CN104636492B - Dynamic data grading method based on fuzzy integral feature fusion - Google Patents
Dynamic data grading method based on fuzzy integral feature fusion Download PDFInfo
- Publication number
- CN104636492B CN104636492B CN201510095450.8A CN201510095450A CN104636492B CN 104636492 B CN104636492 B CN 104636492B CN 201510095450 A CN201510095450 A CN 201510095450A CN 104636492 B CN104636492 B CN 104636492B
- Authority
- CN
- China
- Prior art keywords
- data
- feature
- fusion
- fuzzy
- features
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 238000013145 classification model Methods 0.000 claims abstract description 7
- 238000013507 mapping Methods 0.000 claims abstract description 7
- 238000013500 data storage Methods 0.000 claims abstract description 5
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000012880 independent component analysis Methods 0.000 claims description 5
- 238000007499 fusion processing Methods 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 230000010354 integration Effects 0.000 abstract description 6
- 238000013523 data management Methods 0.000 abstract description 5
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000007788 roughening Methods 0.000 description 1
- 230000011218 segmentation Effects 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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/282—Hierarchical databases, e.g. IMS, LDAP data stores or Lotus Notes
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a dynamic data grading method based on fuzzy integral feature fusion, belonging to the technical field of computer storage and comprising the following steps: firstly, data feature extraction is carried out on training set data to form an initial data feature set, and data features are extracted according to data application and storage characteristics; fusing data characteristics; thirdly, reducing the fused data characteristics; generating a data grading model; data storage hierarchy mapping; the invention improves the accuracy of data classification, fully considers the characteristics of mutual correlation among data characteristics, utilizes fuzzy integration to perform characteristic fusion, establishes a more reasonable data classification model, is suitable for storage level decision of various dynamic data management, improves the processing speed of data classification and improves the storage efficiency.
Description
Technical field
The present invention discloses a kind of dynamic data stage division, belongs to computer memory technical field, specifically a kind of
Dynamic data stage division based on fuzzy integral Fusion Features.
Background technology
With big data, the arrival in cloud storage epoch, cloud data center is developed rapidly so that high-performance, it is low into
This intelligent data management turns into study hotspot.Because the application environment of complexity causes data to have ageing and spatiality, number
According to accessing and handling the features such as complexity, storage requirements for access diversity, so needing to be classified various dynamic datas, dividing
Layer processing, to realize the reasonable mapping between application demand and storage resource, improves the cost performance of storage device.For example, pass through
Data staging model splits data into hot spot data and cold data, and hot spot data is placed into the more excellent storage device of performance
On, access performance is lifted, the cold data often not accessed is placed on low-speed device, reduces carrying cost.
Dynamic data classification substantially belongs to classification problem, more using supervised classification method, i.e., according to point of pre-training
Level model is classified to data.So dynamic data hierarchy model is the key element of data staging/Bedding storage.It is existing
The classification judgment rule of data staging model is mostly the linear combination of each feature, but each access feature of dynamic data is mutually closed
Join, not par wise irrelevance, simple linear relationship can not carry out accurate description to these correlations, thus influence depositing for data
Storage and use.For this problem, the present invention proposes a kind of Intelligent Dynamic data staging side based on fuzzy integral Fusion Features
Method, to improve the accuracy of data staging, characteristic associated with each other between data characteristics is taken into full account, has been carried out using fuzzy integral
Fusion Features, more rational data staging model is established, suitable for the storage hierarchy decision-making of various dynamic data managements, improved
The processing speed of data staging, lift storage efficiency.
The content of the invention
The present invention is interrelated for each access feature of dynamic data, not par wise irrelevance, and simple linear relationship is not
The problem of these correlations can be carried out with accurate description, thus influence the storage and use of data, there is provided one kind is based on fuzzy
The dynamic data stage division of Fusion Features is integrated, realizes and carries out Fusion Features using fuzzy integral, is established more rational
Data staging model, suitable for the storage hierarchy decision-making of various dynamic data managements, the processing speed of data staging is improved, is lifted
Storage efficiency.
