115 results about "Event forecasting" patented technology

The invention belongs to the technical field of wind power climbing event prediction, and particularly relates to a wind power climbing event prediction method considering extreme weather and time-space correlation information. The method comprises the following steps: acquiring historical actually measured meteorological data of each wind power station of a cluster; carrying out cold-wave weather event identification on historical actually measured meteorological data, generating an antagonistic network based on a time sequence, and carrying out cold-wave event sample expansion; an extreme learning machine is constructed, and cold-wave weather prediction is carried out; performing historical climbing event detection on historical power output results of each station of the cluster; dividing the climbing events into various climbing conditions with different severity degrees by using a K-Means clustering algorithm; and carrying out climbing event prediction. According to the method, sample support is provided for training of the climbing prediction model, the climbing events are clustered and divided by fusing the fan operation state and the climbing characteristics, and the harm degrees of different climbing events, especially the climbing events in extreme weather, are finely measured.

This invention discloses an event prediction method and apparatus based on subgraph segmentation and entity relationship model. It transforms an academic knowledge graph into a sequence of temporal historical subgraphs. Segmentation operations determine the segmentation points of the temporal historical subgraph sequence and the number of segments corresponding to different times. The number of continuous segments is also determined. A DQN network is used to perform secondary path search through target operations to obtain the final search path. The entity in the last node is the object to be predicted. This invention achieves this by segmenting the temporal historical subgraph sequence to give the input data a focus, enabling better selection of answer nodes. Secondly, the use of a DQN network better considers the intrinsic connections between entities and relationships. Finally, the global nature of convolutional neural networks captures the global features of the search path, significantly improving prediction accuracy.

The present disclosure provides a threat intelligence information processing method and device and a storage medium, wherein the method comprises: respectively performing fusion processing on vulnerability intelligence information and event intelligence information to generate vulnerability intelligence fusion information and event intelligence fusion information; extracting attribute information from the vulnerability intelligence fusion information, and determining whether the vulnerability intelligence information corresponding to the vulnerability intelligence fusion information is threat vulnerability information based on the attribute information and asset information; processing the event intelligence fusion information by using a trained event prediction model to determine whether the event intelligence information corresponding to the event intelligence fusion information is industry threat event information; and performing security processing on the threat vulnerability information and the industry threat event information. The present disclosure can improve the accuracy of threat intelligence information identification, reduce the situation of missed reporting of threat intelligence and false reporting of invalid threat intelligence, reduce the response time of security operation personnel, improve the network security protection capability, and reduce the operation cost.

The invention discloses an event prediction method and system based on a time sequence hypergraph, and belongs to the technical field of event prediction. The method comprises the following steps: acquiring multi-source heterogeneous data of a target region, defining a unified time index, and generating a time sequence feature sequence of each variable of the region through preprocessing; then, identifying a variable causal relationship in the region based on a frequency domain anti-fact condition mutual information algorithm, and fusing the variable causal relationship with a time sequence evolution relationship to construct a time sequence hypergraph structure representing the interior of the region; a dynamic filter is used for filtering the graph, and deep features of the graph are learned by means of a multi-band spectrum gating mechanism, so that internal complex causal and time sequence modes are effectively captured; and finally, performing dichotomy prediction based on the learned graph representation, and outputting the occurrence probability of future events in the region. According to the method, accurate and explainable event prediction is realized, training and prediction do not need to cross regions, data privacy and calculation efficiency are guaranteed, and stronger robustness is shown for specific data distribution change of the regions.

Aspects described herein may use machine learning models to predict one or more remediation actions to mitigate reoccurrence of an incident that has become restored based upon previous incidents of an entity. Historical incident data is compiled into two incident datasets: one representative of incidents that were assigned a remediation action to mitigate reoccurrence of the incident, and a second representative of incidents that were not assigned a remediation action. A machine learning model matches relationships between data in the two datasets and outputs scores representative of similarities. Based on the scores, one or more remediation actions are mapped to an incident in the second dataset and the remediation action is performed for the incident.

