CN115314345A - Data processing method, data processing apparatus, electronic device, storage medium, and program product - Google Patents
Data processing method, data processing apparatus, electronic device, storage medium, and program product Download PDFInfo
- Publication number
- CN115314345A CN115314345A CN202210945441.3A CN202210945441A CN115314345A CN 115314345 A CN115314345 A CN 115314345A CN 202210945441 A CN202210945441 A CN 202210945441A CN 115314345 A CN115314345 A CN 115314345A
- Authority
- CN
- China
- Prior art keywords
- matrix
- tendency
- matrices
- matrixes
- merging
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/024—Channel estimation channel estimation algorithms
- H04L25/0242—Channel estimation channel estimation algorithms using matrix methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/021—Estimation of channel covariance
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Complex Calculations (AREA)
Abstract
The application provides a data processing method, a data processing device, electronic equipment, a storage medium and a program product, and relates to the technical field of computers. When the method is used for matrix combination, a combination coefficient is determined by utilizing the correlation between a tendency matrix which has larger positive influence on the accuracy of a data processing result and another secondary tendency matrix, and the combination coefficient is used for combination, wherein the combination coefficient can be calculated in a self-adaptive manner according to the tendency in combination, so that the combined result can be more inclined to the tendency matrix.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a data processing method, an apparatus, an electronic device, a storage medium, and a program product.
Background
In a modern communication system, such as Long Term Evolution (LTE), new Radio (NR), wireless Fidelity (WiFi), and the like, there are measurements performed according to different time-frequency resources, signals, and the like, and finally, measurement results of different signal types and different signal positions need to be integrated, and after the final measurement result is obtained, the final measurement result is provided for a post-stage process.
Taking channel measurement in an LTE/NR communication system as an example, the measurement of a channel includes measurement of time-frequency response of channels of different transmit-receive antenna ports, measurement of interference amplitude, and the like, at this time, different covariance matrices are obtained by measurement, such as a statistical covariance matrix of interference noise, a channel covariance matrix representing correlation of channel antennas/data streams, and the like.
The way of integrating the multiple matrices at present is to perform weighted summation on the multiple matrices, and this way depends on reasonable setting of weights, and if the weights are improperly set, the final combining result is greatly affected, for example, when some of the matrices have higher reliability but have smaller weights, the combining result obtained in this way is affected by the rest of the unreliable matrices, and when some of the matrices have larger deviation and have larger weights, this way also has a larger negative effect on the final combining result, thereby causing an effect on the subsequent processing result.
Disclosure of Invention
An embodiment of the present application aims to provide a data processing method, an apparatus, an electronic device, a storage medium, and a program product, so as to solve a problem that an existing method employs a sum-average method to merge multiple matrices, which is unreasonable, and results in a large error influence on a final merging result.
In a first aspect, an embodiment of the present application provides a data processing method, where the method includes:
determining a combination coefficient corresponding to a secondary tendency matrix and a tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix, wherein the tendency matrix refers to one of the two matrices which has a larger positive influence on the accuracy of a data processing result, and the secondary tendency matrix refers to the other one of the two matrices except the tendency matrix;
and combining the two matrixes according to the combination coefficient corresponding to the secondary tendency matrix and the tendency matrix.
In the implementation process, when matrix combination is performed, the method determines a combination coefficient by utilizing the correlation between a tendency matrix which has a larger positive influence on the accuracy of a data processing result and another tendency matrix, and performs combination by utilizing the combination coefficient, wherein the combination coefficient can be adaptively calculated according to the tendency in combination, so that the combined result can be more inclined to the tendency matrix, and the tendency matrix can refer to a matrix with smaller deviation or higher reliability, so that the deviation of the final combination result is smaller, the reliability is higher, namely the accuracy is higher.
Optionally, the determining, by using a correlation between a secondary tendency matrix and a tendency matrix, a combining coefficient of the secondary tendency matrix corresponding to the tendency matrix includes:
calculating the norm of the tendency matrix, and acquiring the inner product of the tendency matrix and the secondary tendency matrix under the definition of the norm;
and calculating to obtain a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the norm of the tendency matrix and the inner product.
In the implementation process, the correlation between the two matrixes can be better obtained by obtaining the norm and the inner product, so that a more reasonable combination coefficient can be determined according to the correlation.
Optionally, the determining, by using a correlation between a secondary tendency matrix and a tendency matrix, a combining coefficient of the secondary tendency matrix corresponding to the tendency matrix includes:
performing characteristic decomposition on the tendency matrix to obtain a characteristic vector and a characteristic value;
projecting the secondary tendency matrix to the characteristic direction of the tendency matrix according to the characteristic vector to obtain a projection matrix;
and calculating and obtaining a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the characteristic value and the projection matrix.
In the implementation process, the correlation size between the two matrixes can be known more accurately by acquiring the projection matrix, and then a more reasonable merging coefficient can be determined according to the correlation.
Optionally, the two matrices are preprocessed by:
acquiring two initial matrixes participating in combination calculation;
if the dimensionalities of the two initial matrixes are different, determining the maximum dimensionality of the two initial matrixes, and performing element expansion on the matrix with the dimensionality not reaching the maximum dimensionality to enable the dimensionality of the matrix to reach the maximum dimensionality to obtain the two matrixes after the element expansion, and/or if the two matrixes belong to different calculation domains, converting the matrix with the real number domain as the calculation domain into a complex number domain to obtain the two matrixes of the complex number domain.
In the implementation process, the dimensionality or the calculation domain of the matrix is transformed and then merged, so that the merging method can be suitable for more matrix merging scenes.
Optionally, the tendency of each matrix is determined based on the number of statistical samples when each matrix is obtained, or when each matrix is a statistical covariance matrix of interference noise obtained by the device based on the measurement result of the interference signal on the CSI-RS, the tendency of each matrix is determined based on the signal-to-noise ratio when each matrix is obtained. The tendency of the matrix is determined in this way, so that the reliability of the matrix can be determined according to the actual situation, and the final combination result can be more inclined to the matrix with high reliability so as to obtain more accurate combination result.
In the implementation process, the dimensionality or the calculation domain of the matrix is transformed and then combined, so that the combining method can be suitable for more matrix combining scenes.
Optionally, the two matrices are matrices characterizing signal-related information or matrices for signal processing.
Optionally, the two matrices are any one of the following matrices:
a statistical covariance matrix of interference noise obtained based on interference signal measurement results on the CSI-RS;
a channel covariance matrix characterizing channel antenna or data stream correlations;
a channel response matrix obtained by channel estimation;
a filter matrix for parametric or signal noise reduction.
In the implementation process, the merging method can be applied to the matrix merging scheme under various scenes.
Optionally, when each of the two matrices is a statistical covariance matrix of interference noise, after obtaining the combined matrix, the method further includes:
and acquiring relevant information of the interference noise according to the merging matrix, wherein the relevant information comprises amplitude and/or direction.
