络性能综合评估模型", 《重庆邮电大学学报(自然科学版)》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108923952A (en) * | 2018-05-31 | 2018-11-30 | 北京百度网讯科技有限公司 | Method for diagnosing faults, equipment and storage medium based on service monitoring index |
CN109587001A (en) * | 2018-11-15 | 2019-04-05 | 新华三信息安全技术有限公司 | A kind of performance indicator method for detecting abnormality and device |
CN109587001B (en) * | 2018-11-15 | 2020-11-27 | 新华三信息安全技术有限公司 | Performance index abnormality detection method and device |
CN109840185A (en) * | 2019-01-11 | 2019-06-04 | 新华三大数据技术有限公司 | A kind of index correlating method and device |
CN111585789A (en) * | 2020-04-13 | 2020-08-25 | 中国联合网络通信集团有限公司 | Data prediction method and device |
CN111585789B (en) * | 2020-04-13 | 2023-02-17 | 中国联合网络通信集团有限公司 | Data prediction method and device |
CN115396929A (en) * | 2022-08-15 | 2022-11-25 | 中国联合网络通信集团有限公司 | Performance data prediction method, device and storage medium |
CN115618076A (en) * | 2022-11-07 | 2023-01-17 | 山东省地质矿产勘查开发局第一地质大队(山东省第一地质矿产勘查院) | Hydrogeological data management method |
Also Published As
Publication number | Publication date |
---|---|
CN106452934B (en) | 2020-02-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106452934A (en) | Analyzing method for network performance index change trend and device for realizing same | |
Wei et al. | Variable importance analysis: A comprehensive review | |
JP7090936B2 (en) | ESG-based corporate evaluation execution device and its operation method | |
CN110688288B (en) | Automatic test method, device, equipment and storage medium based on artificial intelligence | |
CN101556553B (en) | Defect prediction method and system based on requirement change | |
CN103617435B (en) | Image sorting method and system for active learning | |
Riska et al. | An EM-based technique for approximating long-tailed data sets with PH distributions | |
CN107862022A (en) | Cultural resource commending system | |
US20170330078A1 (en) | Method and system for automated model building | |
Zhang et al. | An automatic and effective parameter optimization method for model tuning | |
CN107766316B (en) | Evaluation data analysis method, device and system | |
JP6943242B2 (en) | Analytical instruments, analytical methods, and programs | |
CN115545331A (en) | Control strategy prediction method and device, equipment and storage medium | |
CN116775879A (en) | Fine tuning training method of large language model, contract risk review method and system | |
CN109636212B (en) | Method for predicting actual running time of job | |
CN103425748A (en) | Method and device for mining document resource recommended words | |
CN106487571B (en) | Method and device for evaluating network performance index change trend | |
CN104599062A (en) | Classification based value evaluation method and system for agricultural scientific and technological achievements | |
CN106487570B (en) | A kind of method and apparatus for assessing network performance index variation tendency | |
CN104462215B (en) | A kind of scientific and technical literature based on time series is cited number Forecasting Methodology | |
US11669519B2 (en) | System for generating predicate-weighted histograms | |
CN111027599B (en) | Clustering visualization method and device based on random sampling | |
CN114266496A (en) | Policy landing effect evaluation method and system based on policy completeness analysis | |
CN113408374A (en) | Yield estimation method, device and equipment based on artificial intelligence and storage medium | |
CN113344257A (en) | Mixed prediction method for layer analysis response time in homeland space information service cloud platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 |