160 results about "Column vector" patented technology

The invention provides a vehicle fault root cause diagnosis method, equipment, equipment and a medium, and the method comprises the steps: obtaining a vehicle target fault phenomenon, determining a to-be-detected fault event set based on a preset fault logic relation, constructing an initial scoring matrix, enabling a row vector to correspond to a fault event, enabling a column vector to correspond to the state feature parameters of a plurality of evaluation dimensions, and carrying out the detection of the fault event set; executing the detection task of the highest comprehensive score fault event, obtaining feedback data, updating the initial score matrix parameters based on the feedback data to obtain an updated score matrix, and calculating the comprehensive score of each event based on the updated score matrix again. Iteratively executing the detection task corresponding to the updated highest score event to identify whether the detection event is a fault root cause or not until the fault root cause of the target fault phenomenon is determined; according to the method, through a dynamic priority scheduling mechanism, real-time feedback data is fused in multi-dimensional evaluation, a high-value diagnosis task is executed preferentially, the resource consumption of traditional traversal diagnosis is remarkably reduced, and the fault positioning efficiency is effectively improved.

The invention provides a Cholesky decomposition heterogeneous parallel optimization method and system based on an SW architecture, and relates to the technical field of high-performance computing. The method comprises the following steps: performing sub-block division on a symmetric positive definite matrix based on a distributed parallel distribution scheme, and performing iteration to complete matrix decomposition; each sub-block is distributed to different processes through an MPI programming model, data exchange is carried out between the processes through asynchronous communication, and coarse-grained task-level parallel acceleration is carried out; performing two-stage parallel acceleration on four operations in Cholesky decomposition by utilizing the acceleration parallel characteristic of a master core and a slave core of the SW architecture; wherein for GEMM and SYRK operations, column vectors of a matrix are mapped to a slave core array, and columns are divided according to the number of slave cores; the calculation process is optimized through a double-buffering mechanism, vectorization operation and a loop expansion technology, and the parallel efficiency is improved; and for the TRSM operation, the TRSM operation is decomposed into a plurality of TRSV operations, the TRSV operations are allocated to the slave cores for parallel execution, and a circular reading and data broadcasting mode is adopted to reduce data dependence and realize efficient parallel calculation.

The invention discloses a hydrogeological parameter space-time heterogeneity geostatistics inversion method. The method comprises the following steps: constructing a to-be-solved hydrogeological parameter column vector based on a geologic model and a'space-time equivalence 'principle; a priori covariance matrix of partition parameters in a time domain is assembled in a blocking mode; carrying out dimension reduction on the covariance matrix based on a principal component analysis method; carrying out parameter sensitivity matrix approximate calculation, solving a collaborative Kriging equation, and calculating an estimated value of a to-be-solved parameter by utilizing an iteration mode; and determining a target function, and performing iterative updating according to the target function to obtain a to-be-solved parameter column vector. The invention designs a partition geological statistics dimension reduction inversion method for describing the spatial-temporal heterogeneity of the hydrogeological parameters, and provides an efficient and accurate solution for describing the spatial-temporal heterogeneity of the parameters in a complex spatial-temporal process.

The invention discloses a mixed hexahedron magnetic data inversion method based on random weighting, and belongs to the field of magnetic signal processing, and the method comprises the steps: carrying out the space division of an underground inversion region through employing a structure and non-structure combined mixed hexahedron grid, and discretizing an underground space into N hexahedron units; generating projection geometries with different size parameters and spatial positions through random assignment, embedding the projection geometries into the subdivided grid model, repeating N times to construct N independent N * 1-dimensional model column vectors, and then performing random sequential splicing to obtain an N * N-dimensional random weighting matrix; constructing an inversion objective function based on the random weighting matrix, and solving to obtain weighting domain magnetization intensity; and in combination with a random weighting matrix and a solving result, obtaining spatial distribution of three-dimensional magnetization intensity, and realizing construction of an inversion image. According to the invention, rapid and fine inversion of the geophysical magnetic data under the rugged topography is realized, and the inversion precision and reliability of the magnetic survey data under the complex topographic condition are improved.

The invention discloses a dual robust principal component analysis image denoising method based on a graph, and belongs to the technical field of image processing. The method comprises the following steps: flattening an image to be processed to construct a data matrix in a column vector form; constructing a graph structure by using a K nearest neighbor method, and generating a graph Laplacian matrix based on the graph structure; jointly considering an image reconstruction error, a sparse noise item, a linear mapping error item and a graph structure regular item, and constructing an optimization model; and carrying out variable alternating optimization by adopting an augmented Lagrange multiplier method and an alternating direction solution method, and obtaining an image denoising result according to the low-rank principal component. According to the method, image structure information and a double constraint mechanism are introduced into a robust principal component analysis framework, so that the detail retention capability and the structure consistency of the image are effectively enhanced, the robustness and the visual quality of image denoising are improved, and the method is suitable for application scenes such as image processing, video monitoring and target detection under complex backgrounds.

A communication method includes the following. A terminal device sends a first physical uplink shared channel (PUSCH) and a second PUSCH. The terminal device corresponds to N sounding reference signal (SRS) resource sets, the first PUSCH is associated with a first SRS resource set among the N SRS resource sets, the second PUSCH is associated with a second SRS resource set among the N SRS resource sets, and the first SRS resource set is different from the second SRS resource set. The terminal device includes M antenna ports. A first precoding matrix for the first PUSCH and a second precoding matrix for the second PUSCH belong to a first precoding matrix set, a non-zero element in a column vector of each precoding matrix in the first precoding matrix set is less than or equal to M / N, and M and N are integers greater than or equal to 2.

The invention discloses a large model training method based on second-order matrix optimization, and belongs to the field of deep learning model training technology optimization, and the large model training method based on second-order matrix optimization comprises the following steps: S1, decomposing a second-order matrix into row and column vectors, and carrying out sliding average and distributed block reduction storage; s2, generating a statistical row vector by combining row gradient aggregation with a historical attenuation factor; s3, carrying out block distributed statistics in the column direction and synchronously generating column vectors across equipment; s4, a low-rank matrix is constructed through outer products of row and column vectors, and estimation precision is improved through noise suppression; s5, performing dynamic sparse sampling, performing initial high-density focusing, and stabilizing the sampling rate of a key layer; s6, the sampling points execute time sequence attenuation updating, and asynchronous calculation is carried out to improve the resource utilization rate; s7, performing Gaussian kernel smoothing on a neighborhood value compensation coverage gap in an unsampled region; s8, fusing low-rank estimation and sparse data, and balancing global precision by self-adaptive weight; the method has the beneficial effects of reducing video memory occupation, and improving distributed computing efficiency and balance training precision and speed.

The invention discloses a planetary roller screw dynamic modeling method based on a generalized finite element method, and belongs to the field of dynamic characteristic analysis of planetary roller screws. The method comprises the following steps: carrying out unit division on a planetary roller screw finite element model; constructing a dynamic model of each type of unit; respectively establishing mass matrixes, stiffness matrixes, damping matrixes and load column vectors of the shaft section unit, the lead screw-roller contact side thread meshing unit, the nut-roller contact side thread meshing unit and the roller tooth-inner gear ring gear meshing unit; and assembling a system dynamics model to complete planet roller screw dynamics modeling. According to the method, the dynamic characteristics of the real physical model can be accurately analyzed, and an effective means is provided for analyzing the vibration response of the system.

The invention relates to a matrix decomposition device and method based on a memristor cross array, which are suitable for efficient hardware implementation of singular value decomposition (SVD), the memristor cross array receives a conductance value mapped by a Grubrum matrix constructed by a to-be-decomposed matrix and a bit line input voltage converted by a bit line column vector, and outputs a corresponding current; converting the corresponding current into a voltage vector; performing normalization processing on the voltage vector to obtain a normalized vector; judging whether convergence occurs or not; performing normalization processing on the steady-state voltage vector to obtain a final output vector; and according to the final output vector, main characteristic values are extracted based on a first normalization circuit, and main singular values and corresponding singular matrixes are calculated. Compared with the prior art, the high parallel computing characteristic of the memristor cross array is utilized, efficient operation and hardware acceleration of matrix decomposition are achieved, the method has the advantages of being low in power consumption, high in speed and good in expansibility, and the method is suitable for application scenes such as artificial intelligence, signal processing and large-scale matrix operation.

The embodiment of the invention provides a light spot feature calculation method, device and equipment for a dynamic image, and is applied to the technical field of optical image processing, and the method comprises the steps: carrying out the summation of each row of a non-zero pixel matrix, obtaining a column vector, multiplying the column vector by an index row vector, dividing the product sum by the sum of non-zero pixels, and obtaining the light spot feature of the non-zero pixel matrix. Obtaining a y coordinate value; summing each column of the non-zero pixel matrix to obtain a row vector, multiplying the row vector by the index column vector, and dividing the product sum by the sum of the non-zero pixels to obtain an x coordinate value; according to the x coordinate value and the y coordinate value, x-direction centroid coordinates and y-direction centroid coordinates of the light spot are obtained; calculating the variance of the x direction and the y direction according to the x coordinate value and the y coordinate value, and calculating the light spot roundness and the light spot diameter based on a least square method according to the variance of the x direction and the y direction; and calculating a beam waist radius and a beam divergence angle based on the light spot diameter. In this way, the real features of the light spots in the dynamic image can be accurately calculated.

The invention discloses a method for improving photoetching performance of a dipole light source. The method comprises the following steps of: finding out a selected area with an OPC (Optical Proximity Correction) segment coupling effect in a layout graph, selecting a plurality of adjacent selected OPC segments in the selected area, and combining the selected OPC segments to form a segment group; the row number and the column number of the Jacobian matrix are equal to the number of the selected OPC segments, and each element of a row vector of the Jacobian matrix is a derivative of the light intensity of one selected OPC segment corresponding to the row vector to the position coordinate of each selected OPC segment; each element of a column vector of the Jacobian matrix is a derivative of the light intensity of each selected OPC segment to the position coordinates of one selected OPC segment corresponding to the column vector. A Jacobian matrix is added into a cost function, then MB-OPC correction is carried out, and a light source in an OPC model subjected to MB-OPC correction is a dipole light source. According to the invention, the OPC convergence speed of an area with a strong OPC segmented coupling effect can be accelerated, and the contrast ratio can be improved.

The invention discloses a submarine cable temperature rapid calculation method based on intrinsic orthogonal decomposition, and the method comprises the steps: S1, determining a sample space of submarine cable temperature, and obtaining a sample working condition data set; s2, node temperature column vectors of all sample working conditions are calculated based on the submarine cable simulation calculation model, a snapshot matrix is formed, and a base vector and a modal coefficient matrix of the submarine cable temperature field reduced-order model are obtained according to the snapshot matrix; s3, constructing a polynomial response surface model corresponding to the modal coefficient matrix and the sample working condition, and calculating a corresponding modal coefficient according to the input test working condition; and S4, reconstructing a submarine cable temperature field reduced-order model according to the basis vector and the modal coefficient. Compared with a finite element multi-physics field coupling calculation method for the submarine cable temperature, the calculation speed of the submarine cable temperature field reduced-order model is increased by about 800 times, the real-time requirement of cable temperature calculation in digital twin application can be met, and the accuracy is improved.

This invention discloses a method for inverting geophysical magnetic data based on random weighting, belonging to the field of magnetic signal processing. The method includes: spatially partitioning the subsurface inversion area using a hybrid hexahedral mesh combining structured and unstructured elements, discretizing the subsurface space into N hexahedral units; generating projected geometries with different size parameters and spatial locations through random assignment, embedding them into the partitioned mesh model, repeating this process N times to construct N independent N×1 dimensional model column vectors, which are then randomly sequentially concatenated to obtain an N×N dimensional random weighting matrix; constructing an inversion objective function based on the random weighting matrix, and solving for the weighted domain magnetization intensity; combining the random weighting matrix and the solution results to obtain the spatial distribution of three-dimensional magnetization intensity, thus constructing the inversion image. This invention enables rapid and precise inversion of geophysical magnetic data under undulating terrain, improving the accuracy and reliability of magnetic data inversion under complex terrain conditions.

This invention provides a method and apparatus for constructing an LSSS access structure that supports collaborative decryption. The method includes: constructing an LSSS access structure according to a collaborative access strategy, wherein an access matrix M in the LSSS access structure is used to determine a first column vector λ, and the first column vector λ includes a second column vector λ. s ; and the sub-access matrix M used to construct the sub-strategy sub Determine the sub-access matrix M sub The corresponding third column vector λ sub The second column vector λ s Equal to the third column vector λ sub The first column vector λ and the third column vector λ sub Used for encryption; Sub-access matrix M sub and the third column vector λ sub Used for decryption. In this embodiment of the invention, the second column vector λ is used... s Equal to the third column vector λ sub Thus, for the third column vector λ sub The relevant calculation results can be used in the relevant calculations for the first column vector λ, saving computational overhead and thus improving encryption and decryption efficiency.

An arithmetic processing apparatus is a system for matrix multiplication using M arithmetic units connected in series. It processes two M×N-dimensional input matrices (a and b) to produce an M×M-dimensional output matrix (x). Each arithmetic unit stores and processes a specific column vector pair from the input matrices. The calculation occurs through a distributed process where each Mth arithmetic unit stores corresponding column vectors [am, bm] from matrices a and b, column vectors bm are sequentially propagated through the network, each unit calculates dot products between its stored am vector and all incoming b vectors, and the resulting dot products form M-dimensional column vectors (xm) that make up the final output matrix. This distributed approach allows parallel processing of matrix multiplication across the network of arithmetic units, with each unit contributing a portion of the final result.

The present application is applicable to the technical field of integrated circuits and neural networks. Provided are a matrix multiplication apparatus and method based on a systolic array, and an electronic device. The matrix multiplication apparatus based on a systolic array comprises: a systolic array formed by means of arranging processing elements, wherein two adjacent processing elements in each column are connected by a shift, each processing element has transmission channels in a row direction, a column direction and a diagonal direction, the row direction is used for inputting a row vector of a first matrix, and the column direction and the diagonal direction are used for inputting a column vector of a second matrix; a plurality of first control signal channels, which are used for inputting a first control signal for controlling the operating state of each shift; and a plurality of second control signal channels, which are used for inputting a second control signal for controlling each column vector of the second matrix to be input in the column direction or the diagonal direction. Therefore, a hardware structure taking into consideration a sparse matrix multiplication operation and a dense matrix multiplication operation is achieved, and the calculation complexity of matrix multiplication is reduced.

The invention relates to a voice tag generation method and device, electronic equipment and a storage medium, and belongs to the technical field of computers, and the method comprises the steps: carrying out the feature matrix decomposition of a plurality of to-be-processed first voice feature matrixes, and obtaining a decomposed first matrix and a decomposed second matrix; performing hierarchical clustering on column vectors in the first matrix to obtain a tag tree of the first matrix; performing projection processing on the second voice according to the second matrix to obtain a voice feature representation vector of which the dimension is the same as the column number of the first matrix; querying a tag tree, and determining a feature tag of the second voice; wherein the feature tag of the second voice is determined according to the clustering tag in the tag tree mapped by the voice feature representation vector. According to the embodiment of the invention, the construction and maintenance cost of a voice feature tag generation system can be remarkably reduced, and the automation degree and robustness of tag generation are improved.

The invention discloses an RMFE-based distributed matrix multiplication implementation method. The method comprises the following steps: initializing parameters, constructing an RMFE function, splitting a to-be-processed matrix sequence into sequence vectors, sequentially carrying out phi function operation, obtaining a matrix distributed multiplication result through distributed matrix multiplication, and recovering through psi function operation to obtain a multiplication result of a to-be-processed matrix. According to the method, a distributed multiplication problem of a plurality of matrixes on a small domain / Galois ring is converted into a distributed multiplication problem of a single matrix on a large domain / Galois ring, and the calculation overhead on an expansion domain is allocated to the plurality of matrixes, so that the communication overhead required by distributed matrix multiplication is effectively reduced; the method has important application value for a high-speed communication system.

The invention discloses a multi-unmanned aerial vehicle air combat game payment matrix dimensionality reduction method under incomplete information, and belongs to the field of multi-unmanned aerial vehicle air combat dimensionality reduction game decision. An interval matrix game under incomplete information is converted into a matrix game with parameters, then an equivalent determined value matrix game is constructed, an effective convex hull vertex of a row or column vector set of a matrix is found out by using a convex hull vertex algorithm based on a reward value, and a matrix after dimension reduction is established according to the effective convex hull vertex. And then a mapping function is used to establish a relationship between Nash equilibrium of the matrix game before and after dimension reduction, a solution of an equivalent matrix can be obtained by solving the Nash equilibrium of the matrix after dimension reduction, and a Nash equilibrium solution and expected benefits of the interval matrix game are obtained through transformation. Therefore, the problems that the number of the unmanned aerial vehicles is increased, so that dimensionality of a matrix game is rapidly increased, and sensor information is uncertain due to a complex air combat environment are solved.

The present invention relates to the technical field of neural network compression, and proposes a neural network compression method and a neural network accelerator based on a memristor, comprising the following steps: converting the weights of a pre-trained neural network stored in a memristor into a plurality of column vector subsets; performing a pruning operation on each column vector subset to obtain a first pruned column vector subset; searching for an optimal pruning rate layer by layer; obtaining a second pruned column vector subset using the optimal pruning rate, and encoding the second pruned column vector subset to obtain a plurality of mask subsets; obtaining each third pruned column vector subset using each mask subset; solving the optimal third pruned column vector subset, and clustering them to obtain a clustering sub-matrix set; mapping the clustering sub-matrix set to a memristor, and after the mapping is completed, retraining the pre-trained neural network based on the memristor to complete the compression of the memristor-based neural network.

The application discloses a new correlation image recognition and machine learning method based on fuzzy monotony, and belongs to the technical field of artificial intelligence of pattern recognition and machine learning; the method is defined as FMMCA, which evaluates local fuzzy monotone correlation by comparing row vectors and column vectors of an image matrix pair by pair, then the local correlations are weighted and accumulated, and finally the global fuzzy monotone correlation between images is obtained. The application directly uses fuzzy monotone correlation analysis to replace classical correlation analysis for multi-view research, so that the problems existing in classical correlation analysis do not exist, and the fuzzy monotone method feature does not need to be measured statically by distance, but can be measured dynamically by interval change, so that the influence of some noise is offset, the performance is improved, a new fuzzy monotone machine learning method is formed, the optimization of a focus loss function is not needed, the parameters are few, the robustness is good, and the computing power is small.

The invention relates to the technical field of electrical digital data processing, and discloses a Web application automatic generation and data processing method based on spreadsheet mapping, which comprises the following steps: mapping spreadsheet cells into enumeration values to construct a type fingerprint matrix; calculating a type purity index of a column vector, identifying a vertical blocking band, and splitting the matrix into a plurality of orthogonal sub-matrixes which are not communicated with each other; calculating a type distribution difference value of adjacent row vectors in the sub-matrix to determine a horizontal logic partition line so as to define a logic data block; a Web application object model is assembled based on an arrangement mode of logic data blocks, signal crosstalk between transverse side-by-side data entities is cut off by introducing a vertical topology decoupling mechanism based on column entropy, and the technical problem that downlink feature scanning is easily interfered by orthogonal noise in a multi-table mixed arrangement scene is solved. And the deterministic structure reconstruction of the complex two-dimensional heterogeneous data under the condition of no preset template is realized.

A method includes: obtaining a first sequence, where the first sequence includes T first subsequences; determining T encoding vectors based on the T first subsequences; generating T second subsequences based on one or more codebooks and the T encoding vectors, where each first subsequence corresponds to one column vector group in the one or more codebooks, and a column vector group corresponding to at least one of the T first subsequences includes at least two column vectors; and sending the T second subsequences.

The invention discloses a Bayesian learning-based tensor interpolation decomposition method and device, and the method comprises the following steps: S1, expanding a multi-order tensor into a matrix according to one dimension, carrying out the clustering of the matrix, extracting a certain number of column vectors from different clustering types, and organizing the column vectors into a skeleton matrix; s2, regarding the skeleton matrix as a factor matrix in CP decomposition, and regarding the remaining factor matrixes as weight matrixes; carrying out probability modeling on tensor interpolation decomposition by utilizing leaf bass learning, and adding prior to the tensor; s3, setting prior for the weight matrix, drawing a probability graph model according to the prior of the weight matrix, and exporting model posterior distribution; s4, using a Gibbs sampling algorithm to carry out posterior approximate calculation of the weight matrix; and S5, calculating a model error value based on the tensor approximate value, if the model error value is smaller than a set threshold value, outputting a weight matrix, and otherwise, carrying out iteration on the step S4. The method has the effect of high interpretability.

The present application belongs to the field of communications. Provided are a channel state information processing method and apparatus, so as to further improve the channel estimation performance. The method comprises: a first communication apparatus determining some column vectors in an eigenvector matrix corresponding to a channel between a second communication apparatus and the first communication apparatus, and indicating the some column vectors to the second communication apparatus by using first information as a carrier; and the second communication apparatus performing channel estimation on a first channel on the basis of the first information. Compared with performing channel estimation on a channel by means of a second communication apparatus using a complete eigenvector matrix, performing channel estimation on a channel by using some column vectors can effectively filter out noise in the channel, reduce inaccurate information in the eigenvector matrix, further improve a signal-to-interference-plus-noise ratio of the channel, and improve the channel estimation performance.

The invention provides a large language model binarization quantification method and system based on instructive alternate optimization, and relates to the technical field of large language model deploying.The method comprises the steps that a weight matrix is divided into an important area and an unimportant area by evaluating the influence degree of all weight parameters on the performance of a large language model; carrying out binaryzation on all areas of the weight matrix, and alternately optimizing a row vector scaling factor and a column vector scaling factor by adopting a first-order row-column alternate optimization iteration mode to obtain a first-order reconstruction weight matrix; and carrying out binaryzation again on the important region of the first-order reconstruction weight matrix, carrying out optimization by adopting a first-order and second-order row-column alternating optimization iteration mode to obtain a second-order reconstruction weight matrix, and correspondingly taking the second-order reconstruction weight matrix as a weight parameter after the large language model is quantized. The quantization error of the key weight parameter can be effectively reduced, the quantization precision is high, the quantization process completely depends on the internal structure information of the large language model, and the edge deployment compatibility is high.

The application relates to the technical field of electric digital data processing, and discloses a Web application automatic generation and data processing method based on electronic table mapping, which comprises the following steps: mapping electronic table cells into enumeration values to construct a type fingerprint matrix; calculating a type purity index of a column vector and identifying a vertical blocking band, and splitting the matrix into a plurality of mutually unconnected orthogonal sub-matrices; calculating a type distribution difference value of adjacent row vectors in the sub-matrices to determine a horizontal logical segmentation line, so as to define a logical data block; and assembling a Web application object model based on an arrangement mode of the logical data block. The application introduces a vertical topology decoupling mechanism based on column entropy, cuts off signal crosstalk between horizontally arranged data entities, solves the technical problem that a horizontal feature scanning in a multi-table mixed arrangement scene is easily interfered by orthogonal noise, and realizes deterministic structure reconstruction of complex two-dimensional heterogeneous data under the condition of no preset template.

This invention provides a method for solving matrix eigenvalues ​​based on a resistive variable cell array (RVA) circuit. The method includes: setting the node conductance values ​​of the RVA array; writing the initial values ​​of each element of the matrix (A-βI) into the RVA array; inputting the current signal corresponding to the random column vector b through the bit line terminal of the RVA array; simultaneously connecting the RVA unit corresponding to β to the diagonal node of the RVA array via a switch to complete the solution of the linear equation system (A-βI)x = b, and solving and storing the infinite norm of the solution vector a; connecting the RVA unit corresponding to β to a resistance control circuit via a switching circuit to update the resistance value; repeating the above process until the traversal of β is completed, the β values ​​corresponding to all extreme values ​​of the infinite norm are the total eigenvalues ​​of the matrix. This invention employs an in-memory computation method in the analog domain, which can improve computation speed while reducing energy consumption.

Systems and methods for providing an S-box. The methods comprise: implementing a plurality of layers on a field programmable gate array, wherein each of the layers comprises a substitution sublayer and a mixing sublayer. The substitution sublayer comprises a plurality of μ-boxes that are each configured to perform a different randomly chosen non-linear bijective mapping from GF(24) to GF(24). The mixing sublayer is configured to (i) compute a plurality of product terms each representing a product of an element of a maximum distance separable matrix M and an output from one of the plurality of μ-boxes, and (ii) compute a plurality of column vector elements each comprising a sum of respective ones of the plurality of product terms.

The present invention discloses a method and device for generating a statistical data matrix of user behavior, which relates to the field of computer technology. A specific implementation of the method includes: obtaining multiple single behavior data of multiple users for multiple items within a preset time interval; determining the non-zero statistical results of the target dimension data of each user for each item and the key-value pair data formed by the non-zero statistical results and the item identifier based on the single behavior data; storing the key-value pair data of the same user separately in a set, and storing the item identifier in the first subset, and storing the non-zero statistical results in the second subset; converting the item identifier in each set into the column number or row number of the matrix, so that each set becomes a row vector or column vector for forming a statistical data matrix. This implementation can quickly obtain the statistical results of user behavior data without using a structured query language, thereby generating a statistical data matrix that does not waste storage space.