CN110348534A - Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis - Google Patents
Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis Download PDFInfo
- Publication number
- CN110348534A CN110348534A CN201910647142.XA CN201910647142A CN110348534A CN 110348534 A CN110348534 A CN 110348534A CN 201910647142 A CN201910647142 A CN 201910647142A CN 110348534 A CN110348534 A CN 110348534A
- Authority
- CN
- China
- Prior art keywords
- flow
- data
- sequence
- intension
- sample matrix
- 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.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 34
- 238000005206 flow analysis Methods 0.000 title claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 11
- 238000000513 principal component analysis Methods 0.000 claims abstract description 11
- 230000000694 effects Effects 0.000 claims abstract description 10
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 8
- 238000004891 communication Methods 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 34
- 239000013598 vector Substances 0.000 claims description 23
- 239000000203 mixture Substances 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 description 2
- 239000004615 ingredient Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The present invention relates to the Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis of computer communication field, step includes: 1, carries out intension mode decomposition to flow sequence data, obtains intension mode function sequence;2, according to flow effect variable factors and the intension mode function, establish sample matrix;3, principal component analysis is carried out to sample matrix, obtains the main variables sequence of each sample matrix, main variables sequence embodies the influence factor of changes in flow rate and the feature of flow.Result of study shows that compared with common Data Dimensionality Reduction and characteristic analysis method, method proposed by the present invention can accurately analyze the influence factor and traffic characteristic of changes in flow rate.
Description
Technical field
Data Dimensionality Reduction and signature analysis side the present invention relates to computer communication field, in particular in a kind of flow analysis
Method.
Background technique
With computer network, cordless communication network technology fast development and constantly upgrading, network flow it is strong
Degree, type are continuously increased, and data scale and dimension are increasingly complicated, form magnanimity traffic load.For these traffic load data,
It is desirable that finding the influence factor for influencing changes in flow rate, such as due to access customer number variation, the use of user in interchanger
The variation of bandwidth changes or switch traffic is dispatched some equipment fault of behavior or user's habit, or
Person is some order or some media event of service centre, these can be in guide data to the analysis of flow influence factor
The layout strategy of the scheduling of resource of the heart or interchanger realizes better service quality.
Since the factor for influencing traffic load is more, cause input space dimension excessively high, while between influence factor in itself
There is also correlations, also will affect the computational efficiency and analysis precision of traffic load analysis.In order to improve the accuracy of analysis, need
It eliminates the correlation between multifactor and rejects redundancy.Data Dimensionality Reduction and signature analysis extract data by rejecting redundancy
Main feature information, as far as possible with the main information of least message reflection initial data, to improve data mining efficiency.Generally
Ground, data characteristics is more, and the more the information that data include the abundanter, however in some cases, it may be deposited between these features
In potential redundancy, therefore progress Data Dimensionality Reduction and character representation are of great significance.The method of Data Reduction and dimensionality reduction at present
Have very much, such as principal component analysis (Principle Components Analysis, PCA), wavelet transform
(Discrete Wavelet Transform, DWT), singular value decomposition (Singular Value Decomposition,
SVD), manifold learning (Manifold Learning, ML) etc..These traditional methods change relatively slow, burst some
In the few time series data of influence factor, there is more successful application.Then for information flow, especially computer network
For the data traffic of network, rapidly, changes in flow rate mode is very various and different changing patteries are mingled in one for traffic intensity variation
It rises, the mode of existing relatively smooth change, also has tendency changing pattern and cyclically-varying mode, happen suddenly simultaneously in flow
Property it is also very strong, certain burst flows even account for the main component of data on flows, and flow effect factor usually only influence it is a small amount of
The flow rate mode of several aspects, when these flow rate mode features are mixed in together, point of traditional Data Dimensionality Reduction analysis method
Analysis effect is had a greatly reduced quality.
Summary of the invention
It is an object of the invention to overcome the above-mentioned deficiency in the presence of the prior art, the number in a kind of flow analysis is provided
According to dimensionality reduction and characteristic analysis method.
