CN107247576A - A kind of multichannel data piecemeal floating-point quantification treatment framework - Google Patents
A kind of multichannel data piecemeal floating-point quantification treatment framework Download PDFInfo
- Publication number
- CN107247576A CN107247576A CN201710416191.3A CN201710416191A CN107247576A CN 107247576 A CN107247576 A CN 107247576A CN 201710416191 A CN201710416191 A CN 201710416191A CN 107247576 A CN107247576 A CN 107247576A
- Authority
- CN
- China
- Prior art keywords
- data
- floating
- quantification treatment
- piecemeal
- framework
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
- G06F7/38—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
- G06F7/48—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
- G06F7/483—Computations with numbers represented by a non-linear combination of denominational numbers, e.g. rational numbers, logarithmic number system or floating-point numbers
Abstract
The present invention discloses a kind of multichannel data piecemeal floating-point quantification treatment framework.The feature of framework is that structural data, semi-structured data and unstructured data are arranged in into three road arrays；Using Higher-order Singular value decomposition, three road arrays are decomposed into the matrix pattern of second-order tensor；Matrix pattern is transformed into sparse domain again, piecemeal floating-point quantification treatment is carried out.
Description
Technical field
The present invention relates to a kind of multichannel data piecemeal floating-point quantification treatment framework
Background technology
With cloud computing and virtualization, the features such as " extensive ", " high density ", " high energy consumption ", " complication " is showed, is built
If with development New Generation of IDC, lifting data center infrastructure management will become increasingly important, the basis of data center
The new trend that framework fusion management will develop with intelligence as data center.Ultra-large type data center provide from infrastructure to
Data analysis, screening, the whole application service of application below.Be not only data analysis, in addition to public cloud provide it is logical
With changing the different special services of service in the cloud computing of intelligence manufacture, and super computing, this disposal ability just to big data
Propose requirements at the higher level.
It can be represented with data or unified structure, we term it structural data, such as numeral, symbol.Tradition
Relational data model, row data, be stored in database, bivariate table representation can be used.Semi-structured data, is exactly between complete
Full structural data (data in such as relevant database, object-oriented database) and complete structureless data (such as sound,
Image file etc.) between data, XML, html document just belong to semi-structured data.It is usually self-described, data
Structure and content mix, and do not distinguish significantly.Unstructured database refers to that its field length is variable, and each word
The record of section again can be by the database that repeats or not reproducible subfield is constituted, not only can be with processing structure number with it
According to (such as numeral, symbol information) and it is more suitable for processing unstructured data (full text text, image, sound, video display, super matchmaker
The information such as body).
Arrangement of the data along an equidirectional is referred to as array all the way.Scalar is the expression of zero road array, row vector and arrange to
Amount is the array all the way that data edge is both horizontally and vertically arranged respectively, and matrix is that data are arranged along horizontal and vertical directions
Two road arrays.Tensor is that the multichannel array of data is represented, it is a kind of extension of matrix.The most frequently used tensor is three ranks
Amount.Three rank tensors are also referred to as three-dimensional matrice.Three rank tensors of dimension identical pros are referred to as cube.
Three road arrays of three rank tensors are not matched with row vector, column vector etc., and rename as tensor fiber.Fiber is only to retain
One subscript is variable, fix other all subscripts it is constant obtained from array all the way.They are the level fibre of three rank tensors respectively
Dimension, vertical fiber and depth fiber.High order tensor can also use the set expression of matrix.These matrixes form three rank tensors
Dropping cut slice, lateral section and front section., can be by a three rank tensors (three road arrays) in the analysis and calculating of tensor
By reorganizing or arranging, become a matrix (two road arrays).
Matrix has two associated vector spaces：Column space and row space.Singular value decomposition by the two vector spaces just
Friendshipization, and by product that matrix decomposition is three matrixes：Left singular matrix, right singular matrix and middle diagonal singular value matrix.
Because singular value must act on often more important than left and right singular vector, so singular value matrix can be considered the core square of matrix
Battle array.If regarding diagonal singular value matrix as a second-order tensor, singular value matrix is naturally enough the core of second-order tensor
Amount, and three matrix products of matrix can be changed to the n- patterns product of second-order tensor.
Sparse signal refers to that the value in most of sampling instants is equal to zero or is approximately equal to zero, only a small amount of samples moment
Value be substantially not equal to zero signal.Many natural signs are not sparse signal in time domain, but are in some transform domain
Sparse.These transformation tools include Fourier conversion, Instant Fourier Transform, wavelet transformation and Gabor transformation etc..
It is that data are divided into data in group, group bi-directional scaling relative to each other using block floating point algorithm, but can not be with
The member of other groups scales in the same proportion, even if the so simple mathematical operation of such as multiplication.In more complicated matrix
More complicated mathematical operation is needed in situation of inverting, between packet, block floating point processor must be just used.
Piecemeal floating-point quantization algorithm is based on the fact that in a small time interval entropy of data will be less than whole number
According to the entropy of collection.Piecemeal floating-point quantizer is the output stream of a reception analog-digital converter, and by sampled data unified quantization
For a kind of equipment of effective representation of initial data, only require that bit number is less than sample number in quantizing process.
The invention provides a kind of multichannel data piecemeal floating-point quantification treatment framework.The feature of framework is, by structuring number
Three road arrays are arranged according to, semi-structured data and unstructured data；Using Higher-order Singular value decomposition, three road arrays are decomposed
For the matrix pattern of second-order tensor；Matrix pattern is transformed into sparse domain again, piecemeal floating-point quantification treatment is carried out.
The content of the invention
It is an object of the invention to provide a kind of multichannel data piecemeal floating-point quantification treatment framework.The present invention includes following spy
Levy：
Inventive technique scheme
1. a kind of multichannel data piecemeal floating-point quantification treatment framework, the feature of framework：
1) structural data, semi-structured data and unstructured data are arranged in three road arrays；
2) Higher-order Singular value decomposition is used, three road arrays are decomposed into the matrix pattern of second-order tensor；
3) matrix pattern is transformed into sparse domain again, carries out piecemeal floating-point quantification treatment.
Brief description of the drawings
Accompanying drawing 1 is multichannel data piecemeal floating-point quantification treatment Organization Chart.
Embodiment
This multichannel data piecemeal floating-point quantification treatment framework, comprises the following steps feature：
1) structural data, semi-structured data and unstructured data are arranged in three road arrays；
2) Higher-order Singular value decomposition is used, three road arrays are decomposed into the matrix pattern of second-order tensor；
3) matrix pattern is transformed into sparse domain again, carries out piecemeal floating-point quantification treatment.
Claims (1)
1. a kind of multichannel data piecemeal floating-point quantification treatment framework, the feature of framework：
1) structural data, semi-structured data and unstructured data are arranged in three road arrays；
2) Higher-order Singular value decomposition is used, three road arrays are decomposed into the matrix pattern of second-order tensor；
3) matrix pattern is transformed into sparse domain again, carries out piecemeal floating-point quantification treatment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710416191.3A CN107247576A (en) | 2017-06-06 | 2017-06-06 | A kind of multichannel data piecemeal floating-point quantification treatment framework |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710416191.3A CN107247576A (en) | 2017-06-06 | 2017-06-06 | A kind of multichannel data piecemeal floating-point quantification treatment framework |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107247576A true CN107247576A (en) | 2017-10-13 |
Family
ID=60017670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710416191.3A Pending CN107247576A (en) | 2017-06-06 | 2017-06-06 | A kind of multichannel data piecemeal floating-point quantification treatment framework |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107247576A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104063897A (en) * | 2014-06-28 | 2014-09-24 | 南京理工大学 | Satellite hyper-spectral image compressed sensing reconstruction method based on image sparse regularization |
| CN105531725A (en) * | 2013-06-28 | 2016-04-27 | D-波系统公司 | Systems and methods for quantum processing of data |
| CN105721869A (en) * | 2016-01-26 | 2016-06-29 | 上海交通大学 | Structured sparsity-based compression tensor acquisition and reconstruction system |
| CN106503659A (en) * | 2016-10-24 | 2017-03-15 | 天津大学 | Action identification method based on sparse coding tensor resolution |
| CN106529435A (en) * | 2016-10-24 | 2017-03-22 | 天津大学 | Action recognition method based on sensor quantization |
-
2017
- 2017-06-06 CN CN201710416191.3A patent/CN107247576A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105531725A (en) * | 2013-06-28 | 2016-04-27 | D-波系统公司 | Systems and methods for quantum processing of data |
| CN104063897A (en) * | 2014-06-28 | 2014-09-24 | 南京理工大学 | Satellite hyper-spectral image compressed sensing reconstruction method based on image sparse regularization |
| CN105721869A (en) * | 2016-01-26 | 2016-06-29 | 上海交通大学 | Structured sparsity-based compression tensor acquisition and reconstruction system |
| CN106503659A (en) * | 2016-10-24 | 2017-03-15 | 天津大学 | Action identification method based on sparse coding tensor resolution |
| CN106529435A (en) * | 2016-10-24 | 2017-03-22 | 天津大学 | Action recognition method based on sensor quantization |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107247575A (en) | A kind of multichannel data floating point processor prototype | |
| Bertschinger | Simulations of structure formation in the universe | |
| Graovac et al. | A new version of atom-bond connectivity index | |
| Ergul et al. | A hierarchical partitioning strategy for an efficient parallelization of the multilevel fast multipole algorithm | |
| Kepner et al. | Sparse deep neural network graph challenge | |
| Nasrullah et al. | Joint image compression and encryption using IWT with SPIHT, Kd-tree and chaotic maps | |
| CN109785847A (en) | Audio compression algorithm based on dynamic residual network | |
| CN106295690A (en) | Time series data clustering method based on Non-negative Matrix Factorization and system | |
| Giordano et al. | ROI-based on-board compression for hyperspectral remote sensing images on GPU | |
| CN107247576A (en) | A kind of multichannel data piecemeal floating-point quantification treatment framework | |
| CN107247688A (en) | A kind of multichannel data compressed sensing cube framework | |
| Balasubramanian | Nonrigid water octamer: Computations with the 8‐cube | |
| CN107180009A (en) | A kind of multichannel data compressed sensing cube body processor prototype | |
| Wang et al. | 3D-shufflevit: an efficient video action recognition network with deep integration of self-attention and convolution | |
| Mahmud et al. | Dimension reduction using quantum wavelet transform on a high-performance reconfigurable computer | |
| CN107450851A (en) | A kind of multichannel data storage and processing system framework based on space-time adaptive processing | |
| CN107480209A (en) | A kind of multichannel data storage and processing system prototype based on space-time adaptive processing | |
| Mansoori et al. | High level design of a flexible pca hardware accelerator using a new block-streaming method | |
| Jiang et al. | A review on fast direct methods of surface integral equations for analysis of electromagnetic scattering from 3-D PEC objects | |
| Zhang | On numerical orientation averaging with spherical fibonacci point sets and compressive scheme | |
| Yuen et al. | Remote data analysis, visualization and problem solving environment (PSE) based on wavelets in the geosciences | |
| Munjiza et al. | Parallel, distributed and grid computing in discontinua analysis | |
| Belpassi et al. | Parallelization of a relativistic DFT code | |
| CN110334316A (en) | A kind of multichannel data piecemeal floating-point quantification treatment device prototype | |
| Zaman et al. | Utmost travelling wave phenomena to the fractional type nonlinear evolution equation in mathematical physics |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171013 |