320 results about "Series data" patented technology

The invention belongs to the technical field of artificial intelligence, and particularly relates to a wind power prediction method based on multi-objective optimization, and the method comprises the steps: wind power data collection and training data set construction, adaptive sliding window and dynamic fluctuation decomposition of wind power time series data, construction of a short-term power prediction model, and short-term power prediction. According to the method, the historical window length and the decomposition scale can be autonomously adjusted according to the inherent fluctuation characteristics of the data, and multi-scale accurate characterization of the non-stationary power sequence is realized; according to the method, a dynamic space-time diagram fusing geographic distance and instantaneous power correlation is constructed, and a diagram attention network combined with trend similarity gating is designed, so that dynamic refined modeling of a space incidence relation is realized; according to fluctuation intensity self-adaptive loss function dynamic balance point prediction precision and interval prediction reliability, synchronously outputting deterministic and probabilistic prediction results; the rated power limit and the ramp rate constraint are embedded into the model in a soft mode, and it is ensured that the prediction result conforms to the actual operation rule of the wind turbine generator.

A time series data prediction method and apparatus, and a storage medium are provided. The method includes: obtaining current time series data collected in a current time window that is adjacent to and precedes a prediction time window in a current time period, and obtaining a plurality of groups of historical time series data separately collected in a same target time window of a plurality of historical time periods; encoding the plurality of groups of historical time series data by using a plurality of encoders respectively, to obtain a plurality of historical time series features, where each historical time series feature represents relative location information and change trend information of each group of historical time series data in the target time window; and determining, predicted time series data corresponding to a target object in the prediction time window.

The invention discloses an electric power system data anomaly prediction method based on deep learning, and relates to the field of data anomaly prediction.The method comprises the steps that firstly, a graph convolutional network and LSTM are used for deeply mining time series data of an electric power system and spatial-temporal characteristics of a topological structure, and high-precision anomaly detection and positioning are achieved; and when an exception is detected, an exception attribution mechanism is introduced, and exception signal features (such as exception fragments and positioning information) at a mathematical level are converted into a structured query object at a semantic level. And then, a barrier between numerical data and the unstructured operation and maintenance knowledge base is broken through a vector retrieval technology, and related maintenance regulations and historical cases are accurately recalled. And finally, an executable recommended disposal scheme is automatically generated based on recall knowledge in combination with the generation capability of a large language model, so that an intelligent closed loop from anomaly perception and knowledge matching to decision assistance is constructed, and the accuracy and efficiency of power system fault handling are remarkably improved.

The invention discloses a weighing sensor life degradation modeling method based on a bimodal LSTM network, and belongs to the technical field of industrial automation control systems, and the method comprises the steps: obtaining the data of a weighing sensor, generating a bimodal original data set, and carrying out the denoising processing to obtain a bimodal denoised data set; the method comprises the following steps: constructing a dual-channel LSTM-DC-CNN network, fusing stress time sequence data and environment parameter data, performing time sequence feature extraction to obtain a stress degradation feature vector and an environment coupling coefficient vector to generate a fusion degradation feature vector, performing feature enhancement processing on the fusion degradation feature vector, and performing path optimization to generate a life degradation stage judgment result. And a preset historical degradation database is combined to generate a remaining service life index of the weighing sensor. According to the method, two-channel time sequence feature extraction, a dynamic fusion mechanism and an empirical mode compensation technology are adopted, the prediction precision and reliability can be remarkably improved, and efficient and intelligent evaluation of the remaining service life is achieved.

The invention discloses a main beam formwork erection elevation adjustment method based on machine learning, and belongs to the technical field of elevation adjustment. The method comprises the steps that a space-time database is constructed by collecting data of the construction period of a cable-free section; processing the time sequence data by adopting an LSTM-Transform hybrid model, extracting long-period characteristics, and outputting an initial prediction value of the elevation; analyzing nonlinear variables such as a cable force change rate, a sunlight gradient and a material age by using an XG-Boost algorithm, and outputting a compensation factor; fusing and generating a joint prediction value; constructing a reinforcement learning agent, and outputting an adjustment instruction by taking a construction stage as a state space, taking an elevation adjustment amount as an action space, minimizing deviation between a predicted value and a measured value and taking construction stability as a reward function; the formwork erecting elevation is adjusted according to the driving hydraulic system, and the database is updated in real time. According to the invention, through multi-dimensional data fusion and intelligent optimization, the elevation adjustment precision and stability are improved.

The invention discloses a passenger flow multi-source data statistical method for a bus doorless signal, which relates to the technical field of data statistics, and comprises the following steps: based on multi-source time sequence data of a target vehicle in an operation plan time window, identifying a plurality of candidate boundary judgment events, and endowing each event with a quantitative confidence evaluation value; aiming at the current shift, independently generating an ordered candidate event list for the first station operation boundary and the last station operation boundary respectively based on the confidence evaluation values, and arranging events in the list according to the high-low sequence of the confidence evaluation values of the events; constructing an optimized objective function by taking elimination of passenger flow data string between the first station and the last station in a non-operation period as a target; according to the method, the boundary event is identified in a multi-source data fusion mode, and the operation boundary can be accurately positioned even if no door signal exists, so that the problem of data cross in a non-operation period is finally eliminated.

The invention provides a power load prediction method and device, equipment and a medium. Relates to the technical field of power load prediction. The method comprises the following steps: respectively mapping historical time sequence data from different data sources of a to-be-predicted power system into single-source time sequence characteristics representing overall time sequence change characteristics of the corresponding data sources, and modeling a dependency relationship of the single-source time sequence characteristics corresponding to the data sources, obtaining time sequence embedding representing overall dynamic evolution of the to-be-predicted power system; historical time sequence data from different data sources are converted into natural language texts expressing data source attributes, time sequence characteristics and power load associated information, the natural language texts corresponding to the data sources are coded, and prompt embedding including semantic information of the data sources is obtained; fusing the time sequence embedding and the prompt embedding to obtain cross-modal enhanced embedding; and decoding the cross-modal enhanced embedding to obtain a power load prediction result of the to-be-predicted power system.

The invention discloses an open caisson sinking posture prediction method, which belongs to the technical field of open caisson sinking prediction, and comprises the following steps: constructing a primary data matrix according to on-site actual measurement time sequence data; smoothing the first-level data matrix by using a Savitz-Gauge filtering algorithm to obtain a second-level data matrix; constructing a third-level data matrix in combination with the second-level data matrix and the construction excavation information representation matrix; performing correlation analysis on the three-level data matrix by using three analysis methods to obtain a correlation characterization matrix; screening key variables by using the correlation characterization matrix to construct a four-level data matrix; performing dimension reduction processing on the four-level data matrix by using a principal component analysis method to obtain a principal component matrix; and training a long-short-term memory network by using the principal component matrix to obtain an open caisson sinking posture prediction model to perform open caisson sinking posture prediction so as to obtain an open caisson sinking posture prediction result. According to the method, the accuracy and stability of deep water open caisson sinking posture prediction can be improved.

The invention discloses a vehicle lane changing intention prediction method based on comparative learning under a data missing condition, and the method comprises the steps: building a self-supervised comparative learning framework, employing a twin network structure for an upstream task, and employing a graph attention network module for a downstream task; by constructing positive and negative sample pairs, data representation is learned, so that the positive sample pairs are closer in a feature space, and the negative sample pairs are farther; and respectively processing a complete data sample and a missing data sample by adopting a dual-channel trend attention mechanism. The mechanism can effectively capture the global trend and local change in the time sequence data, and improves the accuracy and reliability of feature extraction. According to the method, the prediction robustness is remarkably improved, the complex operation of traditional data filling is avoided, time sequence dynamic and space interaction can be fused, efficient space-time feature modeling is achieved, and the problems of data missing and space-time interaction modeling confronted by vehicle lane changing intention prediction in an intelligent driving scene are effectively solved.

The invention discloses a financial flow situation awareness system and method based on a multi-modal large model, and relates to the technical field of supply chain financial risk control. Aiming at the problems of data isolation, high processing delay and risk perception lagging in the existing storage pledge financing scene, the method comprises the following steps: obtaining a storage digital planar graph, dividing logic sub-regions, collecting and processing multi-source data of each sub-region in parallel, and extracting pledge stock time sequence data and warehouse receipt information by utilizing a visual and text model respectively; constructing a sub-region association map by taking the cargo identifier as a node and fusing the fund flow; and fusing all the sub-maps to construct a global association map, performing cross-regional consistency verification and anomaly recognition, generating a dynamic risk judgment result, and outputting hierarchical situation awareness information. According to the invention, accurate association and panoramic risk perception of the "object-bill-money" state are realized, and the real-time performance and accuracy of risk identification and the expandability of the system are improved.

The invention relates to the technical field of industrial automation and energy management, in particular to a smoke waste heat utilization monitoring system based on big data. Comprising a multi-dimensional time sequence data acquisition and fusion module used for generating a unified data set after cleaning and time alignment processing; the heat source-load prediction module based on time sequence deep learning is used for generating a predicted supply and demand curve in a future scheduling period; the multi-target dynamic optimization and scheduling decision module is used for generating an optimal heat energy distribution scheduling scheme; and the real-time feedback and self-adaptive control module takes the optimal heat energy distribution scheduling scheme as a dynamic set value of the control system, and returns the actual distribution power measured and calculated in real time to the multi-dimensional time sequence data acquisition and fusion module for periodically retraining the attention mechanism long and short term memory network model. According to the system, the precision of supply and demand matching is remarkably improved, and the problem of supply and demand mismatching caused by untimely response in a traditional mode is effectively solved.

The invention relates to an automatic scheduling method based on load and distributed energy fluctuation, and the method comprises the following steps: data acquisition: obtaining load prediction time sequence data and distributed energy power generation prediction time sequence data, and generating an initial scheduling plan based on the load prediction time sequence data and the distributed energy power generation prediction time sequence data, the initial scheduling plan comprises a conventional unit output plan and an energy storage charging and discharging plan; data acquisition: acquiring load actual power time sequence data and distributed energy actual power time sequence data in real time; fluctuation decomposition: performing adaptive variational mode decomposition on the actual power time sequence data of the load and the actual power time sequence data of the distributed energy, and extracting a load fluctuation component and a distributed energy fluctuation component; calculating an index according to a load fluctuation component and a distributed energy fluctuation component; according to the invention, fluctuation dynamic scheduling can be accurately quantified to improve the stability and economy of the power grid.

The invention discloses a fault diagnosis and health management prediction method based on a laser gyroscope, and belongs to the technical field of inertial navigation equipment health management. The method comprises the following steps: acquiring and preprocessing a time sequence data set of operation core parameters of the laser gyroscope; performing wavelet packet decomposition on the time sequence data to extract high-frequency fault features, extracting time sequence statistical features by a sliding window method, and combining principal component analysis to perform dimensionality reduction and fusion to obtain a multi-dimensional feature vector; inputting the feature vectors into an improved kernel extreme learning machine model, optimizing parameters through a particle swarm optimization algorithm, and then outputting fault types and grades; on the basis of a fault diagnosis result, an analytic hierarchy process is adopted to endow parameter weights, and a health factor calculation model is constructed to quantify a health state; taking the health factor sequential sequence and the key parameter degradation trend as input, and combining a bidirectional long-short-term memory neural network with health factor sequence constraint to predict residual life and a confidence interval; new data are regularly brought in, and model parameters are updated through transfer learning to realize dynamic iteration. The method solves the problems of early fault recognition lag and low life prediction precision of a traditional method, is suitable for the fields of aerospace, precision navigation and the like, and has remarkable engineering application value.

This application relates to a real-time power supply and demand forecasting method and system based on a cloud-native architecture. The method includes: deploying a data access service in a cloud-native cluster using containerized microservices to receive real-time supply and demand data streams and historical time-series data from a power trading system; writing the data streams to distributed storage and pushing them to the forecasting pipeline via a message queue; performing timestamp alignment, missing data handling, normalization, and smoothing / denoising on the supply and demand data by a preprocessing service to obtain a low-noise supply and demand sequence; updating model parameters in a rolling window by an ARIMA forecasting service and outputting linear forecast values ​​as the first forecast result; calculating the forecast residuals based on the first forecast result and the actual observations, constructing residual time-series samples, and outputting residual forecast values ​​by an LSTM forecasting service; and superimposing the first forecast result and the residual forecast values ​​by a fusion service to obtain the real-time supply and demand forecast result and publishing it to the real-time trading business interface.

The invention discloses an optical scanning radiometer in-orbit dark background long time series data extraction method and system, and the method comprises the steps: obtaining cold air observation data and solar zenith angle data from an optical scanning radiometer, and generating a channel-level cold air observation data set and a channel-level solar zenith angle data set. According to the channel-level cold air observation data set and the channel-level solar zenith angle data set, various screening methods are adopted to generate to-be-selected schemes and screening data sets of the to-be-selected schemes. For each to-be-selected scheme, calculating a statistical mean value and data volatility based on a screening data set of the to-be-selected scheme, calculating a dark background representative value of the to-be-selected scheme, and determining a mean value sorting sequence number and a volatility sorting sequence number; and determining an optimal dark background value scheme by adopting a weighted assignment method according to the mean sorting sequence number and the volatility sorting sequence number. Therefore, according to the in-orbit dark background long-time-sequence data extraction method of the optical scanning radiometer, long-time-sequence stable extraction is achieved, and the remote sensing data radiometric calibration precision is remarkably improved.

The invention provides a supplier grading and classifying method based on multi-dimensional feature fusion, and relates to the technical field of supply chain management. The method comprises the steps of obtaining structured indexes and unstructured text data of suppliers; structural features are extracted through a multi-layer perceptron, and semantic features are extracted through Transform; a cross-modal attention mechanism is adopted, and deep fusion is carried out with the structured features as queries and the semantic features as key values to obtain interaction features; performing dynamic modeling on the historical time sequence data through bidirectional LSTM to generate time sequence comprehensive features; inputting a multi-task hierarchical classifier, combining ordinal regression and contrast learning loss optimization, and outputting categories and grades; according to the management priority, the weight is dynamically adjusted through attention bias, and grading flexible adaptation is achieved; and incremental learning fine tuning is carried out by using actual service data through closed-loop feedback. The method solves the problems that multi-source heterogeneous features are difficult to fuse, and feature space semantic distinguishing and grading standards lack dynamic adaptation, and is applied to intelligent evaluation of industrial internet suppliers.

This invention discloses a real-time processing and database entry method for tunnel boring machine (TBM) construction time-series data based on integrated batch processing. The method includes: acquiring the original time-series data of the TBM and its associated original ring numbers; performing standardization processing based on a pre-built semantic rule base to obtain standardized records; identifying the TBM's operating status and determining the target ring number based on the standardized records and a judgment threshold; performing differentiated data cleaning strategies on the standardized records according to the operating status and evaluating the cleaning results to obtain a comprehensive quality score; and routing the standardized records to matching processing channels based on dynamic diversion conditions including the comprehensive quality score and arrival delay, and writing them into the time-series database using the target ring number as a reference. This invention overcomes the defects of missed physical anomaly detection and reference misalignment, improving the operational condition reproduction accuracy of the entered data.

According to various embodiments, there is provided an operation method of an electronic device including: acquiring a plurality of images of a user, captured by an RGB camera of the electronic device, acquiring a specific area on a specific body part of the user from the plurality of images, generating a plurality of first data associated with an RGB color model on the specific area associated with the plurality of images, generating a plurality of second data associated with a YCrCb color model based on the plurality of first data, generating first time-series data associated with a green channel based on the plurality of first data; generating third time-series data based on combining the first time-series data and the second time-series data; and estimating a pulse of the user based on conversion of the third time-series data into a frequency domain.

To efficiently collect teacher data for generating a learned model for predicting a behavior of a control target that may change along a time axis.SOLUTION: Acquiring a trained model for predicting output time-series data indicating a temporal change of an output parameter of a control target from input time-series data indicating a temporal change of one or more input parameters of the control target, acquiring a plurality of pieces of pattern data in which temporal changes of the one or more input parameters are different from each other, repeating a prediction process for predicting the output time-series data from the pattern data using the trained model for each of the plurality of pieces of pattern data while changing one or more invalidation nodes in the trained model, and calculating a variation of the output time-series data for each of the plurality of pieces of pattern data; Based on these variations, one or more pieces of pattern data are selected for updating the learned model.SELECTED DRAWING: Figure 1

This invention discloses a method and apparatus for accurate bridge damage identification and location based on multi-source data fusion. The method includes: performing spatiotemporal registration and data fusion on the time-series data of the bridge's global planar deformation obtained by ground-based synthetic aperture radar and the time-series data of the bridge's point-like three-dimensional deformation obtained by the BeiDou satellite navigation system to obtain a fused deformation field of the entire bridge area; calculating multiple damage characteristic indicators reflecting different deformation physical mechanisms of the bridge based on the fused deformation field, and determining the corresponding preliminary damage level based on the damage judgment threshold of each damage characteristic indicator; using the preliminary damage levels of each damage characteristic indicator as independent evidence sources, performing evidence fusion to resolve indicator conflicts, and determining the bridge damage location and damage level based on the fusion result. This method can effectively improve the accuracy and reliability of bridge damage identification and location, providing a scientific basis for bridge operation and maintenance decisions.

The invention discloses a time series data analysis method and device, a storage medium and electronic equipment, and relates to the technical field of the Internet of Things. By establishing a composite index structure combining a time index and an equipment index, under the composite index structure, corresponding index entries are dynamically created for a target data flow, the framework of the composite index structure is not changed, and the time series data analysis efficiency is improved. The content is dynamically filled along with the data, and finally a complete composite index result capable of being efficiently retrieved is formed. And the time index adopts a B-Tree structure, so that data in a specific time range can be quickly positioned. Equipment indexing is realized based on a hash table, data generated by specific equipment can be efficiently retrieved, the throughput bottleneck of mass data storage is broken through, and second-level writing and aggregation query of ten millions of data are realized. And in combination with artificial intelligence AI, distributed time series data management and an Internet of Things multi-source data fusion technology, in the face of massive time series data, the analysis processing capability of the time series data is improved to meet the demand of time series data analysis.

The embodiment of the invention provides an electricity price prediction method and system based on LSTM and Transform, and belongs to the technical field of electricity price prediction. The electricity price prediction method comprises the following steps: acquiring multi-source historical data of a power system; performing type integration division on the multi-source historical data to obtain cost time sequence data, meteorological time sequence data and economic index time sequence data; preprocessing the cost time series data, the meteorological time series data and the economic index time series data; constructing a training set according to any two items in the cost time sequence data, the meteorological time sequence data and the economic index time sequence data to obtain three groups of cross training sets; a mode of constructing a training set by crossing multiple kinds of time series data and acquiring a real-time prediction sensitivity coefficient of each group of models is adopted, so that the influence weight of the time series data / models with a large electricity price influence degree can be effectively enhanced, and the prediction precision is improved.

This application provides a method and system for monitoring the operational status of reservoirs based on deep learning, relating to the field of water conservancy project operation monitoring. The method includes: collecting multi-source time-series data on reservoir operation, including hydrological time-series data, engineering time-series data, and environmental time-series data; constructing a multi-source feature sequence of the reservoir's operational status based on the multi-source time-series data; inferring the multi-source feature sequence using a deep learning model to obtain reservoir operational status assessment results and trend prediction results; the deep learning model consists of convolutional neural network branches, recurrent neural network branches, and attention mechanism branches connected in parallel; and determining the reservoir's operational status based on the reservoir operational status assessment results and trend prediction results. This application, used in reservoir operational status monitoring, solves the technical problem of low accuracy in existing reservoir operational status monitoring methods.

This application relates to a feature extraction apparatus, device, and storage medium for time series data, belonging to the field of artificial intelligence technology. The method includes: acquiring a time-domain signal corresponding to a user's time-series data, wherein the time-series data represents the user's economic behavior data; performing N-level frequency domain decomposition on the time-domain signal using an N-level filter bank to obtain multiple time-series sub-band signals of different frequency bands, wherein N is a positive integer greater than or equal to 1; and performing feature extraction on each of the time-series sub-band signals to obtain the target features of the time-series data. This application can acquire the performance of time-series data in different frequency bands and can extract more features from the time-series data.

This invention belongs to the field of intelligent transportation, specifically relating to a MaaS-IoT integrated method and system for crowd risk perception and intervention. The method first collects environmental data and travel orders for the target area through IoT devices and an online platform; then, it constructs a path decision model based on the Bandit algorithm; combining shared and unique feature vectors to achieve path recommendation. Next, it constructs a crowd risk event dataset based on historical data and generates an easily searchable vector database, and fine-tunes the LLM to obtain a crowd risk inference model. Then, it tokenizes the time-series data, and uses the crowd risk inference model to perform chain reasoning and Monte Carlo-Shapley attribution to generate corresponding results; finally, it generates intervention instructions based on the outputs of the two models, and issues them for execution after verification. This invention solves the problem that existing methods struggle to achieve timely risk perception and efficient intervention in public places where crowds gather.

Provided are a method and system for cleaning time-series data for power generation applications, a medium, and a device. The method comprises: traversing a fundamental time-series data set for power generation applications to construct a feature vector of the data set; performing outlier detection on the data feature vector and marking an outlier as a missing value; and on the basis of a generative adversarial network, learning the distribution of the time-series data set, and replacing missing data with generated time-series data to obtain a complete time-series data set.

The invention discloses a central heating and refrigerating system power adjusting method, terminal equipment and a storage medium. All available time data are mapped into a hidden space. Data points located on the manifold between the points associated with the available training time data are selected so that new data can be generated that complies with a new time dependency. And the enhanced data is input into the graph structure network, so that the model prediction performance is improved. And lagging relationship representation is obtained by performing lagging relationship calculation on the time series data, future values of the time series data are predicted in an auxiliary manner, and the capturing capability of the model on the large inertia characteristic of the system is enhanced. The improved ST-GNN prediction model and an MPC optimization controller are deeply integrated to form a complete closed loop from perception, prediction, optimization, execution and feedback. A model online updating mechanism is introduced, when the prediction deviation is increased due to the change of system characteristics, the model can be quickly adjusted based on a small amount of new data, and the model performance is prevented from attenuating along with time.

The purpose of this invention is to address the issue of accuracy in processing spinning time-series data streams, which are characterized by high noise levels and distinct time-domain features generated during the operation of spinning workshops, by employing a fuzzy hierarchical clustering method that integrates time-domain characteristics. This method reduces the impact of noise during the classification process and considers both time-domain features and noise effects. The technical solution of this invention is to provide a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain characteristics. This invention proposes a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain features. It iterates between the DTS feature matrix and the MTS feature matrix, considering the time-domain characteristics and noise effects in the spinning time-series data. Without increasing time complexity, it incorporates the time-frequency characteristics and noise effects in the spinning time-series data during the iteration process. Compared with the latest methods, this invention can more accurately process newly generated time-series data in spinning manufacturing.

A method and system for anomaly detection from time-series input data. A Gaussian Mixing Model (GMM) learns distribution parameters in an offline learning stage using sample data. The data used for offline learning and for a subsequent online anomaly detection stage are time-series data collected for multiple parameters of a machine operation, such as a robot performing a repetitive set of operations. The method includes aligning the data with a known good reference data file and taking a difference from it before providing the data to the GMM. In an online anomaly detection stage, the GMM calculates a probability that each time-series data point fits the distribution, and then a log-sum calculation is performed on each data file to determine the likelihood that the file contains anomaly data.The log likelihood of the file is compared with previous values, and an alert is issued if there are statistical deviations from the historical data.

This application discloses a training method, apparatus, electronic device, and storage medium for a prediction model. The method includes: acquiring raw time-series data; preprocessing the raw time-series data to obtain exogenous variables, which represent the extension of periodic data corresponding to the raw time-series data to a first preset time period; inputting the raw time-series data and exogenous variables into a preset model, and processing the raw time-series data based on the exogenous variables and a locality-sensitive hash function to obtain predicted time-series data corresponding to the first preset time period; training the preset model based on the predicted time-series data and the preset time-series data until the preset model meets preset training conditions, thereby obtaining a trained prediction model. According to the embodiments of this application, the problem of low accuracy in data prediction can be effectively solved.