4 results about "Structural correlation" patented technology

The invention relates to the field of financial data, and discloses a medium-and-long-term adaptive transaction signal system based on a large model, which comprises the following steps: after obtaining target signal data, inquiring a hidden variable association graph of the target signal data, analyzing a structure association element and generating an evolution path, and by analyzing a path response sequence, identifying a sequence transition node and calculating a node chaos index, obtaining the target signal data; the method comprises the following steps: performing modal decomposition on target signal data to obtain a signal sub-data cluster, performing label coding to obtain a primary function label set, querying a sparse representation coefficient, marking a weight vector, coding the weight vector to a historical memory module, dynamically adjusting a hyper-parameter combination according to a convergence state, calculating a generalization performance ratio, generating a distribution characteristic key, and finally, performing data distribution on the basis of the distribution characteristic key. And constructing a self-adaptive transaction signal mechanism according to various identification elements in the secret key. According to the invention, the transaction signal adaptation accuracy can be improved.

The present invention discloses a multi-scale feature extraction and association analysis method, which constructs a spatial multi-scale association graph system with multiple hierarchical association graphs, each association graph corresponds to a specific spatial scale, and there is a hierarchical convergence relationship between association graphs at different levels; for each node of the bottom-level association graph, its time series indicator data is extracted, and the time series indicator data is decomposed into periodic, trend and burst components; for association graph nodes other than the bottom-level, convergence features are extracted using statistical convergence and pattern convergence; based on the extracted time series indicator data and the extracted convergence features, the attribute correlation and graph structure correlation of node pairs are calculated in each level, and the two are fused to obtain the intra-level correlation index of the corresponding level; the hierarchical convergence correlation index is defined according to the intra-level correlation index to realize inter-level correlation propagation; a hierarchical attention mechanism is introduced to fuse the correlation information of different levels and calculate the spatial multi-scale correlation.

The invention discloses a multi-scale feature extraction and association analysis method, which comprises the following steps: constructing a spatial multi-scale association graph system with a plurality of hierarchical association graphs, each association graph corresponding to a specific spatial scale, and having a hierarchical convergence relationship among different hierarchical association graphs; extracting time sequence index data of each node of the bottommost association graph, and decomposing the time sequence index data into periodic, trending and sudden components; for association graph nodes not at the bottommost layer, extracting convergence features by using two modes of statistical convergence and mode convergence; based on the extracted time sequence index data and the extracted convergence features, calculating attribute correlation and graph structure correlation of the node pairs in each hierarchy, and fusing the attribute correlation and the graph structure correlation to obtain an intra-hierarchy correlation index of the corresponding hierarchy; defining hierarchical convergence correlation indexes according to the intra-layer correlation indexes to realize inter-layer correlation propagation; a hierarchical attention mechanism is introduced, correlation information of different hierarchies is fused, and spatial multi-scale correlation is obtained through calculation.

The application discloses a large language model query optimization system and method for graph data, and belongs to the technical field of large language models and is applied to graph data processing in a resource-intensive environment. The application implements the method as follows: 1. a connection relationship in a text attribute graph is constructed, and a node classification task is assigned to an unlabeled node set; 2. token pruning is performed on a query set, specifically, text insufficiency is used to evaluate the query set nodes; neighbor text of a text sufficient query is removed to reduce overall token consumption; 3. query enhancement is performed on the query set after token pruning; specifically, the query set is executed in rounds in a cyclic iteration mode, and pseudo-labels generated in a previous round are used to enhance neighbor text of a subsequent query. Compared with the prior art, the application improves the accuracy of large-scale queries of a large language model in a resource-intensive environment graph data scenario by evaluating the text insufficiency of query nodes and utilizing the structural correlation characteristics of graph data.