74 results about "Time series data analysis" patented technology

The invention relates to the technical field of exception prediction, in particular to a storage cabinet exception trend prediction system based on time series data analysis, which comprises a state monitoring module, an interval sensing module, a path reconstruction module, a symptom activation module and an evolution prediction module. According to the method, the state vectors including the temperature, the voltage, the current and the door lock state are constructed and combined with the timestamp information to form the time sequence data sequence, and the dynamic expression mode of state change is established; a jump characteristic is analyzed by using a ratio of a time interval to a state change amplitude, a short-time disturbance path and a trend evolution path are distinguished by combining a jump rate statistical index, and an evolution activation signal is identified based on trend maintenance and non-fallback characteristics. On the basis, a neural network structure with long-time dependent learning ability is introduced to capture an aperiodic thermal anomaly trend in a state sequence, and the accuracy and timeliness of anomaly recognition are improved through multi-dimensional parameter cooperative processing and path construction logic.

The invention relates to the technical field of hydrological monitoring, and discloses a regional groundwater reserve change dynamic monitoring method based on gravity satellite time series data analysis. According to the method, the multi-source data scale effect and observation noise are effectively eliminated, the physical reliability and resolution of groundwater signal extraction are improved, and the recognition precision and early warning perspectiveness of the abnormal loss trend are enhanced. A full-chain monitoring closed loop of multi-source satellite observation, high-precision physical inversion and dynamic statistical benchmark is constructed, multi-dimensional environment data is aligned, a component separation strategy based on multi-element physical mechanism constraint is established based on a water balance principle, and anomaly diagnosis is performed in combination with a self-adaptively updated long-time-sequence historical database.

The invention discloses a federated learning time series data analysis method based on shuffling differential privacy, which comprises the following steps that: a central server initializes global model parameters and distributes an initial model to all federated learning clients; each client uses a local time sequence data training model, calculates a gradient, then cuts the gradient, and adds Laplacian noise; the client uploads the disturbed gradients to a shuffling device, and the shuffling device allocates random delay time to each gradient and disrupts an uploading sequence to realize privacy amplification; the central server aggregates the gradients processed by the shuffling device to generate a global gradient, and broadcasts the updated model parameters to the client; repeating the above steps to perform iterative training, and tracking privacy budget consumption in real time through a privacy accounting method until a preset privacy budget threshold value or model convergence is reached; according to the invention, privacy amplification is realized, and noise injection is reduced.

The invention discloses a remote sensing time series data analysis method and system considering geographic object space-time correlation, and belongs to the technical field of data analysis. The method comprises the following steps: acquiring and preprocessing a multi-source multi-temporal remote sensing image and auxiliary data to generate a consistent time sequence data stack; dividing the image into voxels in a three-dimensional time-space domain, carrying out adjacent tracking, identifying a geographic process object with an evolution behavior, and recording the attribute of the geographic process object; judging a space-time topological relation between geographic process objects, and generating a unified space-time relation table through composite reasoning; constructing a geographic process object space-time diagram model by taking the geographic process object as a node and the space-time relationship as an edge, and extracting and standardizing attribute features of the node and the edge; a graph convolutional network and a topology dynamic mechanism are used for joint modeling, and the importance index of each node is calculated. According to the method, closed-loop analysis from data processing, relation modeling and importance evaluation to decision support is realized, and the interpretability of remote sensing time sequence change detection and decision effectiveness are improved.

The invention discloses a measuring result stability analysis method and system of a metering instrument, and relates to the technical field of time sequence data analysis. The method comprises the steps of measurement data flow alignment, measurement result stability baseline feature construction, three-state feature decomposition and measurement result stability analysis of the metering instrument. Extracting repeatability features, trend features and structural features of the measurement result, and constructing a measurement result stability baseline feature vector; a three-state feature decomposition method based on baseline constraint is adopted to decompose the stability change of the measurement result into a drift feature, a noise feature and a jump feature; comprehensively analyzing the stable state and the change trend of the measurement result of the metering instrument based on the baseline feature and the tri-state feature decomposition result; according to the method, the stability of the measurement result of the metering instrument can be subjected to refined and cause-distinguishable analysis under the complex working condition and the multi-state operation condition, and a reliable basis is provided for the evaluation of the operation state of the instrument and the judgment of the credibility of the measurement result.

The invention relates to a fault prediction method and system for parameter fine tuning and adaptive structure pruning based on business influence driving, and belongs to the technical field of artificial intelligence, time series data analysis and intelligent operation and maintenance. Comprising the following steps: S1, business influence data modeling: integrating multi-source operation and maintenance data and historical business fault event data, and constructing a business influence quantitative model; s2, efficient fine tuning of service influence guide parameters: loading the pre-training fault prediction model, and performing efficient fine tuning of the parameters on the basis of service influence signals; s3, business influence driven adaptive structure pruning; S4, dynamic reasoning and alarm generation: realizing a dynamic reasoning mechanism of a pruned fault prediction model, and generating a structured and business value oriented fault prediction alarm according to a reasoning result; and S5, performing closed-loop optimization and continuous evolution. According to the invention, the configuration accuracy of the operation and maintenance resources is improved.

The invention discloses a frequency converter abnormal state detection method based on time series data analysis, which belongs to the technical field of electric power, and comprises the steps of: finally generating a structured abnormal event through acquisition, preprocessing, feature extraction, anomaly detection, anomaly level judgment and trend analysis of data of a multi-source sensor; accurate prediction and early warning of the operation state of a complex system are achieved, the technical problems of high real-time performance, low data redundancy and multi-source feature fusion of abnormal monitoring of the frequency converter are solved, redundant data are reduced, the data volume is reduced, the communication bandwidth pressure is reduced, it is ensured that high-frequency, low-frequency and event data are aligned on the same time axis, and the accuracy of abnormal monitoring of the frequency converter is improved. Abnormal data are transmitted preferentially, abnormal information is guaranteed to be complete and reliable, suspicious anomalies are screened rapidly, local and remote alarms are generated according to the abnormal levels, an abnormal database and a periodic summary report are stored, and operation and maintenance and remote monitoring are facilitated.

The invention provides a full-period follow-up visit patient risk dynamic management method based on time series data analysis. The method comprises the steps of performing standardization processing on multi-source follow-up visit data of a patient; data acquisition is dynamically optimized through an adaptive strategy based on reinforcement learning, and multi-scale time sequence features are extracted; performing dynamic risk assessment by using the attention mechanism enhanced double-layer LSTM model, and outputting a risk probability; and calculating a comprehensive risk index by integrating the risk probability, the change trend and the volatility, and realizing dynamic risk layering and automatic intervention based on a rule engine. According to the method, a traditional fixed acquisition mode is converted into a dynamic optimization process capable of responding to patient risks, equipment states and system loads in real time, on the premise that high-risk patient monitoring is guaranteed, system resource consumption is remarkably reduced, and long-term optimal balance of data quality, system performance and resource overhead is achieved.

The invention discloses a non-contact screening system for motion coordination based on computer vision, which relates to the field of computer vision, and comprises the following steps: acquiring a human body motion image through a non-contact camera, extracting human body key point coordinates frame by frame by using a deep learning algorithm after noise reduction and distortion correction, and constructing a time sequence data set; performing smooth jitter removal and interpolation processing to form standardized data; calculating track, rhythm and stability multi-dimensional motion coordination quantitative characteristics based on the data, and performing comparative analysis with a personal historical database to identify a motion function degradation trend; and based on age groups and scenes, customizing screening parameters, coordinating workflows of the components, and realizing full-process automatic screening. The method has the advantages that invasive and cross infection risks are avoided through non-contact acquisition, accurate and objective assessment of the exercise coordination is realized by means of deep learning and time series data analysis, exercise function degradation can be early warned, and an efficient and reliable scheme is provided for various exercise coordination screening scenes.

The invention provides an enterprise production data processing system based on artificial intelligence big data analysis. The system comprises a real-time data acquisition module; the streaming data processing engine is connected with the real-time data acquisition module and used for efficient processing based on the real-time data stream; the time sequence data analysis module is connected with the streaming data processing engine and is used for capturing a long-term dependency relationship among data and realizing accurate prediction of equipment faults; the image data analysis module is connected with the streaming data processing engine and is used for improving the accuracy and efficiency of product quality detection and carrying out product quality inspection image analysis; and the association relationship analysis module is connected with the streaming data processing engine and is used for constructing an association graph of the equipment process materials and mining hidden relationships. By applying the technical scheme of the invention, the problems of insufficient real-time performance, data processing capability and intelligent degree in the prior art can be effectively solved, especially in the aspects of real-time data processing, deep intelligent analysis and adaptive optimization.

The application discloses an optical instrument performance detection system and method based on multi-source data fusion analysis, and relates to the technical field of optical instrument detection, to solve the technical problem that in the prior art, the signal characteristics of environmental light interference and product defects are prone to confusion, and specifically, through the cooperative work of an instrument self-calibration module, a calibration evaluation unit, an environmental light interference detection unit and a health state prediction unit, a full-process intelligent detection system is constructed, the application of multi-source data fusion technology breaks through the limitation of traditional single data detection, multi-modal feature matching, time series data analysis, grid fine monitoring and other technologies improve the accuracy and comprehensiveness of detection, accurate differentiation of environmental light interference and product defects reduces misjudgment and missed judgment, and the proactive maintenance of the hardware health state reduces downtime loss.

This invention discloses a method and system for stability analysis of metrological instrument measurement results, relating to the field of time series data analysis technology. The method includes measurement data stream alignment, construction of baseline features for measurement result stability, three-state feature decomposition, and stability analysis of metrological instrument measurement results. Repeatability, trend, and structural features of the measurement results are extracted to construct a baseline feature vector for measurement result stability. A three-state feature decomposition method based on baseline constraints is used to decompose the stability changes of the measurement results into drift, noise, and jump features. Based on the baseline features and the three-state feature decomposition results, a comprehensive analysis of the stability state and changing trends of the metrological instrument measurement results is performed. This invention enables refined and causally distinguishable analysis of the stability of metrological instrument measurement results under complex operating conditions and multi-state operation, providing a reliable basis for instrument operating status assessment and measurement result reliability judgment.

The application discloses a traffic sequence prediction method based on multivariate time series data analysis, adopts a random graph diffusion attention mechanism to extract global and local spatial features of a traffic sequence, uses time attention to extract time features, improves prediction accuracy, reduces memory usage of a model, and improves the effect of the model on long-term prediction. The traffic sequence prediction method based on multivariate time series data analysis solves the problems of insufficient short-term prediction accuracy, high calculation complexity, large memory occupation, and insufficient lightweight of the existing model while maintaining the accuracy of long-term prediction.

The invention belongs to the technical field of disease and pest early warning, and discloses a change detection disease and pest early warning method fusing multispectral time sequence images and deep learning, which comprises a data acquisition module, a data preprocessing module, a model design module, a disease and pest recognition module and an early warning module, the data preprocessing module comprises an automatic preprocessing process and a high-quality cloud mask algorithm, the model design comprises an input layer, a feature extraction module, a time sequence modeling module, a change detection head and an output layer, and the early warning module comprises time sequence data analysis and uncertainty analysis. The early-stage, accurate and automatic early warning capability is realized; experimental results show that the method shows excellent performance in a plurality of crops and areas, and has good popularization and application values.

This deep learning-based method for predicting lung cancer recurrence using time-series data, comprises the steps in which: an analysis device receives, from a patient, inputs of static data collected before surgery and time-series dynamic data collected after the surgery; the analysis device pre-processes the static data and basic dynamic data among the time-series dynamic data; the analysis device inputs, into a trained prediction model, the pre-processed static data, the pre-processed basic dynamic data, and a medical image analysis sentence included in the time-series dynamic data; and the analysis device predicts the possibility of lung cancer recurrence of the patient within a predetermined period on the basis of a value output by the prediction model.

Artificial intelligence for time-series data analytics is provided. A first time-series data set is provided to a pre-trained recurrent neural network trained based on a second time-series data set. A prediction of a ground truth state of the first time-series data set is received therefrom. The first time-series data set is provided to a dynamical recurrent neural network trained based on the second time-series data set and the pre-trained recurrent neural network. A noise-reduced prediction of a ground truth system state of the first time-series data set is received therefrom. An estimate of sensor noise is read. The estimate of sensor noise is generated based on the second time-series data set and the pre-trained recurrent neural network. A prediction of a state of the system is generated based on the pre-trained recurrent neural network, the dynamical recurrent neural network, and the estimate of sensor noise.

The application provides a method for extending time series database analysis capability, and relates to the field of data analysis, which comprises the following steps: constructing a time series data analysis system for providing time series data analysis service independently of a time series database system, and extending, upgrading and maintaining the time series data analysis system without affecting the normal operation of the time series database system; adding an adaptive interface in the time series data analysis system, and calling different time series data analysis models; establishing an analysis operator for the time series data analysis service in the time series database system, and the analysis operator calling the corresponding time series data analysis model through the adaptive interface to complete time series data analysis in the execution of an SQL query process, so that interaction between the time series database system and the time series data analysis system which are isolated from each other in the processing flow is realized, and the complexity of the analysis and processing process is shielded.

Systems and methods are provided for implementing cloud distributed database capacity planning and / or adjustment, using time-series data analysis. In various embodiments, a computing system may be used to analyze collected throughput data associated with consumption of provisioned throughput resources of a distributed cloud database over one or more past periods by an entity. Based on the analysis, a set of predicted throughput data may be determined or generated for the entity over a future upcoming period. In some cases, based on a determination that adding physical partitions would be required, the computing system may adjust the set of predicted throughput data to reduce or minimize a number of physical partitions to be added. The provisioned capacity of the distributed cloud database may then be dynamically adjusted based at least in part on one of the unadjusted or adjusted set of predicted throughput data.

The application belongs to the technical field of time series data analysis, and discloses an online time series event analysis method and system based on heterogeneous memory fusion, which comprises the following steps: acquiring streaming time series data and preprocessing, extracting features from the preprocessed data to obtain time series feature sequences; long-term feature memory, event state probability memory and event-level semantic memory are constructed and fused to form a fusion memory; an event query representation is initialized, and the event query representation is enhanced through a short-term memory and a fusion memory in turn to generate a final event query representation; based on the final event query representation, an instantaneous event state and a candidate event instance at the current time are generated, the candidate event instance is de-redundant and cross-time deduplication processed, and a final event set is output. By constructing and fusing heterogeneous historical memories, the application realizes effective utilization of multi-granularity time series knowledge, thereby significantly improving the accuracy, stability and real-time performance of online event analysis under strict causal constraints.

The present application relates to the technical field of data acquisition, in particular to a kind of intelligent network connection automobile test bench road resistance time series data analysis method.The method comprises: obtaining the oscillation stationary characteristic value and response decay time characteristic value of road resistance data;According to the difference between different road resistance data sequences, the oscillation stationary characteristic value, the response decay time characteristic value of the marked road resistance data, the target metric distance between the test bench road resistance data in the test bench road resistance data sequence and the road test road resistance data in the road test road resistance data sequence is obtained, and the difference between the test bench road resistance data sequence and the road test road resistance data sequence is obtained according to the target metric distance difference evaluation result.The present application can improve the accuracy and reliability of the difference or similarity evaluation between the test bench test road resistance data and the real road test road resistance data.

The application relates to a fault prediction method and system based on service influence driven parameter fine tuning and adaptive structure pruning, and belongs to the technical fields of artificial intelligence, time series data analysis and intelligent operation and maintenance. The application comprises the following steps: S1: service influence data modeling: integrating multi-source operation and maintenance data and historical service fault event data, and constructing a service influence quantification model; S2: service influence guided parameter efficient fine tuning: loading a pre-trained fault prediction model, and performing parameter efficient fine tuning on the model based on a service influence signal; S3: service influence driven adaptive structure pruning; S4: dynamic reasoning and alarm generation: realizing a dynamic reasoning mechanism of the pruned fault prediction model, and generating a structured and service value oriented fault prediction alarm according to a reasoning result; and S5: closed loop optimization and continuous evolution. The application improves the configuration accuracy of operation and maintenance resources.

The invention discloses an industrial time series data analysis method and device, equipment and a storage medium. Continuous industrial time series data is converted into a data set containing a high-dimensional vector, an adaptive data format is provided for subsequent model processing, and rich original information support is provided for model learning. An architecture model with Lma as a base is built, global dependency relationships of different time points in industrial time sequence data can be captured in parallel, and the defect that a traditional model processes long-term time sequence dependency is overcome. Through supervision and fine tuning of the time series data set, the model can be specialized in an industrial safety parameter prediction task, and the prediction accuracy of the model in a target industrial scene is improved. Real-time data is analyzed through the safety early warning model, a safety analysis result is rapidly output, real-time monitoring and potential risk identification of the operation state of the industrial system are achieved, a timely and effective decision basis is provided for industrial safety early warning, and industrial production safety is guaranteed.

The invention discloses an equipment pre-maintenance time series data analysis method and system based on industrial Internet of Things, and relates to the technical field of data transmission processing, and the method comprises the steps: obtaining a historical record period, and obtaining a historical maintenance interval sequence; obtaining a standard maintenance interval, correcting the standard maintenance interval and generating a target maintenance interval; a historical effective maintenance interval is acquired, a historical interval trend value is acquired, a plurality of maintenance equipment sets respectively comprising a plurality of different Internet of Things equipment are acquired, and the difference value of the historical interval trend values of any two Internet of Things equipment in each maintenance equipment set is smaller than a preset threshold value; and obtaining a pre-maintenance interval sequence of each maintenance equipment set according to the plurality of Internet of Things equipment in each maintenance equipment set, and obtaining a pre-maintenance time sequence of the plurality of Internet of Things equipment in each maintenance equipment set according to the pre-maintenance interval sequence of each maintenance equipment set. The method has the advantages that the historical data trend is accurately quantified, collaborative maintenance is achieved, and the maintenance cost is reduced.

The invention discloses a multivariate time series data generation and deletion completion method based on a diffusion model, and relates to the technical field of time series data analysis and modeling, and the method comprises the following steps: S1, generating a data embedding tensor and a mask embedding tensor based on double-branch coding; s2, extracting an initial time sequence latent representation tensor based on a bidirectional cross attention mechanism; s3, extracting a generative latent representation tensor; s4, performing statistic calculation; s5, constructing a whitening transformation matrix and a coloring transformation matrix; s6, extracting the aligned generative latent representation tensor; and S7, generating target completion data based on a double-track fusion refining decoding mechanism. According to the method, the limitations of high result uncertainty, insufficient data internal mode capture and inconsistent original data distribution in a traditional single generation type completion method are overcome, and an efficient and accurate solution is provided for multivariate time series data analysis in the fields of finance, medical treatment, meteorology and the like.

The invention relates to the technical field of time sequence data analysis, and particularly provides a continuous time sequence prediction method based on reinforcement learning and clustering experience playback. The method comprises the following steps: obtaining a structured experience tuple through real-time prediction and experience packaging; experience clustering and clustering experience knowledge base construction are carried out according to the experience tuple; based on the clustering experience knowledge base, intelligent playback decision making is carried out, and an optimal playback strategy is generated; performing mixed training and collaborative updating according to the optimal playback strategy; after mixed training and collaborative updating, continuous self-evolution and closed-loop optimization are carried out, and the method improves the active decision-making ability of time series data prediction and realizes efficient and accurate continuous learning.

The application discloses an epidermal growth factor adding effect prediction method based on growth performance data, relates to the technical field of time series data analysis, and comprises the following steps: collecting target object data and structuring the data, establishing an original data set containing growth performance time series, environmental variables and epidermal growth factor adding dose records; calculating residual errors after preprocessing, performing wavelet decomposition, and combining input features; inputting a double-track learning model to obtain individualized causal effect estimation and dose-time-growth performance prediction mapping, then performing new observation of the target object to establish a correction factor, and outputting a prediction result after compensation. The application solves the technical problem that traditional epidermal growth factor adding effect prediction lacks consideration of dynamic changes in the time dimension and individual differences of different target objects, and cannot meet the precise evaluation and individualized regulation, and achieves the technical effect of accurate and comprehensive prediction of the epidermal growth factor adding effect.

The application relates to the technical field of anomaly prediction, in particular to a storage cabinet abnormal trend prediction system based on time series data analysis, which comprises a state monitoring module, an interval sensing module, a path reconstruction module, a symptom activation module and an evolution prediction module.In the application, a state vector containing temperature, voltage, current and door lock state is constructed, time stamp information is combined to form a time series data sequence, a dynamic expression mode of state change is established, the ratio of time interval to state change amplitude is used to analyze the jump characteristic, and a jump rate statistical index is combined to distinguish a short-time disturbance and a trend evolution path, an evolution activation signal is identified based on trend retention and non-falling characteristics, a neural network structure with long-time dependent learning ability is introduced to capture the aperiodic thermal abnormal trend in the state sequence, and the accuracy and timeliness of abnormal identification are improved through multi-dimensional parameter collaborative processing and path construction logic.

The application discloses a kind of remote sensing time series data analysis method and system considering the spatio-temporal correlation of geographical object, belong to data analysis technical field.The method includes: obtaining multi-source multi-temporal remote sensing image and auxiliary data and pre-processing, generate consistent time series data stack;In three-dimensional space-time domain, image is divided into voxel and is adjacent tracking, identify geographical process object with evolution behavior and record its attribute;Determine the spatio-temporal topological relationship between geographical process object, generate unified spatio-temporal relationship table by composite reasoning;With geographical process object as node, spatio-temporal relationship as edge constructs geographical process object spatio-temporal graph model, extracts and standardizes the attribute characteristics of node and edge;Importance index of each node is calculated by using graph convolution network and topological dynamic mechanism joint modeling.The application realizes closed-loop analysis from data processing, relationship modeling, importance evaluation to decision support, improves the interpretability and decision effectiveness of remote sensing time series change detection.

According to the method for evaluating the acute kidney injury risk of the patient with the abnormal creatinine level, a traditional single-index evaluation mode is broken through in multi-dimensional risk evaluation, and a comprehensive AKI risk evaluation system is established by integrating multi-dimensional clinical data such as medical advice, medication, operation, nursing and blood transfusion. A time sequence data analysis technology is adopted, and the creatinine level of a patient is continuously and dynamically monitored. A creatinine interference model is innovatively constructed, and the influence degree of various clinical diagnosis and treatment items on the creatinine concentration is quantified. And according to a risk assessment result, automatic grading early warning is carried out, detailed analysis is provided for a high interference condition, and a nephrology consultation template is generated. Based on historical data and real-time monitoring information, the future creatinine change trend of the patient can be predicted, and potential risks can be recognized in advance. Hospitals and doctors need more accurate AKI risk assessment tools to improve clinical decision-making efficiency, and treatment schemes can be adjusted in time according to dynamic monitoring of creatinine level changes of patients. The system provides an accurate, dynamic and full-process acute kidney injury risk assessment solution, and the AKI early recognition and intervention capability is remarkably improved.

The invention relates to the field of resource optimization, in particular to a LoRA fine-tuning computing power resource dynamic allocation method, which comprises the following steps of: obtaining a video memory state of a computing cluster and a computing unit load data stream, and analyzing according to historical time sequence data to obtain prior characteristics representing a resource fragmentation degree and a load level; bottleneck labels are set for all kinds of calculation tasks; when the task is executed, differentiated scheduling optimization is carried out according to the label type and the real-time resource characteristics, and by monitoring the real-time video memory state characteristics, the video memory bottleneck task is temporarily deducted until the video memory bottleneck task is reduced to a preset threshold value and then executed; by monitoring the real-time load state characteristics, the bottleneck calculation task is started when the load meets the condition. According to the method, the sensing response to the cluster dynamic resource state and the targeted scheduling of the task bottleneck type are realized, the blockage and idleness caused by video memory fragmentation and non-uniform resource distribution are relieved, and the resource utilization balance degree and the overall execution efficiency of the LoRA fine tuning task are improved.