Concrete scheme proposed by the present invention is:
A kind of dynamic data stage division based on fuzzy integral Fusion Features, is concretely comprised the following steps:
1. carrying out data characteristics extraction to training set data, primary data characteristic set is formed, according to data application and is deposited
Store up feature extraction data characteristics;
2. data characteristics merges:The fuzzy mearue of each data characteristics combination is calculated, each feature is carried out using fuzzy integral
Fusion, obtain Fusion Features computation model and new data characteristics vector;
3. yojan is carried out to the data characteristics after fusion:Feature reduction is carried out to obtained new data characteristic vector, chosen
Optimal feature subset;
4. data staging model generates:Classification based training, generation data classification mould are carried out according to the optimal feature subset of selection
Type;Fusion Features computation model and optimal feature subset disaggregated model according to obtaining show that the optimum fusion of data to be fractionated is special
Sign vector;
5. data storage level maps:According to data classification model and the obtained optimum fusion feature of data to be fractionated to
Amount, judges data category, the mapping established between data and storage hierarchy to be sorted.
Described data characteristics extraction and application is artificial or machine is carried out, and is dropped primitive character with the method for mapping or conversion
Dimension, the less new feature of the dimension compared with primitive character is transformed to, forms primary data characteristic set.
Described data characteristics fusion, specific fusion process are:Calculate the fuzzy mearue of feature set to be fused:Using artificial
The method specified picks out the feature composition feature set to be fused that there may be correlation, reduces and calculates dimension;Calculate each spy
Levy the fuzzy integral of combination:According to obtained fuzzy mearue, integrated using fuzzy integral Choquet or Sugeno integral and calculating moulds
Paste integration;It is determined that new data characteristics vector:The too low combinations of features of fuzzy integral value is given up according to specified threshold, will be remaining
Combinations of features is combined as new characteristic vector together with reserved monomeric character independent of each other.
The described method that yojan is carried out to the data characteristics after fusion can use principal component analysis PCA methods, independent element
Analyze ICA methods, linear decision analysis LDA methods, Local Features Analysis LFA methods.
Usefulness of the present invention is:The present invention proposes a kind of Intelligent Dynamic data based on fuzzy integral Fusion Features
Stage division, to improve the accuracy of data staging, characteristic associated with each other between data characteristics is taken into full account, has utilized fuzzy product
Divide and carry out Fusion Features, establish more rational data staging model, determined suitable for the storage hierarchy of various dynamic data managements
Plan, the processing speed of data staging is improved, lift storage efficiency.
Brief description of the drawings
The schematic flow sheet of Fig. 1 present invention.
Embodiment
The present invention will be further described.
Fuzzy mearue is a dull and normalized set function, and the additive property in probability measure is replaced with condition by it
Weaker monotonicity, it can be regarded as the extension of probability measure.And fuzzy integral be just defined on the basis of fuzzy mearue one
Kind nonlinear function, has the ability of fusion multiple information, and conventional fuzzy integral has Choquet integrations and Sugeno to integrate.
A kind of dynamic data stage division based on fuzzy integral Fusion Features, is concretely comprised the following steps:
1. carrying out data characteristics extraction to training set data, primary data characteristic set is formed, according to data application and is deposited
Store up feature extraction data characteristics;Data access and storage feature can manually be extracted by expert according to correlation experience more.To carry
The availability and accuracy of high data characteristics, can be by the way of multidigit expert extracts jointly.On the other hand, depth network goes out
Now to automatically extract data characteristics.Therefore, the stage can also use depth network or other machines learning method
Data characteristics is automatically extracted, to improve the automaticity of data staging, or even accuracy.Due to meeting in subsequent treatment
The data characteristics extracted is further analysed and handled, so, the stage feature of extraction can be slightly detailed.
2. data characteristics merges:The fuzzy mearue of each data characteristics combination is calculated, each feature is carried out using fuzzy integral
Fusion, obtain Fusion Features computation model and new data characteristics vector;Described data characteristics fusion, specific fusion process
For:Calculate the fuzzy mearue of feature set to be fused:The feature that there may be correlation is picked out using the method being manually specified
Feature set to be fused is formed, reduces and calculates dimension;Calculate the fuzzy integral of each combinations of features:According to obtained fuzzy mearue, make
With fuzzy integral Choquet integrations or Sugeno integral and calculating fuzzy integrals;It is determined that new data characteristics vector:According to specified threshold
Value gives up the too low combinations of features of fuzzy integral value, by remaining combinations of features together with reserved monomeric character independent of each other
It is combined as new characteristic vector.