The invention discloses a graph reasoning-based judgment document event joint extraction method. According to the method, a lexical graph inference mechanism and a trigger word centralization decoding architecture are fused, the lexical graph inference mechanism strengthens the relevance between trigger words and argument roles and solves the problem of semantic segmentation of entities and events, and the trigger word centralization decoding architecture avoids multi-stage cascade errors and improves the complex logic chain processing capacity. The method comprises the following steps: constructing a joint event extraction model: performing corpus labeling and BIO labeling on a data set, and constructing an event type-role label set; processing the text subjected to BIO labeling into numeric vector representation capable of being input into a model, and constructing a lexical element pair label matrix; the hidden lexical elements output by the text embedding layer are classified through a classification layer, a softmax function is adopted to predict relation marks between lexical element pairs, label distribution corresponding to the lexical elements is generated, and a basis is provided for subsequent event extraction; and establishing an event center graph according to lexical tag distribution, performing graph structure decoding to obtain an event prediction result, and training a model.

Methods and systems for predicting events during hydraulic fracturing of a target subterranean formation, inccluding: obtaining a time-series of measurements associated with the hydraulic fracturing; decomposing the measurements into a plurality of respective components that collectively represent nonstationary and nonlinear properties of the pressure measurements; extracting respective attributes that are representative of events occurring in the target subterranean formation from the respective components; and predicting, based on the extracted attributes, a likelihood of one or more future events being induced by the hydraulic fracturing.

This invention belongs to the field of overflow events, specifically relating to a hierarchical multi-joint learning-based method, system, and equipment for predicting well control overflow events. It aims to address the problems of equipment damage, production interruption, safety accidents, and environmental pollution caused by overflow event risks. The invention includes: training multiple well control overflow event prediction base models based on the training set; obtaining prediction results based on the test set; calculating the accuracy of the base models and selecting retained models to obtain a final feature set; integrating the retained models based on the final feature set to construct a voting-based ensemble model; and inputting the data to be predicted into the ensemble model to obtain the well control overflow event prediction result. This invention improves prediction accuracy and robustness by constructing an ensemble model, optimizes resource utilization, enhances operational safety and decision-making efficiency, effectively prevents well control overflow events, and ensures the safety and economic benefits of oilfield operations.

The invention discloses a rail transit abnormal event prediction method based on spatial-temporal feature fusion, and relates to the technical field of public traffic data analysis and prediction, and the method specifically comprises the following steps: S1, data collection and preprocessing; s2, spatio-temporal feature extraction and fusion; s3, selecting and training a model; s4, abnormal events are predicted in real time; according to the method, a closed-loop process from data collection to prediction feedback is realized, spatial-temporal feature fusion plays a key role in identifying rail transit abnormal events, the influence of the abnormal events on a rail transit system can be effectively reduced, the operation efficiency and safety are enhanced, and the method is suitable for popularization and application. The method achieves the precise prediction of the abnormal events of train delay, equipment faults and passenger flow congestion through the multi-source data fusion and deep learning model through the features of spatial distribution and time dynamic changes in a rail transit system, and provides powerful technical support for the operation safety management of rail transit.

A computer-based system and corresponding method perform feature-based machine learning (ML) model prediction. The system uses an imputation method to produce posterior distributions of unprovided features of a set of retrospective features. The posterior distributions are produced based on the set of retrospective features and provided features of the set of retrospective features. The system employs an ML model to produce a threshold and a risk score distribution of a prediction of an event and selects at least one unprovided feature from a partial set of the unprovided features to improve predictive accuracy of the ML model iteratively. The system outputs a representation of the at least one unprovided feature selected toward approximating a full-feature-capacity (FFC) prediction with a partial set of the retrospective features. The system enables efficient feature acquisition for accurate ML model prediction.