In order to achieve the above-mentioned object of the invention, the present invention provides following technical schemes:
Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis, step include:
S1 carries out intension mode decomposition to flow sequence data, obtains intension mode function sequence;
S2 establishes sample matrix according to flow effect variable factors and intension mode function;
S3 carries out principal component analysis to sample matrix, obtains the main variables sequence of each sample matrix, and principal component becomes
Amount sequence embodies the influence factor of changes in flow rate and the feature of flow.
Step S1 is specifically included:
S11, all extreme points for identifying outflow sequence data;
S12, the upper and lower envelope e for being fitted outflow sequence datasup(t) and elow(t), and envelope up and down is calculated
Average value m (t);
S13, flow sequence data x (t) is subtracted into envelope average value m (t) up and down, obtains c (t);
S14, judge whether c (t) meets preset two conditions, if it is satisfied, then regarding c (t) as the first rank intension mould
State function IMF1(t), it executes step S15 and otherwise regards c (t) as new flow sequence data, return step S11, Zhi Daoman
Preset two conditions of foot, export single order intension mode function IMF1(t);
S15, flow sequence data x (t) is subtracted into IMF1(t), new original signal r (t), return step S11 are obtained;
S16, when meeting the stop condition of successive ignition, obtain each rank intension mode function, the stopping item of successive ignition
Part is cn(t) and cn-1(t) meet formula:
Wherein, ε is minimum reference value, and n is the number of iterations, cn(t) the flow sequence data and packet obtained for nth iteration
The difference of winding thread mean value, cn-1It (t) is the difference of (n-1)th iteration obtained flow sequence data and envelope mean value.
Preset two conditions are as follows:
(1) in entire data segment, the number of the extreme point and number of zero crossing is equal or the number and mistake of extreme point
The difference of the number of zero point is no more than one;
(2) at any time, it the coenvelope line that is formed in c (t) by Local modulus maxima and is formed by local minizing point
The average value of lower envelope line be zero.
The specific steps of S2 are as follows: take the data of 1~n rank intension mode function sequence to be sampled, and by sampled data with
Flow effect factor data is combined into the sample matrix of corresponding each rank intension mode sequence together.
Principal component analysis is to adopt for each intension mode function sampled value of flow sequence data with influence factor variable
What the sample matrix of sample value composition carried out, step includes:
S21, sample matrix is standardized, obtains standardization sample matrix;
S22, according to standardized sample matrix, establish covariance matrix R, and calculate eigenvalue λ and feature vector L;
S23, according to eigenvalue λ, calculate the contribution rate and contribution rate of accumulative total of each principal component, and establish eigenvalue λ, feature to
Measure the one-to-one relationship between L and contribution rate;
S24, the corresponding feature vector of contribution rate according to standardization sample matrix and each principal component, determine each intension mould
The corresponding chief composition series vector of state function.
The formula that sample matrix standardization is used are as follows:
Wherein, X'ijFor standardization after i-th of sample j-th of feature data,It is j-th
The arithmetic mean of instantaneous value of feature,For the standard deviation of j-th of feature, XijFor different modalities function sequence
The sample matrix data of column and influence factor variable composition, m are the number of influence factor variable.
The calculation formula of covariance matrix R are as follows:Covariance matrix R, eigenvalue λ and feature vector L
Relationship are as follows: RL=λ L, wherein X be the different modalities sequence of function and influence factor variable composition sample matrix, m be influence because
The number of plain variable.
The calculation formula of principal component contributor rate are as follows:
Wherein, λi(i=1,2 ..., p) is the specific value of the characteristic value of covariance matrix R.
Chief composition series vector calculation formula are as follows:
Zl=XsLl
Wherein, Xs is the sample matrix after standardization, LlCorrespond to the eigenvalue λ of first of contribution ratelFeature vector.