3. yojan is carried out to the data characteristics after fusion:Feature reduction is carried out to obtained new data characteristic vector, chosen
Optimal feature subset;Feature reduction method can use principal component analysis PCA methods, independent component analysis ICA methods, linear decision analysis
LDA methods, Local Features Analysis LFA methods.Also the machine learning methods such as rough set can be used.Because Choquet integrations inherently have
There is roughening, so being mainly the yojan to independent characteristic based on the new feature vector that Choquet integration fusions obtain.
4. data classification model generates:Classification based training, generation data classification mould are carried out according to the optimal feature subset of selection
Type;Fusion Features computation model and optimal feature subset disaggregated model according to obtaining show that the optimum fusion of data to be fractionated is special
Sign vector;Classification based training is carried out to training set based on optimal feature subset, generates data staging model.Classification based training model can be with
It is arbitrary classification model, including Supervised classification, such as decision tree, neutral net or unsupervised segmentation, such as cluster
Deng.
5. data storage level maps:It is vectorial according to data classification model and the fusion feature of obtained data to be fractionated,
Judge data category, the mapping established between data and storage hierarchy to be sorted.
Claims (4)
- A kind of 1. dynamic data stage division based on fuzzy integral Fusion Features, it is characterized in that concretely comprising the following steps:1. carrying out data characteristics extraction to training set data, primary data characteristic set is formed, it is special according to data application and storage Property extraction data characteristics;2. data characteristics merges:The fuzzy mearue of each data characteristics combination is calculated, each feature is merged using fuzzy integral, Obtain Fusion Features computation model and new data characteristics vector;3. yojan is carried out to the data characteristics after fusion:Feature reduction is carried out to obtained new data characteristic vector, chosen optimal Character subset;4. data staging model generates:Classification based training is carried out according to the optimal feature subset of selection, generates data classification model;Root According to obtained Fusion Features computation model and optimal feature subset disaggregated model draw the optimum fusion features of data to be fractionated to Amount;5. data storage level maps:According to data classification model and the obtained optimum fusion characteristic vector of data to be fractionated, Judge data category, the mapping established between data and storage hierarchy to be sorted.
- 2. a kind of dynamic data stage division based on fuzzy integral Fusion Features according to claim 1, it is characterized in that Described data characteristics extraction and application is artificial or machine is carried out, and with the method for mapping or conversion by primitive character dimensionality reduction, is transformed to The less new feature of dimension compared with primitive character, form primary data characteristic set.
- 3. a kind of dynamic data stage division based on fuzzy integral Fusion Features according to claim 1 or 2, its feature It is described data characteristics fusion, specific fusion process is:Calculate the fuzzy mearue of feature set to be fused:Using what is be manually specified Method picks out the feature composition feature set to be fused that there may be correlation, reduces and calculates dimension;Calculate each combinations of features Fuzzy integral:According to obtained fuzzy mearue, integrated using fuzzy integral Choquet or Sugeno integral and calculatings obscure product Point;It is determined that new data characteristics vector:The too low combinations of features of fuzzy integral value is given up according to specified threshold, by remaining feature Combination is combined as new characteristic vector together with reserved monomeric character independent of each other.