The invention is suitable for the field of marine meteorology, and provides a three-dimensional sea temperature space-time prediction method, which comprises the following steps: obtaining reanalysis data and numerical mode prediction data; and inputting the reanalysis data and the numerical mode prediction data into a pre-trained three-dimensional sea temperature intelligent prediction model to obtain a three-dimensional sea temperature space-time prediction result. According to the method, the spatial-temporal characteristics of the sea temperature change can be effectively captured, so that the prediction precision is improved, a high-precision sea temperature prediction result can be provided, good interpretability is achieved, and more powerful support is provided for scientific research and climate change response.

This disclosure provides a method, apparatus, electronic device, and storage medium for predicting ramp events, relating to the field of computer technology, and particularly to the field of power detection. The specific implementation scheme is as follows: A time window is slid across the wind power sequence, and sequences within the time window are extracted to obtain multiple wind power sub-sequences; among the multiple wind power sub-sequences, multiple first sub-sequences belonging to ramp events are identified; based on the Euclidean distance between each first sub-sequence and other sub-sequences within the multiple first sub-sequences, neighboring sub-sequences of each first sub-sequence are determined; based on the sampling time of wind power in the neighboring sub-sequences of each first sub-sequence, the occurrence time interval of the ramp event is determined; based on the most recent occurrence time of the ramp event and the occurrence time interval, the next occurrence time of the ramp event is determined. Using the technical solution of this disclosure can improve the prediction accuracy of power ramp events.

The invention relates to the technical field of electric digital data processing, in particular to an event prediction and alarm generation method based on multi-modal physiological time sequence data, which comprises the following steps: judging whether a suspected falling trend occurs or not; determining a plurality of preliminary abnormal segments; determining a physiological feature vector corresponding to each preliminary abnormal segment; calculating a physiological comprehensive index, and determining a plurality of final abnormal segments; calculating a physiological tumble index and an environmental slip index; determining an event type corresponding to the suspected tumble trend based on the difference between the physiological tumble index and the environmental slip index; generating a corresponding alarm prompt based on the real prediction type; and adjusting a preset impact amplitude threshold value. According to the method, the problems of low prediction accuracy, high false alarm rate and poor system adaptive capacity caused by single data mode and excessive dependence on static parameters in the prior art are effectively solved through multi-mode time sequence data fusion and a dynamic threshold adjustment mechanism.

This invention discloses a method for predicting extreme positive IOD events based on ENSO and MJO. This invention uses the ENSO and MJO co-signals and extracts the key signals of ENSO and MJO that affect IOD. At the same time, it combines model forecast field information to perform statistical modeling for extreme positive IOD event prediction. This method avoids the systematic underestimation of strong IOD amplitude by dynamic models and can improve the prediction lead time and prediction accuracy of extreme positive IOD events.

The invention discloses an extreme cold event prediction method based on a multi-time-scale climate oscillation signal, and relates to the technical field of meteorological prediction. The method comprises the following steps: acquiring climate oscillation signal prediction data of multiple time scales; carrying out initial judgment according to the ENSO index: if the ENSO index is greater than or equal to 0, judging the risk as a 0-level risk, and outputting a code 000; if the ENSO index is less than 0, entering a PDO index judgment step; respectively entering a positive PDO sub-process or a negative PDO sub-process according to positive and negative PDO indexes; judging whether the MJO is at the second or third phase and the amplitude of the MJO is Agt in each sub-process by combining an MJO index; the method comprises: 1, determining a final risk level and an output code; the output result includes a risk level and an output code representing a trigger condition. The Taiwan strait extreme cold event prediction method can predict the Taiwan strait extreme cold event in China according to the three different time scale climate oscillation signals of the PDO phase, the ENSO phase and the MJO activity, and the prediction precision is improved.

The application discloses a kind of day-ahead wind power power ramp event prediction methods, comprising the following steps: S1.day-ahead wind power prediction: using extreme value driving model, based on historical wind power and meteorological characteristics, produce the wind power prediction of Q time points of next day;S2.predicting confidence interval construction: using wind power oriented conformal inference method, for each future time point, construct confidence interval C;S3.ramp event detection based on confidence interval: the prediction result of next day wind power ramp event is obtained by confidence perception detection algorithm.The application improves the prediction accuracy and reliability of wind power ramp event with significant operation risk by fusing multi-scale time series feature analysis, customized loss function for extreme event and adaptive confidence interval construction technology with statistical guarantee.