In a second aspect, an embodiment of the present application further provides a data processing method, where the method includes:
when N matrixes are combined, the N matrixes are sorted from large to small according to the tendency to obtain the sorted N matrixes, wherein the tendency represents the positive influence on the accuracy of a data processing result, and N is an integer greater than 2;
determining the ith matrix in the N sequenced matrixes as the secondary tendency matrix, sequentially taking values of i from M sequences with tolerance of 1, wherein M is a positive integer, the initial value of the 1 st sequence is 2, the ending value of the M sequence is N, the ending value of the jth sequence is in a value range of [2,N ], the initial value of the j +1 th sequence is in a value range of [1,N-1], the initial value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j is sequentially taken from 1 to M;
and when the value of i is 2 for the first time, determining the 1 st matrix in the N sequenced matrixes as the tendency matrix, otherwise, determining a combined matrix obtained by combining the k matrix for the last time as the tendency matrix, and combining the tendency matrix and the secondary tendency matrix according to the method provided by the first aspect to obtain the combined matrix, wherein k is the last value of i.
In the implementation process, if there is only one sequence, the matrixes are combined in sequence according to the tendency, so that the final combination result is more prone to the matrix with the large tendency, that is, the final combination result contains more direction information of the matrix with the large tendency, and thus, a more accurate combination result can be obtained, and the problem that the other matrixes with large deviation have large negative influence on the combination result is solved. If there are multiple sequences, after merging to a certain matrix in the middle, returning to continue iterative merging, so that the final merged result is more biased toward a certain matrix through multiple times of merging, for example, more biased toward a matrix with a large tendency, so that the final merged result includes more information about the matrix with the large tendency.
In a third aspect, an embodiment of the present application provides a data processing apparatus, where the apparatus includes:
the merging coefficient determining module is used for determining merging coefficients corresponding to a secondary tendency matrix and a tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix, wherein the tendency matrix is a matrix which has a larger positive influence on the accuracy of a data processing result in the two matrices, and the secondary tendency matrix is the other matrix except the tendency matrix in the two matrices;
and the matrix merging module is used for merging the two matrixes according to the secondary tendency matrix and the merging coefficient corresponding to the tendency matrix.
In a fourth aspect, an embodiment of the present application further provides a data processing apparatus, where the apparatus includes: the sorting module is used for sorting the N matrixes from large to small according to the tendency when the N matrixes are combined to obtain the N sorted matrixes, wherein the tendency represents the positive influence on the accuracy of a data processing result, and N is an integer greater than 2;
a value taking module, configured to determine an ith matrix in the N sequenced matrices as the secondary tendency matrix, and sequentially take values of i from M sequences with a tolerance of 1, where M is a positive integer, where a starting value of the 1 st sequence is 2, an ending value of the M th sequence is N, a value range of the ending value of the jth sequence is [2,N ], a value range of a starting value of the j +1 th sequence is [1,N-1], a starting value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j sequentially takes values from 1 to M;
and a merging module, configured to determine a 1 st matrix of the N sequenced matrices as the trend matrix when the value of i is 2 for the first time, otherwise, determine a merged matrix obtained by merging the k-th matrix for the last time as the trend matrix, and merge the trend matrix and the secondary trend matrix according to the method provided in the first aspect to obtain a merged matrix, where k is a last value of i.
In a fifth aspect, an embodiment of the present application provides an electronic device, including a processor and a memory, where the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the method as provided in the first aspect are executed.
In a sixth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs the steps in the method as provided in the first aspect.
In a seventh aspect, an embodiment of the present application provides a computer program product, which includes computer program instructions, and when the computer program instructions are read and executed by a processor, the steps in the method provided in the first aspect are performed.
Additional features and advantages of the present application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the present application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and that those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.
Fig. 1 is a flowchart of a data processing method according to an embodiment of the present application;
fig. 2 is a flowchart of another data processing method according to an embodiment of the present application;
fig. 3 is a block diagram of a data processing apparatus according to an embodiment of the present application;
fig. 4 is a block diagram of another data processing apparatus according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of an electronic device for executing a data processing method according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.
It should be noted that the terms "system" and "network" in the embodiments of the present invention may be used interchangeably. The "plurality" means two or more, and in view of this, the "plurality" may also be understood as "at least two" in the embodiments of the present invention. "and/or" describes the association relationship of the associated objects, meaning that there may be three relationships, e.g., a and/or B, which may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" generally indicates that the preceding and following related objects are in an "or" relationship, unless otherwise specified.
The embodiment of the application provides a data processing method, when matrix combination is carried out, a combination coefficient is determined by utilizing the correlation between a tendency matrix which has a larger positive influence on the accuracy of a data processing result and another matrix, combination is carried out by utilizing the combination coefficient, the combination coefficient can be calculated in a self-adaptive mode according to the tendency in combination, so that the combined result can tend to the tendency matrix more, compared with a mode of carrying out matrix combination according to set weight, the scheme can realize self-adaptive weight adjustment in the matrix combination process, the combined matrix is more accurate, and the influence on subsequent calculation is smaller.
Referring to fig. 1, fig. 1 is a flowchart of a data processing method according to an embodiment of the present application, where the method includes the following steps:
step S110: and determining the merging coefficient corresponding to the secondary tendency matrix and the tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix.
In the embodiment of the present application, when two matrices are merged, a merging coefficient of each matrix needs to be determined first. The two matrices may refer to any two matrices that need to be combined, and in some application scenarios, each matrix may be a matrix representing signal related Information or a matrix used for signal processing, for example, in a communication application scenario, the matrix of the signal related Information may be, for example, various statistical covariance matrices, channel State Information (CSI) measurement Information matrices, channel response matrices, and the like, and the matrix used for signal processing may be, for example, an interference noise covariance matrix, a Channel correlation matrix, a statistical direction matrix, a filter matrix, and the like. It can be understood that in different application scenarios, different matrices may need to be merged and then subsequently calculated, so the method of the present application may also be applicable to other application scenarios where there are matrices that need to be merged, and the matrices that participate in merging may also be various types of matrices.
In other words, the types of the two matrices are different in different application scenarios, for example, the two matrices may be any one of the following matrices:
a statistical covariance matrix of interference noise obtained based on an interference Signal measurement result on a Channel State Information-Reference Signal (CSI-RS);
a channel covariance matrix characterizing channel fill or data stream correlation;
a channel response matrix obtained by channel estimation;
the filter matrix is used for parameter or signal noise reduction, where the parameter may refer to parameters such as amplitude-frequency response, frequency offset, time offset, and the like of a channel.
When the two matrixes are combined, a tendency matrix and a secondary tendency matrix in the two matrixes can be determined firstly, wherein the tendency matrix refers to a matrix which has a larger positive influence on the accuracy of the data processing result in the two matrixes, namely the accuracy of the tendency matrix is larger, and the secondary tendency matrix refers to another matrix except the tendency matrix in the two matrixes, so that when a combination coefficient is determined subsequently, a larger weight can be distributed to the tendency matrix, the final combined result can also contain more information of the tendency matrix, namely the final combined result can be more accurate.
For example, two matrices include a matrix 1 and a matrix 2, and in the process of combining the matrix 1 and the matrix 2 to obtain the matrix 12, a trend matrix may be determined from the matrix 1 and the matrix 2, where the trend matrix may be understood as a matrix in which direction information of the matrix 12 obtained after combining is more prone to the trend matrix, that is, a matrix having a greater positive influence on accuracy of the matrix 12, if the trend matrix is the matrix 1, the secondary trend matrix is the matrix 2, the obtained matrix 12 includes more direction information of the matrix 1, and if the matrix 1 includes more real information, the combining result obtained after combining may also include more real information.
Step S120: and combining the two matrixes according to the combination coefficient of the secondary tendency matrix and the corresponding tendency matrix.
The merge coefficient may be understood as a weight for merging two matrices, for example, when matrix 1 and matrix 2 are merged, a merge coefficient of each of the two matrices may be calculated according to a correlation between the two matrices, for example, if matrix 1 corresponds to merge coefficient a, and matrix 2 corresponds to merge coefficient b, merge matrix 12= matrix 1*a + matrix 2*b.
In the scheme, the tendency matrix is taken as a reference, the correlation between the two matrixes is utilized, and then the merging coefficient is determined, so that the merging matrix obtained after merging contains more information of the tendency matrix, such as direction information, if the reliability of the tendency matrix is higher, and if the reliability of the secondary tendency matrix is lower, the reliability of the obtained merging matrix is also higher, and the merging matrix is not greatly influenced by the secondary tendency matrix with lower reliability.
In the implementation process, when matrix combination is performed, the method determines a combination coefficient by utilizing the correlation between a tendency matrix which has a larger positive influence on the accuracy of a data processing result and another tendency matrix, and performs combination by utilizing the combination coefficient, wherein the combination coefficient can be calculated in a self-adaptive manner according to the tendency in combination, so that the combined result can be more inclined to the tendency matrix, and the tendency matrix can refer to a matrix with smaller deviation or higher reliability, so that the deviation of the final combination result is smaller and the reliability is higher.
On the basis of the above embodiment, the correlation between the two matrices may be characterized by a norm inner product, so in the determining the combining coefficient according to the correlation, the method may include: and calculating the norm of the tendency matrix, acquiring the inner product between the tendency matrix and the secondary tendency matrix under the definition of the norm, and calculating and acquiring the merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the norm and the inner product of the tendency matrix.
For example, two matrices R 1 And R 2 If the tendency matrix is R 1 The sub-inclination matrix is R 2 Calculating the matrix R 1 Norm of (R) 1 The Frobenius norm of the matrix can be chosen here, which is calculated as:
wherein the content of the first and second substances,
represents R 1 The conjugate transpose of (c), trace () represents the Trace of a matrix.
Then calculating to obtain a tendency matrix R under the definition of the norm 1 And the sub-trend matrix R 2 Inner product R between 1 ,R 2 >R is defined as the Frobenius norm 1 And R 2 Inner product of<R 1 ,R 2 >Is defined as follows:
then, according to the result of the norm and inner product of the tendency matrix, a corresponding projection coefficient can be calculated, wherein the projection coefficient can be understood as a merging coefficient, and the calculation formula is as follows:
after the merging coefficient is obtained by calculation, a calculation formula for merging the two matrixes is as follows:
R 12 ＝c 1 R 1 +c 2 R 2 。
it should be understood that the expression form of the norm and the inner product thereof can be flexibly adjusted according to actual situations, and is not limited to the expression form.
In the implementation process, the correlation between the two matrixes can be better obtained by obtaining the norm and the inner product, so that a more reasonable combination coefficient can be determined according to the correlation.
In other embodiments, the manner of determining the merging coefficient according to the correlation between the two matrices may also be obtained by performing eigen decomposition on the matrices, and the specific implementation manner includes: and calculating and obtaining a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the characteristic value and the projection matrix.
Taking two matrices as covariance matrices, i.e. matrix R 1 And matrix R 2 Is a covariance matrix of two complex forms, if the trend matrix is the matrix R 1 The sub-inclination matrix is R 2 First, R is added 1 And (3) decomposing the characteristic value in a way shown by the following calculation formula:
wherein, U 1 Represents R 1 Of feature vectors, i.e. U 1 Can be understood as a feature vector, Σ 1 Represents U 1 The feature vector in (2) is corresponding to a diagonal matrix formed by non-negative eigenvalues, namely eigenvalues.
Will matrix R 1 After performing the feature decomposition, the matrix R is 2 Directional matrix R 1 And (3) projecting in the characteristic direction to obtain a projection matrix Q:
therefore, the matrix R 1 And matrix R 2 The correlation between can be based on R 1 The eigenvalues obtained after the eigen decomposition are measured with the norm of the column vector in the projection matrix Q, and the merging coefficient can be obtained specifically by the following formula:
therein, sigma 1,i,i Represents sigma 1 Row ith and column element of (2), Q i,i I row and i column element, Σ, representing Q k Represents a diagonal matrix sigma 1 The sum of the median diagonal elements.
The calculation formula for combining the two matrices is as follows:
R 12 ＝c 1 R 1 +c 2 R 2 。
the method for obtaining the combination coefficient by performing eigenvalue decomposition by the method can also be popularized to the combination of any matrix, not necessarily the covariance matrix, and any matrix R with the same dimension 1 And R 2 The same applies, if the matrix R 1 To trend the matrix, the matrix R can be aligned first 1 Singular value decomposition is performed, such as:
R 1 ＝U 1 ∑ 1 V 1 H ；
wherein, U 1 And V 1 Are each R 1 Of the left and right eigenvectors, which may also be referred to as eigenvectors, Σ 1 Is U 1 And V 1 The feature vector in (2) is corresponding to a diagonal matrix formed by non-0 singular values, namely, the feature value.
The matrix R may then be applied 2 To U 1 And V 1 To obtain a projection matrix:
the merging coefficients are then calculated:
therein, sigma 1,i,i Represents sigma 1 The ith line ofi columns of elements, Q i,i I row and i column element, Σ, representing Q k Represents the diagonal matrix sigma 1 The sum of the median diagonal elements.
The calculation formula for combining the two matrices is as follows:
R 12 ＝c 1 R 1 +c 2 R 2 。
in the implementation process, the correlation size between the two matrixes can be known more accurately by acquiring the projection matrix, and then a more reasonable combination coefficient can be determined according to the correlation.
On the basis of the foregoing embodiment, the two matrices to be combined may be two matrices in N matrices, or may be an intermediate combined matrix obtained by initially combining two matrices in N matrices and another matrix, for example, N is an integer greater than 2, where the N matrices include a matrix 1, a matrix 2, and a matrix 3, the two matrices in the foregoing embodiment may refer to any two matrices in the three matrices that need to be combined, or, for example, the matrix 1 and the matrix 2 may be combined first, and then the matrix 12 is combined, and then the matrix 12 and the matrix 3 are combined, and a final combined matrix is obtained, in which case, the two matrices in the foregoing embodiment may be understood as the matrix 12 and the matrix 3. That is, the two matrices in the above embodiments may not only be understood as including the matrix that needs to be merged initially, but also include the intermediate merged matrix obtained in the merging process.
When N matrices are merged, in order to obtain a better merging effect, merging may also be performed in a certain order, and a specific implementation manner includes:
mode (1): the N matrixes are sequenced from large to small according to the tendency to obtain the sequenced N matrixes, wherein the tendency can be understood as the size of positive influence on the accuracy of a data processing result, then the ith matrix in the sequenced N matrixes is determined as a secondary tendency matrix, i is sequentially valued from 2 to N, when i =2, the 1 st matrix in the sequenced N matrixes is determined as a tendency matrix, the tendency matrix and the secondary tendency matrix are combined according to the method to obtain a combined matrix, when i =2, the combined matrix obtained by combining the (i-1) matrixes is determined as a tendency matrix, and the tendency matrix and the secondary tendency matrix are combined according to the method to obtain the combined matrix.
For example, when the value of i is greater than 2, the merged matrix obtained by merging the first i-1 matrices and the ith matrix may be merged, and when the merged matrix is merged to i = N, the final merged matrix is obtained.
For example, for N matrices, the magnitude of the tendency of each matrix may be determined first, and then the matrices are sorted according to the order of the tendency from large to small or from small to large, for example, the N matrices include 4 matrices, which are sorted according to the order of the tendency from large to small into matrix 1-matrix 2-matrix 3-matrix 4, and sorted according to the order of the tendency from small to large into matrix 4-matrix 3-matrix 2-matrix 1, at this time, it can be understood that the positive influence of matrix 1 on the accuracy of the data processing result is the largest, the positive influence of matrix 4 on the accuracy of the data processing result is the smallest, that is, the accuracy of matrix 1 is larger, the result after subsequent merging also tends to matrix 1, and the accuracy of the result after merging is also larger.
When merging is performed in the descending order of tendencies, the matrix 1 and the matrix 2 may be merged to obtain an intermediate merged matrix 12, where the tendency matrix when the matrix 1 and the matrix 2 are merged is the matrix 1 and the secondary tendency matrix is the matrix 2, and then the intermediate merged matrix 12 (i.e., the merged matrix obtained by merging the 2 nd matrix) is merged with the matrix 3 to obtain an intermediate merged matrix 123, where generally, the intermediate merged matrix 12 is the tendency matrix and the matrix 3 is the secondary tendency matrix. It can be determined in other ways, for example, in this case, the tendency matrix in the two matrices may be determined according to the tendency of the two matrices, the tendency of the intermediate combining matrix 12 may be an average value of the tendency of the matrices 1 and 2, or the tendency may be determined according to the combining coefficients, for example, the tendency of the intermediate combining matrix 12= the combining coefficient of the tendency matrix 1 + the combining coefficient of the tendency matrix 2 of the matrix 1, then the tendency of the intermediate combining matrix 12 is compared with the tendency of the matrix 3, and then the tendency matrix and the secondary tendency matrix are determined.
After the intermediate combining matrix 123 is obtained, the intermediate combining matrix 123 may be continuously combined with the matrix 4 to obtain a final combining matrix. The determination method of the tendency matrix and the secondary tendency matrix in the intermediate combination matrix 123 and the matrix 4 may also be determined in the same manner as described above, and it can be understood that the determination method of the tendency of the intermediate combination matrix may be flexibly set according to actual requirements, and the determination method is not limited to the above determination method.
In the same manner, when the tendencies are sorted from small to large, the merging order may be that the matrix 4 and the matrix 3 are merged to obtain the intermediate merged matrix 43, where the tendentiousness matrix is the matrix 3 and the secondary tendency matrix is the matrix 4, then the intermediate merged matrix 43 (secondary tendency matrix) is merged with the matrix 2 (tendency matrix) to obtain the intermediate merged matrix 432, and the intermediate merged matrix 432 (secondary tendency matrix) is merged with the matrix 1 (tendency matrix) to obtain the final merged matrix.
Or, during merging, merging may not be performed in sequence, and only any two matrixes need to be merged and then merged with the other matrixes until all the matrixes are merged to obtain a final merged matrix.
In the above manner, if the matrixes are arranged in descending order from large to small according to the tendency, the final combination result obtained in this way is most prone to the information of the first matrix and then the second matrix. The merging process described above can be expressed by the following function:
R out ＝f 2 (R 1 ,R 2 )；
R 1 and R 2 Representing the first two matrices after sorting, the combined matrix for these two matrices can be represented as R out，2 ＝f 2 (R 1 ,R 2 ) Then R is added out，2 And matrix R 3 Merging to obtain new intermediate merged matrix R out，3 ＝f 2 (R out,2 ,R 3 ) Until the N matrixes are combined to obtain a final combined matrix R out，N I.e. for an arbitrary number N of matrices, there areThe following recursive merging approach:
the final merged matrix R out ＝R out，N 。
It can be understood that, in the above merging manner, the matrices are sequentially merged according to the sorting order, so that the related information in the matrix with a large tendency can be effectively merged, so as to better retain more related information of the matrix with a large tendency in the final merged matrix, which is more beneficial to the subsequent calculation result.
In the implementation process, the matrixes are combined in sequence according to the tendency, so that the final combination result is more prone to the matrixes with large tendency, namely, the final combination result contains more direction information of the matrixes with large tendency, a more accurate combination result can be obtained, and the problem that other matrixes with large deviation have large negative influence on the combination result is solved.
In practical application, the following merging modes can be included:
mode (2): the combining may be performed in the above manner sequentially according to a random sorting order instead of sorting according to the tendency size, as long as the tendency matrix and the secondary tendency matrix are determined during each combining, and then the combining is performed according to the combining coefficient of the two matrices.
Mode (3): and sorting according to the tendency or randomly, combining the matrixes pairwise to obtain an intermediate combined matrix, and combining the intermediate combined matrix.
For example, for the 4 matrices exemplified above, the matrix 1 and the matrix 2 may be combined to obtain an intermediate combined matrix 12, the matrix 3 and the matrix 4 may be combined to obtain an intermediate combined matrix 34, and then the intermediate combined matrix 12 and the intermediate combined matrix 34 are combined to obtain a final combined matrix, where, of course, when the intermediate combined matrix 12 and the intermediate combined matrix 34 are combined again here, the intermediate combined matrix 12 may be generally determined as a trend matrix, and the intermediate combined matrix 34 may be determined as a secondary trend matrix, and of course, the trend matrix and the secondary trend matrix may also be determined according to the manner of obtaining the trend. Or, the matrix 1 and the matrix 3 may be combined to obtain the matrix 13, the matrix 2 and the matrix 4 may be combined to obtain the matrix 24, and then the matrix 13 and the matrix 24 may be combined to obtain a final combined matrix, where the trend matrix and the secondary trend matrix in the matrices 13 and 24 may be determined according to tendencies of the two matrices.
It should be understood that, in the above merging manner, each matrix will only participate in one merging, that is, one matrix will only be merged with the rest matrices or the intermediate merging matrix once, and regardless of the merging manner, all the N matrices will be finally merged.
In other merging manners, some matrices may participate in multiple merging, and the embodiment of the present application further provides another data processing method, which may be referred to as a matrix merging manner (4), as shown in fig. 2, including the following steps:
step S210: and sequencing the N matrixes from large to small according to the tendency to obtain the sequenced N matrixes. Wherein, the tendency characterization has positive influence on the accuracy of the data processing result, and N is an integer greater than 2.
Step S220: determining the ith matrix in the N sequenced matrixes as a secondary tendency matrix, sequentially taking values of i from M sequences with tolerance of 1, wherein M is a positive integer, the initial value of the 1 st sequence is 2, the ending value of the M sequence is N, the value range of the ending value of the jth sequence [2,N ], the value range of the initial value of the j +1 th sequence is [1,N-1], the initial value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j sequentially takes values from 1 to M.
Step S230: and when the value of i is 2 for the first time, determining the 1 st matrix in the N sequenced matrixes as the tendency matrix, otherwise, determining the combined matrix obtained by combining the k matrix for the last time as the tendency matrix, and combining the tendency matrix and the secondary tendency matrix according to the combination mode in the embodiment to obtain the combined matrix, wherein k is the last value of i.
Where, when M is 1, it means that there is only one sequence, if N is equal to 4, then the sequence is (2,3,4), and the merging mode is degenerated to the above mode (1), i.e. each matrix participates in one merging. And when M is larger than 1, a plurality of sequences exist, when i takes the initial value in the j +1 th sequence, the matrix position corresponding to the initial value is returned for combination, and the initial value of the j +1 th sequence represents the matrix position of matrix return iteration (which is smaller than the end value of the j th sequence and represents that the matrix sequence of the return iteration is before the matrix currently used for combination). The merging process may be understood as that a matrix (which may be understood as an obtained intermediate merged matrix) obtained by the latest merging of the kth (last value of i) matrix is merged with a matrix corresponding to the current value of i at least once, and after the merged matrix is obtained, the merged matrix is continued to be merged with the subsequent secondary trend matrix.
The M sequences may be preset, where the M sequences are used to determine the position and number of times of combining the matrix for each return iteration (for example, the start value of the sequence is used to determine the matrix position of the return iteration, the number of the sequences is used to determine the number of return iterations, and the end value of the sequence is used to determine the intermediate combining matrix that needs to be returned to the iteration).
For example, N is equal to 4, the N is sorted into a matrix 1, a matrix 2, a matrix 3 and a matrix 4, assuming that the 1 st sequence is (2,3), the 2 nd sequence is (1,2,3) and the 3 rd sequence is (2,3,4), when i takes 2 of the 1 st sequence, the matrix 1 and the matrix 2 are merged to obtain a matrix 12, when i takes 3 (k takes 2), the matrix 12 (i.e., an intermediate merged matrix obtained by merging the 2 nd matrix for the last time) is merged with the matrix 3 to obtain a matrix 123, then i takes 1 of the 2 nd sequence (k takes 3), which indicates that the matrix 123 (i.e., an intermediate merged matrix obtained by merging the 3 rd matrix for the last time) is merged with the matrix 1 to obtain a matrix 1231, then i takes 2 (k takes 1), the matrix 1 (i is merged with the matrix 2 (i is merged with the intermediate merged matrix obtained by merging the last time of the 1 st matrix) to obtain a matrix 1231, then i takes 1 (k takes 1), the matrix 1 (i, i is merged with the intermediate merged matrix obtained by merging the matrix 2 for the last time), and the sequence is merged with the sequence obtained by the sequence 3425, and then the sequence obtained by merging the sequence is merged again, and the sequence obtained by repeating the steps until the sequence is returned to obtain a value of the sequence. Of course, if there are many sequences, after the final merged matrix is obtained, it is still possible to continue to return to merge, for example, the 4 th sequence is (3,4), and then return to the 3 rd matrix for merging after the merging according to the 3 rd sequence is completed.
That is to say, in the embodiment (4), the merging may be performed in the same manner as the sequential merging manner (1) described above, but after one intermediate merged matrix (the matrix 12, the matrix 123, and the like described above) is obtained by merging at a certain time, the intermediate merged matrix may be used as an initial matrix to continue returning to the position of the previous matrix for merging, the number of times of returning may not be limited in this manner, and the position of the returned matrix is also not limited, and may be flexibly set according to actual requirements, and certainly the position of the intermediate merged matrix that needs to be iteratively merged may not be limited, for example, the position of the intermediate merged matrix that needs to be merged may be continuously returned to the position of the matrix 2 after the final merged matrix 1234 is obtained, and the position of the intermediate merged matrix that needs to be iteratively set flexibly according to actual requirements, that is flexibly set, and after one intermediate merged matrix is obtained, the matrix corresponding to the initial value of the next sequence is returned to be merged, and in the merging process, there may also be multiple intermediate merged matrices that need to return to be merged, and one intermediate merged again, and one intermediate merged may be returned multiple times, and the position of the matrices returned each time may be different. In other words, the number of sequences, the start value and the end value of each sequence can be flexibly set to control the intermediate merging matrix which needs to be merged by the return iteration, the matrix position which needs to be merged by the return iteration, the number of times that needs to be merged by the return iteration, and the like.
In other embodiments, it may also be determined whether iterative combining needs to be returned in real time according to the combining result, and the determination manner of the intermediate combining matrix that needs to be returned to continue iterative combining may also be determined according to the value thereof, for example, after the intermediate combining matrix is obtained by combining for a certain time, if the value of the intermediate combining matrix does not meet the setting requirement, returning to a certain set matrix to continue combining, and if the setting requirement is met, not returning, and continuing to combine afterwards.
In the above-mentioned method (4), the N matrices may not be sorted according to the tendency size, or the N matrices may be randomly sorted and then merged according to the merging method in the above-mentioned method (4). In this way, each matrix may participate in multiple combining, and certainly, an intermediate combining matrix among the matrices may also participate in multiple combining, for example, when each iteration combining is returned, a position where a certain intermediate combining matrix is located may also be returned, for example, the intermediate combining matrix 123 may be returned to be combined with the intermediate combining matrix 12, that is, the position where the combining is returned is not limited to a position where any one of the matrices that are previously combined is located, and may also be a position where any one of the intermediate combining matrices that are previously combined is located.
In the implementation process, iterative merging is continued by returning the intermediate merging matrix, so that the final merging result is more biased to a certain matrix direction, for example, to a more biased matrix direction, by multiple times of merging, so that the final merging result includes more information about the more biased matrix.
On the basis of any embodiment of the present application, the tendency of each matrix may be determined based on the number of statistical samples when each matrix is obtained, or, when each matrix is a statistical covariance matrix of interference noise obtained by a device based on an interference signal measurement result on a channel state information reference signal CSI-RS, the tendency of each matrix is determined based on a signal-to-noise ratio when each matrix is obtained.
For example, when the matrix 1 and the matrix 2 are merged, if the number of statistical samples of the obtained matrix 1 is greater than the number of statistical samples of the obtained matrix 2, the matrix 1 may be used as a trend matrix, and certainly, if the number of statistical samples of the two matrices is the same, one matrix may be arbitrarily selected as the trend matrix, where the number of statistical samples may refer to the data amount of related data in an actual application scenario, for example, the number of statistical samples refers to the data amount required for calculating to obtain the matrix 1 or the matrix 2.
Or, when the matrix is a statistical covariance matrix of the interference noise, the tendency of the matrix is determined based on the signal-to-noise ratio of the matrix, for example, the larger the signal-to-noise ratio is, the more statistical covariance information of the interference noise is included in the matrix, that is, the smaller the information of the interference noise is included, so that the matrix is taken as a tendency matrix, and the obtained combination matrix can include more useful information and less noise information.
It should be understood that, in practical applications, there may be other ways to determine the tendency of each matrix, for example, by manually defining the setting according to the accuracy of the matrix, for example, by manually determining which matrix has higher accuracy according to human experience, the tendency is larger, and when two matrices are merged, the corresponding tendency matrix may be manually identified. Alternatively, if the tendency of the matrix cannot be determined according to the number of statistical samples, the signal-to-noise ratio, or the accuracy, one of the tendency matrices may be randomly assigned.
In the implementation process, the tendency of the matrix is determined in this way, so that the reliability of the matrix can be determined according to the actual situation, so that the final combination result can be more inclined to the matrix with high reliability to obtain a more accurate combination result.
On the basis of any of the above embodiments, for N matrices participating in merging (where N may be an integer greater than 1), and any two matrices needing to be merged in the above embodiments, the dimensions of the matrices may be the same, if the dimensions of the matrices are different, the matrices may be integrated into a scene having the same dimensions in a corresponding manner, for example, N initial matrices participating in merging calculation are obtained at the beginning, if the dimensions of the N initial matrices are different, the maximum dimension of the N initial matrices is determined, and a matrix whose dimension does not reach the maximum dimension is subjected to element expansion, so that the dimension of the matrix reaches the maximum dimension, then N matrices whose elements are expanded are obtained, and the N matrices may be merged according to the above merging manner.
For example, by obtaining the dimension of each matrix, if the maximum dimension of a certain matrix is 40 × 40, and the dimensions of the other matrices are smaller than the maximum dimension, zero padding expansion is performed on the elements of each of the other matrices, so that the expanded dimension is 40 × 40, and thus the dimensions of each of the expanded matrices are the same, which can facilitate subsequent merging calculation. In addition, the position of the complementary element may be before the original element of the matrix, after the original element of the matrix, or between some elements of the matrix, and in practical application, the position of the complementary element may be set according to the actual situation.
In addition, the merging scheme of the present application is also applicable to a scenario where the computation domains to which the elements belong are also different due to the limitation of the statistical samples, for example, a domain expansion F is found, so that the computation defined by each matrix in its own domain can be defined in F, that is, the domain expansion F, so that each matrix can be expressed in F, and then after the inner product of the tendency matrix and the norm representation form induced by the inner product can be defined in F, the merging coefficients can be obtained in the above manner, and then the matrices are merged.
That is, if the computation domains to which the N matrices belong are different, the matrix whose computation domain is the real domain is converted into the complex domain to obtain N matrices of the complex domain, that is, the computation domains of the N matrices merged by the merging method are all in the complex domain. For example, if some of the N initial matrices are real number fields and some of the N initial matrices are complex number fields, the matrices in the complex number fields are not subjected to computational domain conversion, and the matrices in the real number fields are converted into complex number fields. The reason why the matrix is converted into the complex domain rather than the real domain is that if the matrix is converted into the real domain, the imaginary part of the matrix in the complex domain is lost, i.e., the conversion of the matrix in the complex domain into the real domain causes part of the information to be lost, which affects the subsequent merging result.
That is to say, the finally obtained N matrices that can be merged by the above merging method or any two matrices that need to be merged are complex-field and/or the dimensions of each matrix are the same, so that the merging calculation can be conveniently performed, which is beneficial to improving the calculation efficiency.
Of course, the scheme of the present application may not limit the dimension of the matrix, for example, when one of the number of columns or rows of the matrix is 1, the merging calculation may be performed in a degenerated vector form, and if both the number of rows and the number of columns are 1, the merging calculation may be performed in a degenerated scalar form. Moreover, the scheme of the present application may not limit a real matrix or a complex matrix, but may also be applicable to other matrices, such as matrices of other non-euclidean domains, where domains to which the respective matrices belong may be different, and dimensions of the matrices may also be different, and when performing the combination calculation, the matrices may be converted into the same domain and processed in the same dimension.
In the implementation process, the dimensionality or the calculation domain of the matrix is transformed and then combined, so that the combining method can be suitable for more matrix combining scenes.
On the basis of any of the foregoing embodiments, after obtaining the combining matrix, in different application scenarios, the combining matrix may be used to perform different processing, for example, if N matrices in the foregoing embodiment or each matrix in two matrices that need to be combined in the foregoing embodiments is a statistical covariance matrix of interference noise, after obtaining the combining matrix, relevant information of the interference noise may also be obtained according to the combining matrix, where the relevant information includes an amplitude and/or a direction.
It is understood that the combining matrix is a weighted sum result of each statistical covariance matrix of the interference noise, and may also be considered as a statistical covariance matrix of the interference noise, and the relevant information of the interference noise can be obtained based on the statistical covariance matrix. For example, eigenvalue decomposition may be performed on the merged matrix to obtain a noise subspace, then a spatial spectrum function may be constructed according to the noise subspace, and information such as the amplitude and/or direction of the interference noise may be obtained according to the spatial spectrum function.
Of course, when the merged matrix is a matrix of another type, different processing may be implemented according to requirements in different application scenarios, which are not listed here.
In addition, the scheme of the application can be applied to processing of modules in various communication systems, such as measurement, estimation, detection and the like, for example, direct combination of measurement matrixes, or combination of measurement matrixes after processing, such as result integration after schemes of amplitude adjustment in advance, matrix decomposition, matrix transportation and the like are performed.
The merging scheme provided by the application can adaptively calculate merging coefficients by taking the tendency matrix as a reference according to the correlation of each matrix in the matrixes to be merged, when one matrix is completely orthogonal to the direction of the matrix to be merged, the matrix with weaker tendency has smaller influence on the final merging result, and when one matrix is completely consistent with the matrix to be merged, the merging scheme can automatically degenerate into average merging, so that the adaptive adjustment of the merging coefficients is realized, rather than setting fixed weights, the scheme can realize more reasonable merging of the matrices, obtain more accurate merging results and reduce the influence on subsequent processing.
Referring to fig. 3, fig. 3 is a block diagram of a data processing apparatus 300 according to an embodiment of the present disclosure, where the apparatus 300 may be a module, a program segment, or code on an electronic device. It should be understood that the apparatus corresponds to the method embodiment of fig. 1, and can perform the steps related to the method embodiment of fig. 1, and the specific functions of the apparatus 300 can be referred to the description above, and the detailed description is omitted here appropriately to avoid redundancy.
Optionally, the apparatus 300 comprises:
a combining coefficient determining module 310, configured to determine a combining coefficient corresponding to a secondary tendency matrix and a tendency matrix by using a correlation between the secondary tendency matrix and the tendency matrix, where the tendency matrix is a matrix of the two matrices that has a greater positive impact on accuracy of a data processing result, and the secondary tendency matrix is another matrix of the two matrices except for the tendency matrix;
and a matrix merging module 320, configured to merge the two matrices according to the secondary tendency matrix and the merging coefficient corresponding to the tendency matrix.
Optionally, the merging coefficient determining module 310 is configured to calculate a norm of the tendency matrix, and obtain an inner product of the tendency matrix and the secondary tendency matrix under the definition of the norm; and calculating to obtain a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the norm of the tendency matrix and the inner product.
Optionally, the merging coefficient determining module 310 is configured to perform feature decomposition on the trend matrix to obtain a feature vector and a feature value; projecting the secondary tendency matrix to the characteristic direction of the tendency matrix according to the characteristic vector to obtain a projection matrix; and calculating and obtaining a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the characteristic value and the projection matrix.
Optionally, the tendency of each matrix is determined based on the number of statistical samples when each matrix is obtained, or when each matrix is a statistical covariance matrix of interference noise obtained by the device based on an interference signal measurement result on the CSI-RS, the tendency of each matrix is determined based on a signal-to-noise ratio when each matrix is obtained.
Optionally, the two matrices are preprocessed by: acquiring two initial matrixes participating in combination calculation; if the dimensionalities of the two initial matrixes are different, determining the maximum dimensionality of the two initial matrixes, and performing element expansion on the matrix with the dimensionality not reaching the maximum dimensionality to enable the dimensionality of the matrix to reach the maximum dimensionality to obtain the two matrixes after the element expansion, and/or if the two matrixes belong to different calculation domains, converting the matrix with the real number domain as the calculation domain into a complex number domain to obtain the two matrixes of the complex number domain.
Optionally, the two matrices are matrices characterizing signal-related information or matrices for signal processing.
Optionally, the two matrices are any one of the following matrices:
a statistical covariance matrix of interference noise obtained based on interference signal measurement results on the CSI-RS;
a channel covariance matrix characterizing channel antenna or data stream correlations;
a channel response matrix obtained by channel estimation;
a filter matrix for parametric or signal noise reduction.
Optionally, when each of the two matrices is a statistical covariance matrix of interference noise, the apparatus 300 further includes:
and the matrix processing module is used for acquiring the relevant information of the interference noise according to the combined matrix, wherein the relevant information comprises amplitude and/or direction.
Referring to fig. 4, fig. 4 is a block diagram of another data processing apparatus 400 according to an embodiment of the present application, where the apparatus 400 may be a module, a program segment, or code on an electronic device. It should be understood that the apparatus corresponds to the above-mentioned embodiment of the method of fig. 2, and can perform the steps related to the embodiment of the method of fig. 2, and the specific functions of the apparatus 400 can be referred to the above description, and the detailed description is appropriately omitted here to avoid redundancy.
Optionally, the apparatus 400 comprises:
the sorting module 410 is configured to, when N matrices are merged, sort the N matrices from large to small according to a tendency to obtain N sorted matrices, where the tendency represents a size of a positive influence on accuracy of a data processing result, and N is an integer greater than 2;
a value module 420, configured to determine an ith matrix of the N sequenced matrices as the sub-trend matrix, and sequentially take values of i from M sequences with a tolerance of 1, where M is a positive integer, where a starting value of the 1 st sequence is 2, an ending value of the M sequence is N, a value range of the ending value of the jth sequence is [2,N ], a value range of a starting value of the j +1 th sequence is [1,N-1], a starting value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j sequentially takes values from 1 to M;
a merging module 430, configured to determine a 1 st matrix of the N sequenced matrices as the trend matrix when the value of i is 2 for the first time, otherwise, determine a merged matrix obtained by merging the k-th matrix for the last time as the trend matrix, and merge the trend matrix and the secondary trend matrix according to the method provided in the foregoing embodiment to obtain a merged matrix, where k is a last value of i.
It should be noted that, for the convenience and brevity of description, the specific working procedure of the above-described apparatus may refer to the corresponding procedure in the foregoing method embodiment, and the description is not repeated herein.
Referring to fig. 5, fig. 5 is a schematic structural diagram of an electronic device for executing a data processing method according to an embodiment of the present application, where the electronic device may include: at least one processor 510, such as a CPU, at least one communication interface 520, at least one memory 530, and at least one communication bus 540. Wherein the communication bus 540 is used for realizing direct connection communication of these components. The communication interface 520 of the device in the embodiment of the present application is used for performing signaling or data communication with other node devices. Memory 530 may be a high-speed RAM memory or a non-volatile memory (e.g., at least one disk memory). Memory 530 may optionally be at least one memory device located remotely from the aforementioned processor. The memory 530 stores computer readable instructions, and when the computer readable instructions are executed by the processor 510, the electronic device executes the method process shown in fig. 1 or fig. 2.
It will be appreciated that the configuration shown in fig. 5 is merely illustrative and that the electronic device may include more or fewer components than shown in fig. 5 or have a different configuration than shown in fig. 5. The components shown in fig. 5 may be implemented in hardware, software, or a combination thereof.
Embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs the method processes performed by an electronic device in the method embodiments shown in fig. 1 or fig. 2.
The present embodiments disclose a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the methods provided by the above-described method embodiments, for example, comprising: determining a combination coefficient corresponding to a secondary tendency matrix and a tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix, wherein the tendency matrix refers to a matrix which has a larger positive influence on the accuracy of a data processing result in the two matrices, and the secondary tendency matrix refers to another matrix except the tendency matrix in the two matrices; and combining the two matrixes according to the combination coefficient corresponding to the secondary tendency matrix and the tendency matrix.
In summary, the embodiments of the present application provide a data processing method, an apparatus, an electronic device, a storage medium, and a program product, where when performing matrix merging, a merging coefficient is determined by using a correlation between a trend matrix that has a greater positive impact on accuracy of a data processing result and another trend matrix, and the merging coefficient is used for merging, where the merging coefficient may be adaptively calculated according to a trend in merging, so that a result after merging may be more likely to be a trend matrix, where the trend matrix may be a matrix with a smaller deviation or higher reliability, so that a deviation of a final merging result is smaller and the reliability is higher.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
In addition, units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
Furthermore, the functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A method of data processing, the method comprising:
determining a combination coefficient corresponding to a secondary tendency matrix and a tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix, wherein the tendency matrix refers to a matrix which has a larger positive influence on the accuracy of a data processing result in the two matrices, and the secondary tendency matrix refers to another matrix except the tendency matrix in the two matrices;
and combining the two matrixes according to the combination coefficient corresponding to the secondary tendency matrix and the tendency matrix.
2. The method according to claim 1, wherein the determining the combining coefficients of the secondary tendency matrix and the tendency matrix by using the correlation between the secondary tendency matrix and the tendency matrix comprises:
calculating the norm of the tendency matrix, and acquiring the inner product of the tendency matrix and the secondary tendency matrix under the definition of the norm;
calculating and obtaining a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the norm of the tendency matrix and the inner product;
or, the determining the combining coefficient corresponding to the secondary tendency matrix and the tendency matrix by using the correlation between the secondary tendency matrix and the tendency matrix includes:
performing characteristic decomposition on the tendency matrix to obtain a characteristic vector and a characteristic value;
projecting the secondary tendency matrix to the characteristic direction of the tendency matrix according to the characteristic vector to obtain a projection matrix;
and calculating and obtaining a merging coefficient corresponding to the tendency matrix and the secondary tendency matrix according to the characteristic value and the projection matrix.
3. The method of claim 1, wherein the two matrices are preprocessed by:
acquiring two initial matrixes participating in combination calculation;
if the dimensionalities of the two initial matrixes are different, determining the maximum dimensionality of the two initial matrixes, and performing element expansion on the matrix with the dimensionality not reaching the maximum dimensionality to enable the dimensionality of the matrix to reach the maximum dimensionality to obtain the two matrixes after the element expansion, and/or if the two matrixes belong to different calculation domains, converting the matrix with the real number domain as the calculation domain into a complex number domain to obtain the two matrixes of the complex number domain.
4. The method of claim 1, wherein the two matrices are matrices characterizing signal-related information or matrices for signal processing.
5. A method of data processing, the method comprising:
when N matrixes are combined, the N matrixes are sorted from large to small according to the tendency to obtain the sorted N matrixes, wherein the tendency represents the positive influence on the accuracy of a data processing result, and N is an integer greater than 2;
determining the ith matrix in the N sequenced matrixes as the secondary tendency matrix, sequentially taking values of i from M sequences with tolerance of 1, wherein M is a positive integer, the initial value of the 1 st sequence is 2, the ending value of the M sequence is N, the ending value of the jth sequence is in a value range of [2,N ], the initial value of the j +1 th sequence is in a value range of [1,N-1], the initial value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j is sequentially taken from 1 to M;
and when the first value of i is 2, determining the 1 st matrix in the N sequenced matrixes as the tendency matrix, otherwise, determining a combined matrix obtained by combining the k matrix for the last time as the tendency matrix, and combining the tendency matrix and the secondary tendency matrix according to the method of any one of claims 1 to 4 to obtain a combined matrix, wherein k is the last value of i.
6. A data processing apparatus, characterized in that the apparatus comprises:
the merging coefficient determining module is used for determining merging coefficients corresponding to a secondary tendency matrix and a tendency matrix by utilizing the correlation between the secondary tendency matrix and the tendency matrix, wherein the tendency matrix is a matrix which has a larger positive influence on the accuracy of a data processing result in the two matrices, and the secondary tendency matrix is the other matrix except the tendency matrix in the two matrices;
and the matrix merging module is used for merging the two matrixes according to the secondary tendency matrix and the merging coefficient corresponding to the tendency matrix.
7. A data processing apparatus, characterized in that the apparatus comprises:
the sorting module is used for sorting the N matrixes from large to small according to the tendency when the N matrixes are combined to obtain the N sorted matrixes, wherein the tendency represents the positive influence on the accuracy of a data processing result, and N is an integer greater than 2;
a value taking module, configured to determine an ith matrix in the N sequenced matrices as the secondary tendency matrix, and sequentially take values of i from M sequences with a tolerance of 1, where M is a positive integer, where a starting value of the 1 st sequence is 2, an ending value of the M th sequence is N, a value range of the ending value of the jth sequence is [2,N ], a value range of a starting value of the j +1 th sequence is [1,N-1], a starting value of the j +1 th sequence is smaller than the ending value of the jth sequence, and j sequentially takes values from 1 to M;
a merging module, configured to determine a 1 st matrix of the N sequenced matrices as the trend matrix when the value of i is 2 for the first time, otherwise, determine a merged matrix obtained by merging the k-th matrix for the last time as the trend matrix, and merge the trend matrix and the secondary trend matrix according to the method of any one of claims 1 to 4 to obtain a merged matrix, where k is a last value of i.
8. An electronic device comprising a processor and a memory, the memory storing computer readable instructions that, when executed by the processor, perform the method of any of claims 1-5.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 5.
10. A computer program product comprising computer program instructions which, when read and executed by a processor, perform the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210945441.3A CN115314345B (en) | 2022-08-08 | 2022-08-08 | Data processing method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210945441.3A CN115314345B (en) | 2022-08-08 | 2022-08-08 | Data processing method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115314345A true CN115314345A (en) | 2022-11-08 |
| CN115314345B CN115314345B (en) | 2023-03-21 |
Family
ID=83860423
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210945441.3A Active CN115314345B (en) | 2022-08-08 | 2022-08-08 | Data processing method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115314345B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110038445A1 (en) * | 2009-08-13 | 2011-02-17 | Qualcomm Incorporated | Communications channel estimation |
| US20190074999A1 (en) * | 2015-12-18 | 2019-03-07 | Intel IP Corporation | Communication terminal and method for channel estimation |
| CN110611626A (en) * | 2018-06-15 | 2019-12-24 | 上海华为技术有限公司 | Channel estimation method, device and equipment |
| CN111443897A (en) * | 2020-03-30 | 2020-07-24 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| CN113271130A (en) * | 2018-05-11 | 2021-08-17 | 华为技术有限公司 | Channel estimation method and device |
| WO2022091228A1 (en) * | 2020-10-27 | 2022-05-05 | 日本電気株式会社 | Eigenvalue decomposition device, wireless communication device, method, and non-transitory computer-readable medium |
| CN114629754A (en) * | 2022-05-13 | 2022-06-14 | 成都爱瑞无线科技有限公司 | Interference noise equalization method, system and storage medium |
-
2022
- 2022-08-08 CN CN202210945441.3A patent/CN115314345B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110038445A1 (en) * | 2009-08-13 | 2011-02-17 | Qualcomm Incorporated | Communications channel estimation |
| US20190074999A1 (en) * | 2015-12-18 | 2019-03-07 | Intel IP Corporation | Communication terminal and method for channel estimation |
| CN113271130A (en) * | 2018-05-11 | 2021-08-17 | 华为技术有限公司 | Channel estimation method and device |
| CN110611626A (en) * | 2018-06-15 | 2019-12-24 | 上海华为技术有限公司 | Channel estimation method, device and equipment |
| CN111443897A (en) * | 2020-03-30 | 2020-07-24 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| WO2022091228A1 (en) * | 2020-10-27 | 2022-05-05 | 日本電気株式会社 | Eigenvalue decomposition device, wireless communication device, method, and non-transitory computer-readable medium |
| CN114629754A (en) * | 2022-05-13 | 2022-06-14 | 成都爱瑞无线科技有限公司 | Interference noise equalization method, system and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115314345B (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Swirszcz et al. | Grouped orthogonal matching pursuit for variable selection and prediction | |
| Wang | Factor profiled sure independence screening | |
| US7864077B2 (en) | System and method for encoding a signal using compressed sensor measurements | |
| CN109633538B (en) | Maximum likelihood time difference estimation method of non-uniform sampling system | |
| CN107450047A (en) | Compressed sensing DOA estimation method based on unknown mutual coupling information under nested battle array | |
| CN114268388B (en) | Channel estimation method based on improved GAN network in large-scale MIMO | |
| CN110932807B (en) | Spectrum sensing method and system for MIMO (multiple input multiple output) system under non-Gaussian noise | |
| Chiancone et al. | Student sliced inverse regression | |
| US7584396B1 (en) | System and method for decoding a signal using compressed sensor measurements | |
| WO2022089217A1 (en) | Key generation method and apparatus, electronic device, and storage medium | |
| US20230353426A1 (en) | Channel Estimation for an Antenna Array | |
| CN115314345B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN113051739A (en) | Robustness self-adaptive processing method based on sparse constraint | |
| Zhang et al. | Subset pursuit for analysis dictionary learning | |
| CN108896967B (en) | Method and device for detecting distance extended target based on clutter covariance matrix estimation | |
| CN116319190A (en) | GAN-based large-scale MIMO system channel estimation method, device, equipment and medium | |
| CN109991578A (en) | Multi -components radar signal Modulation Identification method based on blind compression core dictionary learning | |
| Khare et al. | A convex framework for high-dimensional sparse Cholesky based covariance estimation | |
| CN114362851B (en) | Wireless channel data denoising method based on machine learning | |
| Miller et al. | Independent components analysis by direct entropy minimization | |
| Gonzalez et al. | Robust clustering using tau-scales | |
| Li et al. | Asymptotic normality and confidence intervals for prediction risk of the min-norm least squares estimator | |
| CN111314000B (en) | Radio frequency tomography method based on low-rank data-driven weight model | |
| Su et al. | Underdetermined independent component analysis based on first-and second-order statistics | |
| WO2024027321A1 (en) | Large-dimensional communication signal processing method and apparatus, and electronic device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |