56 results about "Time-series segmentation" patented technology

The invention discloses a deep reinforcement learning-based network traffic optimization scheduling method suitable for periodic traffic characteristics. The method comprises the following stages: collecting historical periodic traffic data; a network state is monitored in real time through an SDN controller, and historical periodic flow data including time delay, bandwidth utilization rate and load information are acquired; data preprocessing: performing cleaning, normalization and time sequence segmentation on the collected historical traffic data; constructing an LSTM model, constructing the LSTM model by using the preprocessed historical data, and defining an input feature and an output tag; training an LSTM model, training the LSTM model by using the training set data, and generating traffic prediction data of a next period; constructing a state space, fusing the predicted traffic data with the real-time traffic data, and constructing the state space; training a DQN model, and training a deep Q network (DQN) model by using the constructed state space; network topology and link state information are obtained, the current network topology is obtained through a topology discovery module of the SDN controller, and the bandwidth utilization rate, time delay and packet loss probability information of a link are collected in real time through a flow monitoring module; large flows and small flows are distinguished; scheduling the small flow by using the ECMP; and scheduling the large flow by using the DQN model.

The invention relates to the technical field of unmanned aerial vehicles, and discloses an unmanned aerial vehicle battery endurance flight capability prediction system, which comprises a multi-dimensional data acquisition module, a feature mapping module, a prediction module, an optimization module and a feedback optimization module, and can be additionally provided with an early warning module. The multi-dimensional data acquisition module acquires battery data and cleans the battery data to generate standardized data; the feature mapping module maps the data to a feature space, and generates a feature sequence cluster containing a multi-dimensional association relationship by using a time sequence segmentation algorithm; the prediction module divides prediction intervals based on a support vector machine algorithm and extracts prediction indexes; the optimization module generates an endurance prediction strategy by predicting and optimizing the network model; and the feedback optimization module performs multi-source data fusion optimization and outputs a prediction instruction. The early warning module can associate the prediction instruction with the battery health degree, output a grading early warning signal and trigger a response mechanism.

The invention discloses a dynamic power distribution network deduction method based on graph neural network perception and a privacy protection type Transform model. The method comprises the steps of collecting source load storage historical data in a power distribution network and performing standardization processing; performing time sequence segmentation; establishing a dynamic topological structure model of the power distribution network; extracting spatial features of the dynamic topological structure of the power distribution network by using a graph neural network; the spatial features are converted into time sequence data, the time sequence data are combined with time sequence data obtained after time sequence segmentation, and a source network load storage time sequence data set is established; differential privacy processing is carried out; training a Transform model, in the training process, using a differential privacy optimization technology, adding noise to the gradient when the gradient is updated each time, and finally obtaining a trained DP-Transform model; and inputting a specific scene, and deducing the state of the power distribution network by using a DP-Transform model. According to the method, the interpretability of deducing the state of the power distribution network by using the Transform is improved, so that the Transform model can fully receive the graph structure input of the network topology structure of the power distribution network, and privacy protection on user information in the deducing process is considered.

An adaptive calibration method and system for an electric energy meter, relating to the technical field of electric energy meter calibration. The method comprises: acquiring basic data and additional data of an electric energy meter (S1); performing time series data alignment, and performing period division on an environment data set to generate a time series segmentation result (S2); acquiring device information of the electric energy meter, performing key feature analysis within a period, and establishing an interference weight factor for each feature (S3); building a performance degradation model (S4); acquiring usage environment data of the electric energy meter by means of a monitoring sensor (S5); then generating an adaptive calibration result on the basis of the performance degradation model (S6); and finally performing electric energy meter calibration management (S7). The present invention solves the technical problem in the prior art where the accuracy of electric energy metering is affected due to untimely calibration caused by an inability to monitor performance degradation of an electric energy meter in real time, thereby implementing effective calibration management for the electric energy meter, ensuring the long-term accuracy and reliability of the electric energy meter, and improving the efficiency of calibration work.

The invention discloses a wind turbine generator equipment fault diagnosis method and system based on a large model. The method comprises the following steps: initializing a feature set according to normal data; selecting a plurality of candidate features according to the mixed score of each candidate feature, and storing the candidate features in a feature set; performing sample division on normal data by using a time sequence segmentation algorithm and constructing multi-dimensional spatial-temporal feature representation; constructing a fault diagnosis model: loading a large language model as an infrastructure, and injecting the multi-dimensional spatio-temporal feature representation into an embedded input layer of the large language model; an adaptive pooling layer is accessed after the output of the large language model, and a double-layer MLP classifier is constructed for realizing the identification and classification of different fault types; the first layer of the double-layer MLP classifier compresses an input feature to half of an original feature dimension and applies GELU activation, and the second layer of the double-layer MLP classifier is mapped to a corresponding fault category space; constructing a knowledge mechanism library, providing prior knowledge of the wind turbine generator for the model, designing a loss function driven by the knowledge of the wind turbine generator, and carrying out model training.

The invention discloses a multi-scale time-frequency network sleep stage classification method based on single-channel electroencephalogram, and belongs to the technical field of medical signal processing. Aiming at the problems of capturing multi-scale time-frequency characteristics, processing variability among subjects and modeling long-range time dependence of a sleep staging method, the method comprises the following steps: firstly, acquiring electroencephalogram signal data, and performing time sequence segmentation, sleep stage category label determination and standardized preprocessing by taking a single-channel electroencephalogram signal as an analysis object; performing data set division by adopting nested cross validation, and constructing a multi-scale time-frequency network model; the model comprises a feature extraction module and a sequence learning module, wherein the feature extraction module comprises a time domain branch and a frequency domain branch; a subject adaptive feature calibration module is proposed to dynamically compensate the influence brought by individual difference and signal quality fluctuation; respectively training a feature extraction module and a sequence learning module by adopting a component type training strategy; and inputting to-be-classified electroencephalogram signal data into the trained model, and outputting a corresponding sleep stage classification result.

The invention provides a multi-parameter adaptive dynamic multilevel decision converter transformer fault diagnosis method and system, and the method comprises the steps: constructing a multi-parameter fuzzy feature space based on temperature, voiceprint, current and oil soluble gas parameter data, carrying out the statistics of fuzzy features through sliding time sequence segmentation, generating a multi-parameter fuzzy time sequence feature vector, and carrying out the fault diagnosis of a converter transformer. Obtaining multi-parameter internal cross characteristics and multi-parameter external cross characteristics, and performing splicing fusion to form multi-parameter correlation characteristics; constructing a multi-stage aggregation classifier, dynamically distributing a weight for each classifier based on verification set performance, fusing probability output fault types of each layer, and adjusting the matching degree of a model prediction probability and a real fault occurrence probability; and dynamically adjusting the depth of the hierarchy according to the matching degree adjustment condition, if the improvement of the fault diagnosis precision of continuous multiple layers is smaller than a set threshold value, terminating the expansion and retaining the current hierarchy, and executing multi-level processing on a prediction result. According to the invention, various parameter information is effectively fused, and the prediction accuracy and reliability are improved.

The invention relates to the technical field of fault diagnosis, in particular to a circuit breaker mechanical fault pre-diagnosis method based on time sequence analysis, which comprises the following steps of: firstly, acquiring mechanical operation data of a circuit breaker in real time by installing a sensor, and preprocessing the data; then, on the basis of standard operation data, a time sequence segmentation strategy is adopted, the data is segmented into independent time sequence sequences, time domain features and frequency domain features are extracted for each independent time sequence sequence, and a time sequence dependence mode is extracted in combination with an autocorrelation function and a cross-correlation function; fusing the time domain feature, the frequency domain feature and the time sequence dependency mode through a time frequency-time sequence collaborative modeling framework; and finally, on the basis of the time sequence features, a federated Bayesian unit is constructed for fault diagnosis, local units provide fault type and confidence prediction, and a global unit integrates models of a plurality of local units through a federated learning framework. According to the invention, circuit breaker mechanical fault pre-diagnosis is realized through time sequence analysis.

The invention relates to the technical field of geological prospecting, in particular to a geological prospecting drilling rig drilling real-time depth measuring system, which comprises a vibration acquisition module, a data processing module and a data processing module, the original vibration signals are divided into 256 time sequence segments, sampling interval values are set, the amplitude value and the phase value of each time sequence segment are calculated, and drilling dynamic characteristics are generated. According to the method, vibration sensors are arranged on the surface of a drilling tool to collect original vibration signals, the signals are segmented to calculate amplitude values and phase values, drilling dynamic characteristics are obtained, difference values of adjacent time sequence segments are calculated and compressed, Fourier transform is carried out to generate frequency spectrum parameters, and the frequency spectrum parameters are projected to Grassmann space to calculate characteristic vectors and characteristic values. And constructing a tangent plane projection function to calculate a geodesic distance to obtain a correction coordinate, and iteratively calculating a depth value based on a Riccati equation. And meanwhile, time sequence segmented compression is adopted, so that data redundancy is reduced, and the utilization efficiency of computing resources is optimized.

The invention discloses a time sequence prediction method and system based on adaptive period segmentation and parallel decoding, and the method comprises the following steps: judging the period length of time sequence data, segmenting the time sequence data according to the period length, and obtaining time sequence segmentation data; adaptively adjusting the hiding dimension of the embedded layer according to the period length, and mapping the time sequence segmentation data into the hiding dimension to obtain a time sequence representation; the obtained time sequence representation is input into a Transform encoder to carry out feature extraction; the decoder input is initialized according to the Transform encoder output, the initialized decoder input and the encoder output are subjected to feature interaction in the decoder based on a cross attention mechanism, and finally prediction output is obtained. According to the method, a more flexible period segmentation scheme and a decoding mechanism are adopted, the reasoning efficiency is improved, accumulative errors are reduced, and the time sequence basic model can achieve a better prediction effect with fewer parameters.

The invention provides a water level prediction method of a TFT model based on hyper-parameter optimization and a related device, and belongs to the technical field of water level prediction. According to the method, all data obtained in real time are preprocessed, noise and abnormal data are removed, and all preprocessed data are obtained; performing standardization processing on all the preprocessed data to obtain standardized data; performing time sequence segmentation on the standardized data to obtain a plurality of pieces of time window data; constructing a training set by using a plurality of time window data; constructing a TFT (thin film transistor) model, and optimizing hyper-parameters of the constructed TFT model to obtain an optimized TFT model; training the optimized TFT model according to the training set to obtain a trained TFT model; and inputting the standardized data into the trained TFT model for future water level prediction to obtain a future water level prediction result. The problem that the accuracy of water level prediction is not high is solved.

The invention relates to a fire-fighting water supply system dynamic equipment fault diagnosis method and system, processing equipment and a storage medium. The method comprises the following steps: acquiring a time sequence of operation of fire-fighting water supply system dynamic equipment; dividing the obtained time sequence into a plurality of non-overlapping subsequences, and dividing the non-overlapping subsequences into different instances; encoding and embedding the divided instances into a feature space to obtain a time sequence vector, and performing feature level comparison and instance level comparison; randomly selecting a text prototype, and aligning and comparing the text prototype to form a new feature matrix; generating a corresponding soft prompt; splicing the soft prompt corresponding to the time sequence and the new feature matrix, and inputting the spliced soft prompt and new feature matrix into a trained fire-fighting water supply system dynamic equipment fault diagnosis large model to obtain a time sequence with the length reaching the expected length; the fire-fighting water supply system dynamic equipment fault diagnosis method can be widely applied to the fields of artificial intelligence and big data application.

The present invention relates to the field of video recognition technology, specifically a method and system for video content structured decomposition and analysis based on time segmentation. First, the video and metadata are acquired, and a first artificial intelligence model integrating a spatiotemporal attention mechanism and a sliding window mechanism is used to perform time segmentation to generate time segments. The video is cut into several video segments using the FFmpeg tool, and a second artificial intelligence model integrating multimodal fusion and timestamp embedding is used to extract text descriptions of each segment to achieve semantic alignment of actions and texts. At the same time, FFmpeg is used to extract key frames to generate picture groups, and a third artificial intelligence model combining target detection and cross-frame consistency constraints performs semantic analysis to generate action labels. Finally, the time segments, key frames, text descriptions, and action labels are integrated to generate a complete structured output. The present invention can improve the recognition accuracy of video content.

The invention provides a real-time customer portrait dynamic update prediction method and system for a retail scene, and relates to the technical field of data analysis, and the method comprises the steps: determining a customer portrait feature set, calculating the information gain values of features under different time windows, and generating a weighted feature vector; performing time sequence division by adopting a sliding time window method; calculating a dynamic correlation coefficient, and generating customer behavior fusion features with a time sequence dependency relationship; calculating a feature relevancy matrix, and predicting customer consumption behaviors; and the customer portrait features are dynamically updated, and personalized recommendation is realized. According to the method, the real-time performance and the accuracy of the customer portrait are improved, and the effectiveness of personalized recommendation is enhanced.

The present application discloses a preparation method of an explosion-proof and protective composite material. By obtaining the real-time hot pressing temperature collected by a temperature sensor and using a data processing and analysis algorithm based on deep learning, the real-time hot pressing temperature is subjected to time series segmentation and local time series correlation, so as to automatically obtain a recommended value of the pressing pressure at the next time point according to the aggregation representation characteristics of each real-time hot pressing temperature in the local time series. In this way, by monitoring and analyzing the temperature change in the hot pressing process in real time, the pressing pressure can be dynamically adjusted according to these subtle time series changes to ensure the molding quality of the composite material, improve the consistency of product quality, and thus enhance the degree of automation in the hot pressing process of the material.

The invention discloses a rock climbing action analysis system and method based on multi-dimensional quantitative evaluation, and relates to the technical field of computer vision, and the method comprises the steps: carrying out the spatial registration and skeleton node positioning of a multi-mode rock climbing data set through employing an ICP algorithm in combination with an OpenPose algorithm, forming a skeleton motion state matrix, and carrying out the calculation of the skeleton motion state matrix; inputting the skeleton motion state matrix into a skeleton motion multi-dimensional quantitative model, performing time sequence segmentation and dynamic behavior classification by a time sequence analysis layer, performing multi-dimensional index calculation and weight distribution by an action evaluation layer, outputting a rock climbing action comprehensive score, performing differential operation on the rock climbing action comprehensive score and a standard action score, and obtaining a quantitative offset tensor. According to the invention, through the ICP algorithm, the OpenPose algorithm and the skeleton motion multi-dimensional quantitative model, the spatial registration precision of the multi-modal rock climbing data is improved, and fine dynamic classification and multi-dimensional index evaluation of rock climbing actions are realized.

The application provides an intelligent operation and maintenance learning method and system related to monitoring a baseline, and relates to the technical field of computer operation and maintenance. The method comprises: collecting a plurality of core index monitoring data of a target server cluster in a set historical period to obtain a historical monitoring data set; performing time series segmentation processing on the historical monitoring data set to obtain a time series data segment set; performing trend fitting on the time series data segment set by a linear regression algorithm to obtain a corresponding trend fitting straight line; determining a center data point based on the trend fitting straight line, and determining a bidirectional data fluctuation interval according to a residual error analysis result at the center data point. The application realizes dynamic matching of the baseline with the business, full-process automation, precision improvement, and reduction of false and missed reports.

The application relates to the technical field of fault diagnosis, in particular to a circuit breaker mechanical fault pre-diagnosis method based on time series analysis. First, the operation data of the circuit breaker mechanical are collected in real time through the installation of sensors, and the data are preprocessed; then, the data are divided into independent time series sequences based on standard operation data by adopting a time series segmentation strategy, time domain features and frequency domain features are extracted for each independent time series sequence, and time series dependent patterns are extracted in combination with autocorrelation functions and cross-correlation functions; the time domain features, the frequency domain features and the time series dependent patterns are fused through a time-frequency-time series collaborative modeling framework; finally, based on the time series features, a federal Bayesian unit is constructed for fault diagnosis, wherein a local unit provides fault type and confidence prediction, and a global unit integrates the models of multiple local units through a federal learning framework. The application realizes circuit breaker mechanical fault pre-diagnosis through time series analysis.

This invention belongs to the field of artificial intelligence technology and discloses a traffic flow prediction method, system, and device based on intra-block convolution and axis remapping. The invention divides the input traffic flow time series into multiple data blocks; uses a temporal convolutional network as the backbone to independently extract local temporal features within each data block; transforms the dimension representing the data block index into a channel dimension through axis remapping, thereby reorganizing the discrete block features into a structured multi-channel feature tensor; maps the tensor to the target dimension through a projection head network to obtain an encoded representation; and inputs the encoded representation into a prediction decoder to generate a prediction result for future traffic flow. This invention integrates the local temporal capture capability of TCN with the strong representation capability of modern sequence models, achieving both high computational efficiency and high accuracy.

The invention relates to a method for extracting the power consumption of an individual device within a set of devices (E1, E2, E3,..., En), the method comprising transforming a stream of measurements of an overall power consumption into a time series, and dividing (S1) the time series into a set of sequences, and then, for each sequence, the steps of: - generating (S4) a positional encoding matrix (MEP); - extracting (S5) features from the sequence to form a feature matrix (MF); - projecting (S6) statistical metrics into a statistics vector (MS); - concatenating (S7) these various elements (MEP, MF, MS) to form an input matrix (ME) provided to a transformation neural network designed to infer (S8) a time sequence corresponding to the individual device.

The invention discloses a zero sample fault detection method based on weighted semantic consistency embedding, and belongs to the field of industrial fault diagnosis. The method comprises the following steps that firstly, industrial equipment is used for collecting multi-mode sensor data, and standardized data are obtained through normalization, wavelet denoising and time sequence segmentation preprocessing; determining a fault attribute dimension based on domain expert knowledge, assigning each fault class to form a semantic attribute vector, and constructing a semantic attribute weight matrix through mutual information gain between attributes and fault tags; embedding data and semantic attributes into a shared space by using a data encoder and a semantic attribute encoder, decoding and calculating reconstruction loss by combining a data decoder and a semantic attribute decoder, and introducing weight to construct weighted semantic consistency loss; combining the two types of loss to train an encoder and a decoder, and extracting training data features by using the trained encoder and training a multi-attribute classifier; and finally, in a zero sample scene, extracting embedded features without fault data, inputting the embedded features into an attribute classifier, and matching fault categories through a maximum posterior probability. According to the method, the equal-weight hypothesis defect of an existing semantic consistency embedding method is broken through, key semantics are strengthened through mutual information gain weights, consistency loss is weighted to resist noise interference, and the precision, robustness and generalization ability of zero sample fault diagnosis are remarkably improved.

The invention relates to the related technical field of application performance monitoring, in particular to a low-Token consumption time sequence analysis method based on time sequence data visualization and a multi-modal large model, which comprises the following steps: step S1, cleaning, normalizing and feature extraction are performed on original time sequence data, and features comprise trend features, periodic features and abnormal features; s2, a customized time sequence trend chart is generated according to the features extracted in the step S1, the trend chart comprises a customized coordinate axis, a multivariable visual distinguishing identifier, a feature highlight identifier and a long time sequence segmentation identifier, and the method relates to a long time sequence data efficient analysis technology under the scenes of industrial equipment operation and maintenance, financial risk control, urban traffic scheduling and the like. By customizing the time sequence visual coding compression data Token amount and combining with the time sequence enhancement fine tuning optimization multi-modal model, low-cost and high-precision analysis of long time sequence data is realized, and technical support is provided for time sequence abnormity early warning, trend prediction and correlation diagnosis.