The present application relates to the technical field of wireless communications, and discloses an event prediction method and apparatus, a terminal, and a network side device. The event prediction method in the embodiments of the present application comprises: a terminal acquires a first parameter; the terminal acquires a predicted measurement result; and on the basis of the first parameter and the predicted measurement result, the terminal determines whether a trigger condition for a target event is satisfied.

The invention relates to the technical field of liquefied natural gas storage and transportation, and discloses a tumbling event prediction method, system, equipment and medium for an LNG storage tank, and the method achieves the fundamental transformation of the tumbling event of the LNG storage tank from post-event alarm to pre-event prediction by constructing a tumbling event prediction framework in which a physical model and a machine learning model are fused. Multi-dimensional early warning information such as rolling remaining time and risk probability can be quantitatively output, an effective intervention window is remarkably prolonged, and prediction precision and decision support capability are improved; meanwhile, based on software and hardware collaborative system-level design, data instantaneity and model reliability are guaranteed, and finally a comprehensive early warning solution which is high in perspectiveness, high in reliability and high in engineering practical value is formed.

This invention discloses a business process anomaly detection method based on concept drift discovery, comprising the following steps: 1) collecting data to form an event log; 2) extracting process information using control flow features from the event log, encoding events, and constructing a process feature dataset; 3) building a prediction model for the next event in the business process based on a GRU model, and training the prediction model using the process feature dataset as input data; 4) calculating the anomaly score s of the business process attributes through probability distribution; 5) performing concept drift detection on the anomaly detection results using a concept drift discovery module; and 6) using an incremental learning method to incorporate the drift case set as new knowledge using an event prediction model update module. This method mines process models from event logs without requiring manual judgment to find concept drift cases, enabling more accurate detection of whether anomalies occur in business process instances and locating and determining whether concept drift has occurred.

The invention provides a robot zero-value insulation detection system based on intelligent sensing, relates to the technical field of electrical variable measurement, and aims at detecting the zero-value insulation of a robot by constructing and solving a state evolution model jointly driven by an internal health state and future external stress and generating a future health state track. The accuracy and foresight of critical failure event prediction are improved, and the risk of unplanned shutdown caused by sudden insulator failure is reduced; secondly, deep attribution analysis is performed on the dominant stress causing deterioration of the state trajectory, so that decision support is provided for operation scheduling and maintenance strategies of the power grid, and resources can be more effectively allocated to a link with the highest risk; the robustness and the intelligent level of the whole detection system in a complex and changeable operation environment are enhanced, and transformation of a power transmission line operation and maintenance mode to higher-order predictive maintenance and initiative risk management and control is promoted.

The application discloses an extreme event-oriented main network key element identification method and system, and mainly relates to the technical field of extreme event prediction. The method comprises the following steps: based on the importance of load and the remedial measures of the power system to cope with extreme events, taking the maximum value of the system loss caused by the power cut of the extreme event as a first objective function; using a direct current flow model to model power system operation and establish element constraints; based on the first objective function, power system operation modeling and element constraints, establishing upper and lower two-layer robust models and converting the two-layer robust models into a main problem and a sub-problem; solving the main problem and the sub-problem and outputting the key element combination in the power system; and determining the importance degree of each element in the element combination through the repair sequence of the element. The application has the beneficial effects that it solves the problem that the Monte Carlo simulation method cannot generate the corresponding scene due to the small event occurrence probability, and identifies the key element combination in the power transmission network.

