CN106487570B - A kind of method and apparatus for assessing network performance index variation tendency - Google Patents
A kind of method and apparatus for assessing network performance index variation tendency Download PDFInfo
- Publication number
- CN106487570B CN106487570B CN201510557027.5A CN201510557027A CN106487570B CN 106487570 B CN106487570 B CN 106487570B CN 201510557027 A CN201510557027 A CN 201510557027A CN 106487570 B CN106487570 B CN 106487570B
- Authority
- CN
- China
- Prior art keywords
- index
- matrix
- performance index
- network performance
- desired indicator
- 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
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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of method and apparatus for assessing network performance index variation tendency, n network performance index in geographical granularity is set in the index value at continuous m time point the described method includes: obtaining, obtain the original performance index matrix of n × m dimension of n network performance index, then the mahalanobis distance of each other indexs and desired indicator in n network performance index is calculated by calculating the covariance matrix of original performance index matrix, and then obtain the index with the immediate index of the variation tendency of the desired indicator and with the variation tendency disparity of the desired indicator, wherein, the n is the positive integer greater than 1, the m is the positive integer not less than 1, to realize judging automatically for network performance index variation tendency, it improves the analysis efficiency of the network optimization and the comprehensive of network optimization analysis can be improved Property and scalability.
Description
Technical field
The present invention relates to network communication technology field more particularly to a kind of methods for assessing network performance index variation tendency
And equipment.
Background technique
After communication network is built up, need just to can guarantee that network operates normally by continuing to optimize, so that network quality meets
User demand.The network optimization is by carrying out data acquisition, data analysis, Parameter analysis, firmly to the existing network that has run
Part inspection etc., find out influence network quality the reason of, and by parameter adjustment, network structure regulation, device configuration adjust and its
His technological means, it is ensured that the operation of network high quality makes conventional network resources obtain optimum efficiency.
In general, the network optimization can comprise the following steps that
Step 1: data acquisition passes through test, network management or monitoring system acquisition network performance index;
Step 2: data analysis, i.e., the network performance index of acquisition is analyzed, for example, analysis indexes absolute value and its
Variation etc.;
Step 3: positioning problems carry out existing net in conjunction with the Optimization Experience of network optimization personnel that is, by index analysis result
Problem judgment;
Step 4: scheme is implemented, i.e., according to the possible problem of the existing net judged, formulates prioritization scheme, and it is real to carry out existing net
It applies;
Step 5: recruitment evaluation implements front and back, the situation of change of network performance index, assessment optimization by comparison scheme
Whether scheme is effective, if unsolved problem, needs suboptimization again.
As shown in the above, during the network optimization, network performance index is that network optimization personnel carry out network point
The basis of analysis, network optimization personnel can carry out network problem positioning by the mutation analysis to network performance index, and scheme is real
Shi Hou can also optimize the recruitment evaluation of scheme by the analysis again to network performance index.
Specifically, network optimization personnel are when carrying out network performance index analysis, usually by internetworking predetermined
Energy index output is drawn as trend chart at excel report, and to its index set of concern, then, is become according to index variation
Gesture figure carries out case study;And after obtaining preliminary solution by analysis and implementing, network optimization personnel can be right again
Its index set of concern draws trend chart, and further assesses effect of optimization.
That is, network optimization personnel need to be referred to according to a large amount of network performance during the existing network optimization
Mark report, by way of drawing index change curve, the variation tendency of artificial observation indices.But due to showing net at present
There are many index of definition, network optimization personnel and the change curve progress case study that can not successively draw each index,
It can only choose some key indexes that it is thought according to existing network optimization experience and carry out observation analysis, form preliminary optimization
Scheme.Similarly, network optimization personnel are also impossible to influence of the peep optimization scheme to all indexs after prioritization scheme implementation,
Its variation tendency for thinking to may be subjected to the index of influence can only be observed according to previous experiences, assessment result is obtained, to lead
Cause can have the following problems:
Problem one expends a large amount of manpower and time, reduces the efficiency of the network optimization: since the network optimization is one from net
Network builds up the continual process started after operation, thus, network optimization personnel need constantly to carry out selecting index, refer to
The operation such as observation, index evaluation is marked, if relying solely on eye-observation index, it will it is time-consuming and laborious, influence efficiency.
Problem two, the comprehensive and accuracy for lacking analysis: due to network performance index enormous amount, network optimization people
Member and can not the variation tendency to all indexs all analyze, be also impossible to fully understand prioritization scheme implement after to show
The influence of net quality, it is possible to the index set Long-term change trend as expected that network optimization personnel are concerned about, but disliking occur in other indexs
Change but and undiscovered, these situations will the accuracy of impact analysis and comprehensive.
Problem three, the scalability for lacking analysis: during the network optimization, network optimization personnel are usually according to previous
What the index set that a large amount of Optimization Experiences of accumulation choose its care was observed, and with the new demand of existing network optimization, it will
Can continuous more thinner network performance indexes of definition, at this moment network optimization personnel are for newly defining the pass of index Yu original index
System does not have experience that can follow, and can largely effect on the scalability of analysis, reduce the efficiency of analysis.
That is, due to network performance index variation tendency be usually by manually rule of thumb being judged, thus
Causing to exist reduces the analysis efficiency of the network optimization, network optimization analysis is made to lack asking for comprehensive and scalability etc.
Topic.
Summary of the invention
The embodiment of the invention provides a kind of method and apparatus for assessing network performance index variation tendency, to realize net
Network performance indicator variation tendency judges automatically, with solve existing network optimization analysis mode present in analysis efficiency it is low, with
And the problem of comprehensive and scalability of shortage etc..
The embodiment of the invention provides a kind of methods for assessing network performance index variation tendency, comprising:
N network performance index in the geographical granularity of setting is obtained in the index value at continuous m time point, obtains n net
The original performance index matrix of n × m dimension of network performance indicator;
Calculate the covariance matrix of the original performance index matrix;
For any desired indicator in the n network performance index, according to the original performance index matrix and
The covariance matrix calculates the mahalanobis distance of each other indexs and the desired indicator in the n network performance index;
According to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the n network performance is determined
In index with the immediate index of variation tendency of the desired indicator and with the variation tendency gap of the desired indicator
Maximum index;
Wherein, the n is the positive integer greater than 1, and the m is the positive integer not less than 1.
Optionally, it is calculated by the following formula the covariance matrix ∑ of the original performance index matrix X:
Wherein, μ i is the desired value of i-th of index Xi, (i, j) a element ∑ ij in covariance matrix ∑ be Xi with
The covariance of j-th of index Xj, i, j are respectively any positive integer for being not more than n.
Optionally, any index Xj and the predetermined finger being calculated by the following formula in the n network performance index
Mark the mahalanobis distance Dij of Xi:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and i, j are respectively not more than any just whole of n
Number.
Optionally, according to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the n net is determined
Becoming in network performance indicator with the immediate index of variation tendency of the desired indicator and with the variation of the desired indicator
Potential difference is away from maximum index, comprising:
According to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the geneva of 1 × (n-1) dimension is obtained
Distance matrix;
According to the size of each matrix element in the mahalanobis distance matrix, according to sequence from big to small to the geneva
Each matrix element in distance matrix is ranked up;
The most preceding index of sorting as in the n network performance index with the variation tendency of the desired indicator most
Close index, the last index that will sort is as the index of the variation tendency disparity with the desired indicator.
Optionally, before the covariance matrix for calculating the original performance index matrix, the method also includes:
Operation is normalized to each index value of the n network performance index got.
Optionally, in the mahalanobis distance according to each other indexs and the desired indicator being calculated, the n are determined
The variation with the immediate index of variation tendency of the desired indicator and with the desired indicator in network performance index
After the index of trend disparity, the method also includes:
According to the determining change with the immediate index of desired indicator variation tendency and with the desired indicator
The index of change trend disparity, whether the influence for determining that the adjustment of network parameter changes index is reasonable, to judge that network is joined
Whether number adjustment is reasonable.
Correspondingly, the embodiment of the invention also provides a kind of equipment for assessing network performance index variation tendency, comprising:
Index selection module, for obtaining n network performance index in the geographical granularity of setting at continuous m time point
Index value obtains the original performance index matrix of n × m dimension of n network performance index;Wherein, the n is just whole greater than 1
Number, the m are the positive integer not less than 1;
Covariance computing module, for calculating the obtained original performance index matrix of the index selection module
Covariance matrix;
Index distance calculation module, any desired indicator for being directed in the n network performance index, according to described
The institute that the obtained original performance index matrix of index selection module and the covariance computing module are calculated
Covariance matrix is stated, the mahalanobis distance of each other indexs and the desired indicator in the n network performance index is calculated;
Indicator analysis module, other each indexs for being calculated according to the index distance calculation module with it is described pre-
The mahalanobis distance for determining index determines immediate with the variation tendency of the desired indicator in the n network performance index
Index and index with the variation tendency disparity of the desired indicator.
Optionally, the covariance computing module is specifically used for being calculated by the following formula the original performance index matrix
The covariance matrix ∑ of X:
Wherein, μ i is the desired value of i-th of index Xi, (i, j) a element ∑ ij in covariance matrix ∑ be Xi with
The covariance of j-th of index Xj, i, j are respectively any positive integer for being not more than n.
Optionally, the index distance calculation module refers to specifically for being calculated by the following formula the n network performance
The mahalanobis distance Dij of any index Xj and the desired indicator Xi in mark:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and i, j are respectively not more than any just whole of N
Number.
Optionally, the indicator analysis module, specifically for according to other each indexs and the predetermined finger being calculated
Target mahalanobis distance obtains the mahalanobis distance matrix of 1 × (n-1) dimension;And according to each matrix element in the mahalanobis distance matrix
The size of element, is ranked up each matrix element in the mahalanobis distance matrix according to sequence from big to small;And it will arrange
The most preceding index of sequence as the immediate index of variation tendency with the desired indicator in the n network performance index,
The last index that will sort is as the index of the variation tendency disparity with the desired indicator.
Optionally, the equipment further includes normalization module:
The normalization module, for calculating the association side of the original performance index matrix in the covariance computing module
Before poor matrix, operation is normalized to each index value for the n network performance index that the index selection module is got.
Optionally, the equipment further includes index evaluation module:
The index evaluation module, for the indicator analysis module according to other each indexs being calculated with it is described
The mahalanobis distance of desired indicator determines closest with the variation tendency of the desired indicator in the n network performance index
Index and with after the index of the variation tendency disparity of the desired indicator, according to determining with the desired indicator
The immediate index of variation tendency and index with the variation tendency disparity of the desired indicator, determine network parameter
Adjustment influence that index is changed it is whether reasonable, to judge whether network parameter adjustment reasonable.
The present invention has the beneficial effect that:
The embodiment of the invention provides a kind of method and apparatus for assessing network performance index variation tendency, the method packets
It includes: obtaining n network performance index in the geographical granularity of setting in the index value at continuous m time point, obtain n network performance
The original performance index matrix of n × m dimension of index, then calculates n by calculating the covariance matrix of original performance index matrix
The mahalanobis distance of each other indexs and desired indicator in a network performance index, and then obtain the variation with the desired indicator
The immediate index of trend and index with the variation tendency disparity of the desired indicator, wherein the n is greater than 1
Positive integer, the m is positive integer not less than 1, to realize judging automatically for network performance index variation tendency, is improved
The analysis efficiency of the network optimization and the comprehensive and scalability that network optimization analysis can be improved.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill in field, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
The process that Fig. 1 show the method that network performance index variation tendency is assessed described in the embodiment of the present invention one is shown
It is intended to;
The structure that Fig. 2 show the equipment that network performance index variation tendency is assessed described in the embodiment of the present invention two is shown
It is intended to.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
It is described in detail to one step, it is clear that described embodiments are only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
All other embodiment, shall fall within the protection scope of the present invention.
Embodiment one:
The embodiment of the present invention one provides a kind of method for assessing network performance index variation tendency, as shown in Figure 1, it is
The flow diagram of the method for network performance index variation tendency, the assessment internetworking are assessed described in the embodiment of the present invention one
The method of energy index variation tendency can comprise the following steps that
Step 101: obtaining n network performance index in the geographical granularity of setting in the index value at continuous m time point, obtain
To the original performance index matrix of n × m dimension of n network performance index;Wherein, the n is the positive integer greater than 1, and the m is
Positive integer not less than 1.
Wherein, the geographical granularity of the setting can carry out flexible setting according to the actual situation, such as may be configured as the cell of setting,
The region of setting or the whole network etc.;In addition, the value of described n, m can also carry out flexible setting according to the actual situation, e.g., in order to make
Network analysis result it is more accurate, the all-network performance that the value of the n usually may be configured as setting in geographical granularity refers to
Target total number, the value of the m may be configured as relatively large numerical value etc.;Further more, the time in the m time point
Point can value be according to the actual situation hour or day etc., and the embodiment of the present invention does not repeat this.
Optionally, n network performance index in geographical granularity is set in the finger at continuous m time point according to what is got
The original performance index matrix X of scale value, n × m dimension of the obtained n network performance index can be as follows:
Wherein, all achievement datas of each time point can be used as an analysis sample, that is, can get m of n index
Sample.In addition, (i, j) a element Xij in the original performance index matrix X may refer to i-th (i=1 ..., n) a net
I-th index value of the network performance indicator in the index value at jth (j=1 ..., m) a time point, i.e. j-th of sample.
For example, if wishing the network performance index for daily analyzing certain cell, that is, if in the geographical granularity of the setting got
Index value of the n network performance index at continuous m time point is the continuous 30 days fingers of all n network performance indexes of certain cell
It marks sampled value (such as every 15 minutes sampled points), then the primitiveness of n × m dimension of the obtained n network performance index
Energy index matrix X is daily to handle a large amount of index sampled value of each network performance index, obtained all nets of the cell
Network performance indicator continuous 30 days, size be index value original matrix that N row 30 arranges, the embodiment of the present invention do not repeat this.
Step 102: calculating the covariance matrix of the original performance index matrix.
Optionally, in embodiment of the present invention, it can be calculated by the following formula the original performance index matrix X's
Covariance matrix ∑:
Wherein, μ i is the desired value of i-th of index Xi, i.e. μ i=E (Xi);(i, j) a member in covariance matrix ∑
Plain ∑ ij is the covariance of Xi and j-th of index Xj, and herein, i, j are respectively any positive integer for being not more than n.In addition, it is necessary to say
It is bright, index Xi, Xj described herein usually may refer to include corresponding m index value indicator vector, the present invention is real
It applies example and this is not repeated.
Step 103: for any desired indicator in the n network performance index, according to the original performance index
Matrix and the covariance matrix calculate other each indexs and the desired indicator in the n network performance index
Mahalanobis distance.
Wherein, mahalanobis distance is proposed by India's statistician's Mahalanobis, can indicate the covariances of data away from
From, and be a kind of method of similarity for effectively calculating two unknown sample collection.Unlike Euclidean distance, mahalanobis distance
It is contemplated that between various characteristics connection (for example, the information about height can bring the information about weight, because
It is related for the two) and be that scale is unrelated (scale-invariant), i.e., independently of measurement scale.Thus, with general
All uniformly the treating per one-dimensional and do not consider that the Euclidean distance of the connection between feature is compared of sample, can avoid sample
Certain information missing so that the calculated result of similarity is more accurate.
Optionally, it in embodiment of the present invention, can be calculated by the following formula in the n network performance index
The mahalanobis distance Dij of any index Xj and the desired indicator Xi:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and, i, j described herein are respectively little
In any positive integer of n;In addition, it is necessary to which explanation, index Xi, Xj described herein usually may refer to include corresponding m
The indicator vector of a index value, the embodiment of the present invention do not repeat this.
Step 104: according to the mahalanobis distance of each other indexs and the desired indicator for being calculated, determining the n
The variation with the immediate index of variation tendency of the desired indicator and with the desired indicator in network performance index
The index of trend disparity.
Optionally, according to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the n net is determined
Becoming in network performance indicator with the immediate index of variation tendency of the desired indicator and with the variation of the desired indicator
Potential difference is away from maximum index, it may include:
According to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the geneva of 1 × (n-1) dimension is obtained
Distance matrix is to get the mahalanobis distance vector tieed up to 1 × (n-1);It should be noted that if can be examined when calculating mahalanobis distance
Consider the mahalanobis distance between the desired indicator and the desired indicator itself, then obtained mahalanobis distance matrix can be 1 × n
The mahalanobis distance matrix of dimension;
According to the size of each matrix element in the mahalanobis distance matrix, according to sequence from big to small to the geneva
Each matrix element in distance matrix is ranked up;
The most preceding index of sorting as in the n network performance index with the variation tendency of the desired indicator most
Close index, the last index that will sort is as the index of the variation tendency disparity with the desired indicator.
Certainly, it should be noted that can also be according to sequence from small to large to each matrix in the mahalanobis distance matrix
Element is ranked up, and the last index that will sort is as the change with the desired indicator in the n network performance index
The immediate index of change trend, the most preceding index that will sort is as the finger of the variation tendency disparity with the desired indicator
Mark.
That is, the variation between the smaller index and desired indicator of mahalanobis distance is closer, mahalanobis distance is bigger
Difference in change between index and desired indicator is away from bigger.
Optionally, since different network performance indexes has different dimensions, thus, for each index of lateral comparison
Variation, first can be normalized operation to each achievement data sample, that is, after obtaining the original performance index matrix,
And before the covariance matrix for calculating the original performance index matrix, the method may also include that
Operation is normalized to each index value of the n network performance index got, so that subsequent carry out geneva
When the calculating of distance, each normalization index Xjnorm can be calculated to the mahalanobis distance of desired indicator Xinorm.
Optionally, following formula can be used, behaviour is normalized to each index value of the n network performance index got
Make, each index normalized into [0,1] section:
Wherein, Xij may refer to i-th (i=1 ..., n) a network performance index at jth (j=1 ..., m) a time point
Index value, XijnormIt may refer to the normalized value of the Xij;Min (Xi) may refer to i-th (i=1 ..., n) a internetworking
The corresponding minimum index value of energy index, max (Xi) may refer to the corresponding maximum of i-th (i=1 ..., n) a network performance index
Index value.
Further, in the mahalanobis distance according to each other indexs and the desired indicator being calculated, the n is determined
The change with the immediate index of variation tendency of the desired indicator and with the desired indicator in a network performance index
After the index of change trend disparity, the method be can comprise the further steps of:
According to the determining change with the immediate index of desired indicator variation tendency and with the desired indicator
The index of change trend disparity, whether the influence for determining that the adjustment of network parameter changes index is reasonable, to judge that network is joined
Whether number adjustment is reasonable.
That is, it is assumed that index i is decline in observation cycle, and network optimization personnel need in the analysis process
Solving which index is also decline, and which index can rise, and whether the involved KPI of network adjustment reaches expected, and
Whether other indexs have impliedly been fed through to.Optimization personnel extract a collection of KPI data, are defined as " experience collection according to Optimization Experience
Close ", it is that adjustment personnel are necessarily required to grasp.
But for the strong correlation index for being not belonging to " experience set ", network optimization personnel then can be according to mahalanobis distance matrix
D is found out in all n-1 network performance indexes, with the finger that index i variation tendency similarity is big or variation tendency similarity is small
Mark analyzes whether the influence that the adjustment of parameter change index is reasonable, finds out the outer shadow that may be omitted of network optimization personnel experience
The factor of sound, to improve the accuracy of the network optimization.
The embodiment of the present invention one provides a kind of method for assessing network performance index variation tendency, and the method can wrap
It includes: obtaining n network performance index in the geographical granularity of setting in the index value at continuous m time point, obtain n network performance
The original performance index matrix of n × m dimension of index, then calculates n by calculating the covariance matrix of original performance index matrix
The mahalanobis distance of each other indexs and desired indicator in a network performance index, and then obtain the variation with the desired indicator
The immediate index of trend and index with the variation tendency disparity of the desired indicator, wherein the n is greater than 1
Positive integer, the m is positive integer not less than 1.
That is, in the technical solution described in the embodiment of the present invention one, it can be by data mining, to based on heterogeneous networks
The network performance index data of parameter are associated, cluster, to grope and find the unknown variations trend similarity between index, in advance
Phase solves the problems such as potential network performance failure, to realize judging automatically for network performance index variation tendency, thus relatively
For existing way, since network optimization personnel are when carrying out the analysis of index variation tendency, without again passing by picture index
The operation such as tendency chart and eye-observation improves analysis efficiency so as to greatly save manpower and time;In addition,
Due to method provided by according to embodiments of the present invention, network optimization personnel can the variation quickly to all indexs become
Gesture is all analyzed, and is chosen the limitation that limited index is analyzed by previous Optimization Experience to breach, can be accomplished
Influence after analysis network index situation and overall understanding prioritization scheme are implemented comprehensively to existing net quality, so that network also can be improved
Optimize the comprehensive and scalability of analysis.
Embodiment two:
Based on inventive concept identical with the present embodiment one, the embodiment of the present application two provides a kind of assessment network performance and refers to
The equipment for marking variation tendency, the specific implementation of the equipment can be found in the associated description in above method embodiment one, repeat place
It repeats no more, as shown in Fig. 2, the equipment is main can include:
Index selection module 21 can be used for obtaining n network performance index in the geographical granularity of setting in the continuous m time
The index value of point obtains the original performance index matrix of n × m dimension of n network performance index;Wherein, the n is greater than 1
Positive integer, the m are the positive integer not less than 1;
Covariance computing module 22 can be used for calculating the obtained original performance index of the index selection module 21
The covariance matrix of matrix;
Index distance calculation module 23 can be used for any desired indicator being directed in the n network performance index, according to
The obtained original performance index matrix of index selection module 21 and the covariance computing module calculate
The covariance matrix arrived calculates the geneva of each other indexs and the desired indicator in the n network performance index
Distance;
Indicator analysis module 24, other each indexs that can be used for being calculated according to the index distance calculation module 23 with
The mahalanobis distance of the desired indicator, determine in the n network performance index with the variation tendency of the desired indicator most
Close index and the index with the variation tendency disparity of the desired indicator.
Optionally, the covariance computing module 22, which is particularly used in, is calculated by the following formula the original performance index
The covariance matrix ∑ of matrix X:
Wherein, μ i is the desired value of i-th of index Xi, (i, j) a element ∑ ij in covariance matrix ∑ be Xi with
The covariance of j-th of index Xj, i, j are respectively any positive integer for being not more than n.
Optionally, the index distance calculation module 23, which is particularly used in, is calculated by the following formula the n internetworking
The mahalanobis distance Dij of any index Xj and the desired indicator Xi in energy index:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and i, j are respectively not more than any just whole of n
Number.
Optionally, the indicator analysis module 24 is particularly used in is calculated according to the index distance calculation module 23
Each other indexs and the desired indicator mahalanobis distance, obtain 1 × (n-1) dimension mahalanobis distance matrix;And according to described
The size of each matrix element in mahalanobis distance matrix, according to sequence from big to small to each square in the mahalanobis distance matrix
Array element element is ranked up;And will sort most preceding index as in the n network performance index with the desired indicator
The immediate index of variation tendency, the last index that will sort is as the variation tendency with the desired indicator with the biggest gap
Index.
Optionally, the equipment may also include normalization module 25:
The normalization module 25 can be used for calculating the original performance index matrix in the covariance computing module 22
Covariance matrix before, each index value for the n network performance index that the index selection module 21 is got is returned
One changes operation.
Optionally, the equipment may also include index evaluation module 26:
The index evaluation module 26 can be used in the indicator analysis module 24 according to the index distance calculation module
The mahalanobis distance of 23 each other indexs and the desired indicator being calculated, determine in the n network performance index with
The immediate index of the variation tendency of the desired indicator and index with the variation tendency disparity of the desired indicator
Later, become according to determining with the immediate index of desired indicator variation tendency and with the variation of the desired indicator
Potential difference is away from maximum index, and whether the influence for determining that the adjustment of network parameter changes index is reasonable, to judge network parameter tune
It is whole whether reasonable.
It will be understood by those skilled in the art that the embodiment of the present invention can provide as method, apparatus (equipment) or computer journey
Sequence product.Therefore, complete hardware embodiment, complete software embodiment or combining software and hardware aspects can be used in the present invention
The form of embodiment.Moreover, it wherein includes the calculating of computer usable program code that the present invention, which can be used in one or more,
The computer program implemented in machine usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The present invention be referring to according to the method for the embodiment of the present invention, the flow chart of device (equipment) and computer program product
And/or block diagram describes.It should be understood that each process in flowchart and/or the block diagram can be realized by computer program instructions
And/or the combination of the process and/or box in box and flowchart and/or the block diagram.It can provide these computer programs to refer to
Enable the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to generate
One machine so that by the instruction that the processor of computer or other programmable data processing devices executes generate for realizing
The device for the function of being specified in one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although preferred embodiments of the present invention have been described, it is created once a person skilled in the art knows basic
Property concept, then additional changes and modifications may be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the scope of the invention.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (10)
1. a kind of method for assessing network performance index variation tendency characterized by comprising
N network performance index in the geographical granularity of setting is obtained in the index value at continuous m time point, obtains n internetworking
The original performance index matrix of n × m dimension of energy index, wherein the n is the positive integer greater than 1, and the m is just not less than 1
Integer;
Calculate the covariance matrix of the original performance index matrix;
For any desired indicator in the n network performance index, according to the original performance index matrix and described
Covariance matrix calculates the mahalanobis distance of each other indexs and the desired indicator in the n network performance index;
According to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the n network performance index is determined
In it is with the biggest gap with the immediate index of variation tendency of the desired indicator and with the variation tendency of the desired indicator
Index;
Become according to determining with the immediate index of desired indicator variation tendency and with the variation of the desired indicator
Potential difference is away from maximum index, and whether the influence for determining that the adjustment of network parameter changes index is reasonable, to judge network parameter tune
It is whole whether reasonable.
2. the method as described in claim 1, which is characterized in that be calculated by the following formula the original performance index matrix X
Covariance matrix ∑:
Wherein, μ i is the desired value of i-th of index Xi, and (i, j) a element ∑ ij in covariance matrix ∑ is Xi and j-th
The covariance of index Xj, i, j are respectively any positive integer for being not more than n.
3. the method as described in claim 1, which is characterized in that be calculated by the following formula in the n network performance index
Any index Xj and the desired indicator Xi mahalanobis distance Dij:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and i, j are respectively any positive integer for being not more than n.
4. the method as described in claim 1, which is characterized in that according to other each indexs and the desired indicator being calculated
Mahalanobis distance, determine in the n network performance index with the immediate index of variation tendency of the desired indicator with
And the index with the variation tendency disparity of the desired indicator, comprising:
According to the mahalanobis distance of each other indexs and the desired indicator for being calculated, the mahalanobis distance of 1 × (n-1) dimension is obtained
Matrix;
According to the size of each matrix element in the mahalanobis distance matrix, according to sequence from big to small to the mahalanobis distance
Each matrix element in matrix is ranked up;
The most preceding index that will sort is as closest with the variation tendency of the desired indicator in the n network performance index
Index, the last index that will sort is as the index of the variation tendency disparity with the desired indicator.
5. the method as described in claim 1, which is characterized in that in the covariance matrix for calculating the original performance index matrix
Before, the method also includes:
Operation is normalized to each index value of the n network performance index got.
6. a kind of equipment for assessing network performance index variation tendency characterized by comprising
Index selection module sets n network performance index in geographical granularity in the index at continuous m time point for obtaining
Value obtains the original performance index matrix of n × m dimension of n network performance index;Wherein, the n is the positive integer greater than 1, institute
Stating m is the positive integer not less than 1;
Covariance computing module, for calculating the association side of the obtained original performance index matrix of the index selection module
Poor matrix;
Index distance calculation module, any desired indicator for being directed in the n network performance index, according to the index
Obtain the association that the obtained original performance index matrix of module and the covariance computing module are calculated
Variance matrix calculates the mahalanobis distance of each other indexs and the desired indicator in the n network performance index;
Indicator analysis module, other each indexs and the predetermined finger for being calculated according to the index distance calculation module
Target mahalanobis distance determines the immediate index of variation tendency with the desired indicator in the n network performance index
And the index with the variation tendency disparity of the desired indicator;
The index evaluation module, for according to it is determining with the immediate index of desired indicator variation tendency and with
The index of the variation tendency disparity of the desired indicator, determines whether the adjustment of network parameter closes the influence that index changes
Reason, to judge whether network parameter adjustment is reasonable.
7. equipment as claimed in claim 6, which is characterized in that the covariance computing module is specifically used for passing through following formula
Calculate the covariance matrix ∑ of the original performance index matrix X:
Wherein, μ i is the desired value of i-th of index Xi, and (i, j) a element ∑ ij in covariance matrix ∑ is Xi and j-th
The covariance of index Xj, i, j are respectively any positive integer for being not more than n.
8. equipment as claimed in claim 6, which is characterized in that the index distance calculation module is specifically used for passing through following public affairs
Formula calculates the mahalanobis distance Dij of any index Xj and the desired indicator Xi in the n network performance index:
Wherein, ∑ is the covariance matrix of the original performance index matrix X, and i, j are respectively any positive integer for being not more than n.
9. equipment as claimed in claim 6, which is characterized in that
The indicator analysis module, specifically for according to the geneva of other each indexs for being calculated and the desired indicator away from
From, obtain 1 × (n-1) dimension mahalanobis distance matrix;And according to the size of each matrix element in the mahalanobis distance matrix, press
Each matrix element in the mahalanobis distance matrix is ranked up according to sequence from big to small;And the finger that sequence is most preceding
Be denoted as the immediate index of variation tendency with the desired indicator in the n network performance index, will sequence it is last
Index of the index as the variation tendency disparity with the desired indicator.
10. equipment as claimed in claim 6, which is characterized in that the equipment further includes normalization module:
The normalization module, for calculating the covariance square of the original performance index matrix in the covariance computing module
Before battle array, operation is normalized to each index value for the n network performance index that the index selection module is got.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510557027.5A CN106487570B (en) | 2015-09-02 | 2015-09-02 | A kind of method and apparatus for assessing network performance index variation tendency |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510557027.5A CN106487570B (en) | 2015-09-02 | 2015-09-02 | A kind of method and apparatus for assessing network performance index variation tendency |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106487570A CN106487570A (en) | 2017-03-08 |
CN106487570B true CN106487570B (en) | 2019-06-25 |
Family
ID=58238163
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510557027.5A Active CN106487570B (en) | 2015-09-02 | 2015-09-02 | A kind of method and apparatus for assessing network performance index variation tendency |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106487570B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110555546B (en) * | 2019-07-31 | 2022-04-08 | 烽火通信科技股份有限公司 | Updating method and system for optical performance degradation trend prediction |
CN114286360B (en) * | 2020-09-27 | 2023-09-05 | 中国移动通信集团设计院有限公司 | Wireless network communication optimization method and device, electronic equipment and storage medium |
CN115334559B (en) * | 2022-08-19 | 2024-07-02 | 中国联合网络通信集团有限公司 | Network detection method, device, equipment and medium |
Citations (4)
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 |
CN103560925A (en) * | 2011-05-04 | 2014-02-05 | 成都勤智数码科技股份有限公司 | IT operation and maintenance index forecasting method utilizing relevance |
CN103780415A (en) * | 2012-10-22 | 2014-05-07 | 华为技术服务有限公司 | Method and apparatus for monitoring key performance indicator |
JP2015001823A (en) * | 2013-06-14 | 2015-01-05 | ヤンマー株式会社 | Prediction apparatus, prediction method, and computer program |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7620204B2 (en) * | 2006-02-09 | 2009-11-17 | Mitsubishi Electric Research Laboratories, Inc. | Method for tracking objects in videos using covariance matrices |
-
2015
- 2015-09-02 CN CN201510557027.5A patent/CN106487570B/en active Active
Patent Citations (4)
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 |
CN103560925A (en) * | 2011-05-04 | 2014-02-05 | 成都勤智数码科技股份有限公司 | IT operation and maintenance index forecasting method utilizing relevance |
CN103780415A (en) * | 2012-10-22 | 2014-05-07 | 华为技术服务有限公司 | Method and apparatus for monitoring key performance indicator |
JP2015001823A (en) * | 2013-06-14 | 2015-01-05 | ヤンマー株式会社 | Prediction apparatus, prediction method, and computer program |
Also Published As
Publication number | Publication date |
---|---|
CN106487570A (en) | 2017-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2020207214A1 (en) | Data processing method and apparatus, electronic device and storage medium | |
CN110298249A (en) | Face identification method, device, terminal and storage medium | |
CN104462184B (en) | A kind of large-scale data abnormality recognition method based on two-way sampling combination | |
CN107038167A (en) | Big data excavating analysis system and its analysis method based on model evaluation | |
CN111091247A (en) | Power load prediction method and device based on deep neural network model fusion | |
CN108805213B (en) | Power load curve double-layer spectral clustering method considering wavelet entropy dimensionality reduction | |
CN106487570B (en) | A kind of method and apparatus for assessing network performance index variation tendency | |
CN104408667A (en) | Method and system for comprehensively evaluating power quality | |
CN110097088A (en) | A kind of dynamic multi-objective evolvement method based on transfer learning Yu particular point strategy | |
CN103559303A (en) | Evaluation and selection method for data mining algorithm | |
CN103888541A (en) | Method and system for discovering cells fused with topology potential and spectral clustering | |
CN111178533B (en) | Method and device for realizing automatic semi-supervised machine learning | |
CN104994366A (en) | FCM video key frame extracting method based on feature weighing | |
CN112800231A (en) | Power data verification method and device, computer equipment and storage medium | |
CN116644184B (en) | Human resource information management system based on data clustering | |
CN114912720A (en) | Memory network-based power load prediction method, device, terminal and storage medium | |
CN110110339A (en) | A kind of hydrologic forecast error calibration method and system a few days ago | |
CN103095534A (en) | Network attack effect evaluation method for two-dimensional K-means entropy | |
CN116365519B (en) | Power load prediction method, system, storage medium and equipment | |
Dong | Application of Big Data Mining Technology in Blockchain Computing | |
Hai-Jun et al. | Fuzzy entropy clustering using possibilistic approach | |
CN114155578A (en) | Portrait clustering method, device, electronic equipment and storage medium | |
CN106487547B (en) | A kind of method and device for assessing network performance index variation tendency | |
CN111222688A (en) | Daily load prediction method for commercial building | |
CN110110906A (en) | A kind of survival risk modeling method based on Efron near-optimal |
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 |