A kind of system of Data Dimensionality Reduction and characteristic analysis method using in flow analysis, including at least one processor,
And the memory being connect at least one processor communication;Memory is stored with the finger that can be executed by least one processor
It enables, instruction is executed by least one processor, so that at least one processor is able to carry out any of the above-described method.
Compared with prior art, beneficial effects of the present invention:
The invention proposes a kind of flows point that Data Dimensionality Reduction and signature analysis are carried out for complicated network traffic patterns
Analysis method.The decomposition of mode is changed to data on flows first, isolates the flow mode sequence with different versions,
Then traffic characteristic and main changing factor are analyzed using principal component analytical method.Result of study shows to drop with common data
Peacekeeping characteristic analysis method is compared, method proposed by the present invention can accurately analyze changes in flow rate influence factor and
Traffic characteristic.
Detailed description of the invention
Fig. 1 is the flow chart of the Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis.
Specific embodiment
Below with reference to test example and specific embodiment, the present invention is described in further detail.But this should not be understood
It is all that this is belonged to based on the technology that the content of present invention is realized for the scope of the above subject matter of the present invention is limited to the following embodiments
The range of invention.
Embodiment 1
The flow chart of Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis is as shown in Figure 1.First to load flow
It measures data and carries out intension mode decomposition, then each modal data is adopted by time interval identical with flow effect factor
Sample obtains the data sequence of each flow mode, on this basis, establishes the sample matrix of flow and flow effect factor data,
Principal component analysis is carried out to the sample matrix of reflection different mode, to obtain the main changing pattern of different mode.Specific steps
It is as follows:
Step 1: by flow sequence x (t) as initial original signal, carrying out intension mode decomposition.Original is identified first
All extreme points of beginning signal fit the upper and lower envelope e of signal respectivelysup(t), elow(t), it is flat to calculate envelope up and down
Mean value:
M (t)=[esup(t)+elow(t)]/2 (1)
Step 2: original signal x (t) being subtracted into m (t) and is obtained: c (t)=x (t)-m (t).It is following to judge whether c (t) meets
Two conditions:
1) in entire data segment, the number of extreme point and the number of zero crossing it is equal or difference be no more than one.
2) it the coenvelope line that at any time, is formed in c (t) by Local modulus maxima and is formed by local minizing point
The average value of lower envelope line is zero, i.e., upper and lower envelope is relative to time shaft Local Symmetric.
If it is satisfied, then regarding c (t) as the first rank intension mode function IMF1(t), otherwise, c (t) is regarded as new original
Beginning signal repeats step 1) and 2), until meeting above-mentioned two condition, then exportable single order intension mode function IMF1(t)。
Step 3: obtaining IMF1(t) after signal, original function x (t) is subtracted into IMF1(t) it obtains: r (t)=x (t)-
IMF1(t), r (t) is regarded as to new original signal, repeats step (1), (2).The above process passes through successive ignition, can be obtained each
Rank intension mode function IMF2(t), IMF3(t) ..., IMFn(t), until meeting following iteration stopping conditions:
Assuming that repeating by n times, the difference of obtained original signal and envelope mean value is cn(t), (n-1)th repetition obtains
Original signal and envelope mean value difference be cn-1(t).If cn(t) and cn-1(t) meet following formula:
Then iteration stopping, ε are the number of a very little, such as 0.01.N rank mode function IMF is obtained simultaneouslyn(t)=cn(t), and
Residual error is rn(t)=n-1 rank original signal IMFn(t)。
Step 4: take the data of 1~n rank intension mode function sequence to be sampled, and by sampled data and flow effect because
Prime number is according to the data sample matrix for being combined into corresponding each rank intension mode sequence together.
Step 5: principal component analysis being carried out to each rank sample matrix, is obtained a series of by the different modalities sequence of function and influence
The main variables sequence of the composite sequence of variable factors composition.