- 4. a kind of dynamic data stage division based on fuzzy integral Fusion Features according to claim 3, it is characterized in that The described method that yojan is carried out to the data characteristics after fusion can use principal component analysis PCA methods, independent component analysis ICA Method, linear decision analysis LDA methods, Local Features Analysis LFA methods.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510095450.8A CN104636492B (en) | 2015-03-04 | 2015-03-04 | Dynamic data grading method based on fuzzy integral feature fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510095450.8A CN104636492B (en) | 2015-03-04 | 2015-03-04 | Dynamic data grading method based on fuzzy integral feature fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104636492A CN104636492A (en) | 2015-05-20 |
CN104636492B true CN104636492B (en) | 2017-12-05 |
Family
ID=53215238
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510095450.8A Active CN104636492B (en) | 2015-03-04 | 2015-03-04 | Dynamic data grading method based on fuzzy integral feature fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104636492B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105303324B (en) * | 2015-11-10 | 2019-11-19 | 中国建设银行股份有限公司 | A kind of information system parameter management method and device |
CN107426315B (en) * | 2017-07-24 | 2020-07-31 | 南京邮电大学 | Distributed cache system Memcached improvement method based on BP neural network |
CN109214514A (en) * | 2018-08-14 | 2019-01-15 | 浪潮通用软件有限公司 | A kind of data analysing method based on Rough Set |
CN111166294B (en) * | 2020-01-29 | 2021-09-14 | 北京交通大学 | Automatic sleep apnea detection method and device based on inter-heartbeat period |
CN117909507B (en) * | 2024-03-19 | 2024-05-17 | 金盾检测技术股份有限公司 | AI-based data classification system |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101872424A (en) * | 2010-07-01 | 2010-10-27 | 重庆大学 | Facial expression recognizing method based on Gabor transform optimal channel blur fusion |
-
2015
- 2015-03-04 CN CN201510095450.8A patent/CN104636492B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101872424A (en) * | 2010-07-01 | 2010-10-27 | 重庆大学 | Facial expression recognizing method based on Gabor transform optimal channel blur fusion |
Non-Patent Citations (2)
Title |
---|
《基于多分类器多模糊积分的信息融合方法》;段宝彬,孙梅兰;《 重庆科技学院学报(自然科学版)》;20080630;第10卷(第3期);第87-89页 * |
《基于模糊积分的多分类器融合方法研究》;赵志伟;《中国优秀硕士学位论文全文数据库》;20091215;第三章 * |
Also Published As
Publication number | Publication date |
---|---|
CN104636492A (en) | 2015-05-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104636492B (en) | Dynamic data grading method based on fuzzy integral feature fusion | |
CN106372648A (en) | Multi-feature-fusion-convolutional-neural-network-based plankton image classification method | |
CN106503148B (en) | A kind of table entity link method based on multiple knowledge base | |
CN106909643A (en) | The social media big data motif discovery method of knowledge based collection of illustrative plates | |
CN106709754A (en) | Power user grouping method based on text mining | |
CN104090882B (en) | A kind of quick clustering method of advertisement order and system, server | |
CN103902988B (en) | A kind of sketch shape matching method based on Modular products figure with Clique | |
CN110322453A (en) | 3D point cloud semantic segmentation method based on position attention and auxiliary network | |
CN103186538A (en) | Image classification method, image classification device, image retrieval method and image retrieval device | |
CN104036255A (en) | Facial expression recognition method | |
CN105654196A (en) | Adaptive load prediction selection method based on electric power big data | |
CN107562947A (en) | A kind of Mobile Space-time perceives the lower dynamic method for establishing model of recommendation service immediately | |
CN103942606A (en) | Residential electricity consumption customer segmentation method based on fruit fly intelligent optimization algorithm | |
CN105678590A (en) | topN recommendation method for social network based on cloud model | |
CN109657063A (en) | A kind of processing method and storage medium of magnanimity environment-protection artificial reported event data | |
CN112528639B (en) | Object recognition method and device, storage medium and electronic equipment | |
CN104915354A (en) | Multimedia file pushing method and device | |
CN101398846A (en) | Image, semantic and concept detection method based on partial color space characteristic | |
CN105808665A (en) | Novel hand-drawn sketch based image retrieval method | |
CN104216993A (en) | Tag-co-occurred tag clustering method | |
CN109783805A (en) | A kind of network community user recognition methods and device | |
CN104008177A (en) | Method and system for rule base structure optimization and generation facing image semantic annotation | |
CN104765852B (en) | Data digging method based on fuzzy algorithmic approach under big data background | |
CN104361135A (en) | Image retrieval method | |
CN103440651A (en) | Multi-label image annotation result fusion method based on rank minimization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |