8 results about "Low-rank approximation" patented technology

The invention discloses a wind turbine generator blade vibration mode feature recognition method, which mainly comprises the following steps of: arranging three-axis piezoelectric acceleration sensors at equal intervals from a blade root to a maximum chord length position, and adaptively selecting an optimal channel as a target input vibration signal according to envelope dispersion and a period proportion; introducing a vibration signal denoising method based on regenerative phase shift sine-assisted empirical mode decomposition, and constructing a Fisher ratio-based multi-dimensional fusion index to remove a noise component; estimating a system order range according to a singular entropy jump value theory, and designing modal similarity and a confidence index to accurately estimate a real order of a blade system; introducing three types of constraints of structure maintenance, modal sparsity and energy smoothness to jointly optimize a low-rank approximation strategy so as to realize optimal reconstruction of the Hankel matrix; constructing a fitness function selected by a clustering center by combining the point set density of the sample and Euclidean distance information, and optimizing a modal extraction result by adopting inter-class dispersivity and an intra-class sample number; according to the method, the environmental noise can be effectively removed, the system order can be accurately determined, and finally the modal parameters of the system can be accurately identified.

The invention provides a user portrait recommendation method based on a high-order structure and semantic enhancement, and the method specifically comprises the following steps: S1, introducing a multi-hop adjacent matrix to capture a high-order behavior pattern in an interaction graph for the interaction graph of a user and a project through a high-order structure maintenance module of user grouping, and employing a low-rank approximation and clustering method to obtain a high-order behavior pattern in the interaction graph; grouping the users based on the behavior similarity; s2, extracting a representative keyword set from items interacted by the same group of users, and obtaining group-level keywords of the users; s3, through a portrait perception recommendation module based on cross-view comparative learning, user semantic embedding and project semantic embedding based on keywords are constructed; obtaining user structure embedding and project structure embedding according to a collaborative structure between a user and a project in the interaction graph; then, alignment of project semantic embedding and project structure embedding is achieved through cross-view comparative learning; and S4, calculating a user-project combination score and generating a recommendation result. According to the invention, the recommendation performance is enhanced.

The invention relates to a compressed and quantized Internet of Vehicles privacy protection security hierarchical federated learning method, which designs a vehicle end side encryption parameter compression and integer quantization method, a vehicle side completes model training on the basis of local data, and performs flattening, low-rank approximate compression and integer quantization on parameter tensors to obtain a compressed and quantized vehicle end side encryption parameter. And the transmission quantity is obviously reduced under the condition of ensuring the model precision. The quantized parameters are subjected to end-side homomorphic encryption at a vehicle end, and a ciphertext is kept invisible in transmission and aggregation stages. And secondly, a method for setting anonymous authentication batch verification and session key negotiation security encryption between the vehicle end and the fog layer is designed, the concurrent authentication pressure is reduced through a batch verification structure, and the communication security and the access efficiency are improved. And finally, designing a hierarchical ciphertext aggregation updating and Bloom filtering revocation control method, performing homomorphic domain aggregation on ciphertext parameters uploaded by different vehicles by a fog layer, executing one-time decryption by a center end and recovering a floating point model weight, and completing global model updating.

The invention relates to an interpolation method and an interpolation model for scRNA-seq data. An interpolation method for scRNA-seq data comprises the following steps: performing dropout interpolation on the scRNA-seq data through an automatic encoder of a Markov self-attention mechanism and a self-adaptive threshold low-rank approximation method to obtain an interpolation matrix XCG and an interpolation matrix XMF; the interpolation matrix XCG and the interpolation matrix XMF are integrated through a cell-level interpolation matrix integration method, and an interpolation result is obtained; and taking the interpolation result as the input of the next iteration until the iteration condition is met, and obtaining the final interpolation result after the loop is ended. According to the interpolation method and the interpolation model for the scRNA-seq data, the global heterogeneity of the scRNA-seq data is ensured, meanwhile, the local structure of the scRNA-seq data is captured, the interpolation value is highly consistent with the real data in trend, and the optimal numerical precision is also obtained.

The application discloses a style legal consultation question and answer method and system, a storage medium and equipment, relates to the technical field of natural language processing, and the method comprises the following steps: collecting text data and audio and video data in the legal consultation field and converting the text data and audio and video data into texts; generating structured text data sets from the texts by using a base model; dividing the structured text data sets into a basic legal knowledge base and a stylized knowledge base according to the style categories of seed instructions; adopting different styles of vertical field labeled training sets, and obtaining a pre-training fine-tuning model by fine-tuning the base model based on a low-rank approximation fine-tuning method; and generating a corresponding style answer from the pre-training fine-tuning model according to a received user legal consultation question. Through the technical scheme of the application, the stylized model can be iterated more quickly, the low-delay question and answer demand is realized, the legal consultation reply can be provided according to the style preferred by the user, and the work efficiency of legal consultation is improved.

The application discloses a kind of based on low rank contrast learning and reconstruction graph anomaly detection method, comprising: 1) obtain graph data, and generate low rank graph data by SVD singular value decomposition dimension reduction, using restart random walk algorithm to carry out subgraph sampling, obtain original view and low rank view;2) on original view and low rank view, construct contrast pair and carry out contrast learning, obtain contrast learning loss;3) low rank attribute reconstruction is carried out on original view and low rank view, and the final reconstruction loss of original view and low rank view is obtained;4) the graph neural network model is trained in combination with contrast learning loss and reconstruction loss;5) according to the trained graph neural network model, the abnormality of node in the graph to be measured is judged, and the potential abnormal node is determined.The application generates low rank view by low rank approximation to node attribute and topological structure, effectively filters the interference of abnormal node and noise on the basis of retaining the original structure of graph.

The present application belongs to the technical field of block iterative matrix solution in computational electromagnetics, in particular to a block iterative matrix solution method based on matrix dimension reduction and preprocessing technology. The present application uses two technologies of matrix dimension reduction and matrix preprocessing: the right end item matrix is processed by using the matrix dimension reduction, through low rank approximation, which can effectively ensure the accuracy of the solution, significantly reduce the calculation cost, and facilitate the processing of multiple right end linear systems; and the BCOCG algorithm is applied to the preprocessing technology to further improve the matrix behavior and improve the convergence and stability of the algorithm. For the system equation generated by multiple physical excitation sources or multiple scanning angles in electromagnetic problems, the memory cost is shortened from the perspective of short recursive Krylov subspace iteration method, and the block iteration BCOCG algorithm can calculate multiple column right end item vectors at the same time, avoiding the matrix ill-conditioning problem caused by the correlation between columns and columns.

The application relates to a cluster clustering and positioning method and device based on multi-view fusion, equipment and medium. The method comprises the following steps: acquiring a sparse distance matrix among nodes in a cluster; based on the same, constructing multi-feature views such as an original distance matrix, a Gaussian similarity, node degree centrality and a Laplace feature vector; constructing a self-representation matrix according to the node similarity of each view and stacking the same into a three-order tensor; performing low-rank approximation to construct a fusion affinity matrix, and dividing the cluster into multiple subnets through spectral clustering; extracting a distance submatrix from each subnet, obtaining local relative coordinates through multidimensional scaling transformation; identifying a shared node between subnets as an anchor point, splicing each local coordinate to a global coordinate system through Procrustes analysis, and obtaining a global positioning result. By using the method, the clustering criterion and the positioning geometric configuration requirement can be kept consistent under the condition of non-full-connection sparse distance observation, so that error transmission is inhibited and the global positioning accuracy is improved.