If being X by the sample sequence that some IMF sequence of function samples11, X12, X13..., X1n, by influence factor variable
Obtained sample sequence is respectively Xj1, Xj2, Xj3..., Xjn, j ∈ [2, m] then can establish m × n dimension data sample matrix:
The detailed process for carrying out principal component analysis to above-mentioned sample matrix is described as follows:
1) data normalization
Data-standardizing formula are as follows:
Wherein, X'ijFor standardization after i-th of sample j-th of feature data, while define standardization after by institute
There is X'ijThe matrix of composition is Xs;WithThe arithmetic of respectively j-th feature
Average and standard deviation.
2) covariance matrix R is established, eigenvalue λ and feature vector L are calculated
RL=λ L (6)
3) it calculates each principal component contributor rate and contribution rate of accumulative total, contribution rate is mainly used to judge the corresponding feature of each characteristic value
Weight of the vector in all feature vectors, while contribution rate is also used to judge the corresponding master of the corresponding feature vector of each characteristic value
Weight of the component vector in all principal component vectors.
The contribution rate of first of principal component are as follows:
Contribution rate of accumulative total:
Wherein λi(i=1,2 ..., p) is the specific value of the eigenvalue λ of covariance matrix R, i.e. the variable λ feature that is R
The General Expression of value, and λiIt is the specific value of the characteristic value of R matrix.
4) chief composition series of corresponding each rank IMF mode function and all influence factors are determined.
For a certain IMFiSequence, first of chief composition series vector are determined by following formula:
Zl=XsLl (9)
Wherein Xs is the sample matrix after X standardization, LlCorrespond to the eigenvalue λ with first of contribution ratelFeature
Vector, l=1 ... p.Take how many a principal components that can be determined by user, generally, principal component vector only takes the first two contribution rate
Maximum vector.
The cardinal principle of the above process is after carrying out intension model analysis to original flow sequence data x (t), to obtain
The the i-th rank intension mode sequential sampling data X arrivedi1, Xi2, Xi3..., Xin, (i=1,2 ..., n), be have certain basic frequency at
The approximate period delta data sequence divided, frequency change from high to low with the order i of intension mode function.These intension mode
The sequence of function reflects the ingredient of the variation of the different cycles in former data sequence x (t), and frequency is higher, and the periodicity of variation is got over
By force;Frequency is lower, and the tendency of variation is more significant.But the sequence of these approximate periods variation is also very coarse, may still mix
There are the data component of mutability and the data of other frequency contents.Therefore further principal component analysis is carried out.Pass through principal component
Analysis, obtains p chief composition series vector Z of the rank intension mode function sample sequencel(1), Zl(2) ..., Zl(n), l=
1 ... .p, each chief composition series have corresponded to the rank intension mode function sampled value sequence (Xi1, Xi2, Xi3..., Xin) in
The change frequency of one subsequence, the subsequence can be obtained by Fourier analysis, and the frequency of different chief composition series is not
Together.The ingredient with different change frequencies has been separated from the rank intension mode function sequence thus and has been come out, and has been obtained
Specific gravity of this changing pattern in i-th of intension mode function sequence.
Claims (10)
1. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis, which is characterized in that step includes:
S1 carries out intension mode decomposition to flow sequence data, obtains intension mode function sequence;
S2 establishes sample matrix according to flow effect variable factors and the intension mode function;
S3 carries out principal component analysis to the sample matrix, obtains the main variables sequence of each sample matrix, described
Main variables sequence embodies the influence factor of changes in flow rate and the feature of flow.
2. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as described in claim 1, which is characterized in that step
S1 is specifically included:
S11, all extreme points for identifying the flow sequence data;
S12, the upper and lower envelope e for fitting the flow sequence datasup(t) and elow(t), and envelope up and down is calculated
Average value m (t);
S13, flow sequence data x (t) is subtracted into envelope average value m (t) up and down, obtains c (t);
S14, judge whether c (t) meets preset two conditions, if it is satisfied, then regarding c (t) as the first rank intension mode letter
Number IMF1(t), it executes step S15 and otherwise regards c (t) as new flow sequence data, return step S11, until meeting
Preset two conditions are stated, single order intension mode function IMF is exported1(t);
S15, the flow sequence data x (t) is subtracted into IMF1(t), new original signal r (t), return step S11 are obtained;
S16, when meeting the stop condition of successive ignition, obtain each rank intension mode function, the stopping item of the successive ignition
Part is cn(t) and cn-1(t) meet formula:
Wherein, ε is minimum reference value, and n is the number of iterations, cn(t) the flow sequence data and envelope obtained for nth iteration
The difference of mean value, cn-1It (t) is the difference of (n-1)th iteration obtained flow sequence data and envelope mean value.
3. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 2, which is characterized in that described
Preset two conditions are as follows:
(1) in entire data segment, the number of the extreme point and number of zero crossing is equal or the number and zero crossing of extreme point
Number difference be no more than one;
(2) at any time, it the coenvelope line that is formed in c (t) by Local modulus maxima and is formed down by local minizing point
The average value of envelope is zero.
4. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 3, which is characterized in that S2's
Specific steps are as follows: take the data of intension mode function sequence described in 1~n rank to be sampled, and by sampled data and flow effect
Factor data is combined into the sample matrix of corresponding each rank intension mode sequence together.
5. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 4, which is characterized in that described
Principal component analysis is each intension mode function sampled value and influence factor variable the sampled value composition for flow sequence data
Sample matrix carry out, step includes:
S21, the sample matrix is standardized, obtains standardization sample matrix;
S22, according to the standardized sample matrix, establish covariance matrix R, and calculate eigenvalue λ and feature vector L;
S23, according to the eigenvalue λ, calculate the contribution rate and contribution rate of accumulative total of each principal component, and establish eigenvalue λ, feature to
Measure the one-to-one relationship between L and contribution rate;
S24, according to the corresponding feature vector of contribution rate of the standardization sample matrix and each principal component, determine in each
Contain the corresponding chief composition series vector of mode function.
6. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 5, which is characterized in that by institute
State the formula that sample matrix standardization uses are as follows:
Wherein, Xi'jFor standardization after i-th of sample j-th of feature data,For j-th of feature
Arithmetic mean of instantaneous value,For the standard deviation of j-th of feature, XijFor the different modalities sequence of function with
The sample matrix data of influence factor variable composition, m are the number of influence factor variable.
7. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 5, which is characterized in that described
The calculation formula of covariance matrix R are as follows:The pass of the covariance matrix R, eigenvalue λ and feature vector L
System are as follows: RL=λ L, wherein X is the sample matrix of the different modalities sequence of function and influence factor variable composition, and m is influence factor
The number of variable.
8. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 5, which is characterized in that described
The calculation formula of principal component contributor rate are as follows:
Wherein, λi(i=1,2 ..., p) is the specific value of the characteristic value of covariance matrix R.
9. Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis as claimed in claim 5, which is characterized in that described
Chief composition series vector calculation formula are as follows:
Zl=XsLl
Wherein, Xs is the sample matrix after standardization, LlCorrespond to the eigenvalue λ of first of contribution ratelFeature vector.
10. a kind of system of Data Dimensionality Reduction and characteristic analysis method using in flow analysis, which is characterized in that including at least one
A processor, and the memory being connect at least one described processor communication;The memory be stored with can by it is described extremely
The instruction that a few processor executes, described instruction are executed by least one described processor, so that at least one described processing
Device is able to carry out method described in any one of claims 1 to 9.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2019106320167 | 2019-07-12 | ||
CN201910632016 | 2019-07-12 |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110348534A true CN110348534A (en) | 2019-10-18 |
Family
ID=68175702
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910647142.XA Pending CN110348534A (en) | 2019-07-12 | 2019-07-17 | Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110348534A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113034940A (en) * | 2019-12-25 | 2021-06-25 | 中国航天系统工程有限公司 | Fisher ordered clustering-based single-point signalized intersection optimization timing method |
CN113836756A (en) * | 2021-11-29 | 2021-12-24 | 山东华尚电气有限公司 | Intelligent monitoring method and system for annealing process of three-dimensional wound core transformer |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102778355A (en) * | 2012-08-07 | 2012-11-14 | 北京交通大学 | Rolling bearing state identification method based on empirical mode decomposition (EMD) and principal component analysis (PCA) |
CN106646096A (en) * | 2016-11-15 | 2017-05-10 | 国网四川省电力公司广安供电公司 | Transformer fault classification and identification method based on vibration analysis method |
-
2019
- 2019-07-17 CN CN201910647142.XA patent/CN110348534A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102778355A (en) * | 2012-08-07 | 2012-11-14 | 北京交通大学 | Rolling bearing state identification method based on empirical mode decomposition (EMD) and principal component analysis (PCA) |
CN106646096A (en) * | 2016-11-15 | 2017-05-10 | 国网四川省电力公司广安供电公司 | Transformer fault classification and identification method based on vibration analysis method |
Non-Patent Citations (1)
Title |
---|
李洪安: "《信号稀疏化与应用》", 31 October 2017 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113034940A (en) * | 2019-12-25 | 2021-06-25 | 中国航天系统工程有限公司 | Fisher ordered clustering-based single-point signalized intersection optimization timing method |
CN113836756A (en) * | 2021-11-29 | 2021-12-24 | 山东华尚电气有限公司 | Intelligent monitoring method and system for annealing process of three-dimensional wound core transformer |
CN113836756B (en) * | 2021-11-29 | 2022-03-18 | 山东华尚电气有限公司 | Intelligent monitoring method and system for annealing process of three-dimensional wound core transformer |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Dalal et al. | Halo assembly bias in hierarchical structure formation | |
CN107704971A (en) | A kind of data processing method and device of real-time estimate airport security number | |
CN110348534A (en) | Data Dimensionality Reduction and characteristic analysis method in a kind of flow analysis | |
CN106326913A (en) | Money laundering account determination method and device | |
CN116994709B (en) | Personalized diet and exercise recommendation method and system and electronic equipment | |
CN105719045A (en) | Retention risk determiner | |
CN108416032A (en) | A kind of file classification method, device and storage medium | |
CN109299087A (en) | Data cache method, device, computer equipment and storage medium | |
CN112084752B (en) | Sentence marking method, device, equipment and storage medium based on natural language | |
CN102522709A (en) | Decision-making method and decision-making system for state overhaul of transformers | |
CN110197207A (en) | To not sorting out the method and relevant apparatus that user group is sorted out | |
CN108228684A (en) | Training method, device, electronic equipment and the computer storage media of Clustering Model | |
CN110390587A (en) | A kind of customer evaluation method and system | |
CN108229358A (en) | Index establishing method and device, electronic equipment, computer storage media, program | |
CN114139931A (en) | Enterprise data evaluation method and device, computer equipment and storage medium | |
CN110457478A (en) | Text compliance inspection method and device, electronic equipment and computer-readable medium | |
CN109377436A (en) | A kind of accurate monitoring and managing method of environment and device, terminal device and storage medium | |
CN106209443B (en) | Client relation management method based on cloud computing | |
CN114861800A (en) | Model training method, probability determination method, device, equipment, medium and product | |
CN110262795B (en) | Application system deployment architecture modeling and verification method | |
CN114462524A (en) | Clustering method for data center batch processing operation | |
CN106682273A (en) | Method for determining service life importance of series-parallel connection mixed spaceflight equipment system | |
CN106372670A (en) | Loyalty index prediction method based on improved nearest neighbor algorithm | |
CN110458383A (en) | Demand handles implementation method, device and the computer equipment of serviceization, storage medium | |
US20230041611A1 (en) | Method and apparatus for building database for retrieval, device and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191018 |
|
RJ01 | Rejection of invention patent application after publication |