The invention discloses a time sequence knowledge graph reasoning method based on autoregression condition diffusion generation, and belongs to the technical field of artificial intelligence and dynamic knowledge graphs. According to the method, entity structure dependency in a single time snapshot is aggregated through a relation-aware graph neural network, time-gated loop unit autoregression modeling is combined to capture cross-snapshot time sequence evolution characteristics, and dynamic embedding is output; simulating future uncertainty through a forward diffusion process under the condition of the embedding, and generating future representation aligned with real distribution through reverse denoising iteration to cover a multivariate potential evolution path; and designing an entity and relation fusion gate, adaptively balancing historical experience and generating a prediction weight, and combining a ConvTransE decoder to realize high-precision link prediction. According to the method, adaptive capacity to uncertainty is enhanced through generative modeling, modeling comprehensiveness is improved through multi-granularity feature fusion, history and generated information are balanced through a dynamic gating mechanism, prediction flexibility and robustness are remarkably improved, and the method is suitable for sequential reasoning tasks in multiple fields such as event prediction and a dynamic recommendation system.

The invention discloses a multivariable time point process analysis method based on two-dimensional local differential privacy, and relates to the technical field of federated learning and privacy protection. The method comprises the following steps: setting a total privacy budget on a client side and distributing the budget as a time and type dimension budget; a delta t-SAHRP-LDP order-preserving perturbation mechanism is adopted in the time dimension, interval scale adaptive noise adding, minimum right lower bound constraint and dynamic anchor point resetting are combined, and a time sequence after perturbation is generated; and generating a disturbed type sequence in the type dimension by adopting a random response mechanism. A client uploads an intermediate representation and an alignment scale based on disturbance data, and a server establishes a cross-end alignment and neighborhood relationship only according to the intermediate representation and the alignment scale, generates extension and neighborhood context features and returns the extension and neighborhood context features. And the client performs prediction and training in combination with the local real representation and the cross-end context. According to the method, on the premise of not exposing an original sequence, the leakage risk of sensitive time rhythm and type distribution is effectively reduced, and meanwhile, the accuracy of event prediction is improved by utilizing cross-client collaborative information.

The application provides a method and device for event prediction based on a timing diagram rule, the method comprising: obtaining timing diagram data, and adding event prediction results of a machine learning model to the timing diagram data to obtain timing diagram extension data; extracting a subgraph for verifying prediction rule reliability from the timing diagram extension data according to a verification ratio; training a rule creator according to the timing diagram extension data and the subgraph to obtain a timing event prediction rule set; and obtaining a prediction event set according to the timing event prediction rule set and the timing diagram data. By embedding the machine learning model for event prediction into a rule-based association relationship system as a predicate, the association rule can not only use the existing machine learning model event prediction result, but also use a logical condition to improve the prediction result of the machine learning model, has stronger expression ability, and does not have to constrain constant interval time or have a common focus constraint on a graph pattern.

This application relates to a method, model, system, and related equipment for predicting target events. The method for predicting target events includes: acquiring several interaction events of a target user, wherein the interaction events constitute a behavioral sequence of the target user; encoding each interaction event to form an event feature vector; performing feature extraction and serialization processing on each event feature vector to form a sequence feature vector of the behavioral sequence, which serves as the latent state of the target user; and fusing the latent state and static attributes of the target user to obtain a prediction result of the occurrence of the target event. This method can cover the interaction chain of the target user, improving the reliability of target event prediction.

The embodiment of the invention provides a risk early warning method for medicine production, electronic equipment and a medium, and belongs to the technical field of medicine production and artificial intelligence. The method comprises the following steps: acquiring at least one of personnel behavior data, equipment operation data, process data, material data and environment data in a drug production area, performing risk detection on each source data, and generating a reference event according to other source data in multi-source data to predict the risk of a risk event, and executing an early warning operation matched with the risk event and the event risk degree corresponding to the source data. According to the embodiment of the invention, real-time risk event detection and early warning operation are carried out on the drug production area, so that the final influence of risk events such as non-standard operation on the drug quality can be reduced fundamentally.

This invention discloses a method, system, storage medium, and device for selecting emergency medical equipment, applied in the field of information processing technology. The emergency medical equipment selection system performs gridding processing on a target area to obtain multiple grid cells and acquires grid data for each cell, including current emergency medical equipment, traffic, weather, social factors, and historical out-of-hospital emergency events. Based on this, it acquires the grid characteristics, spatial connectivity characteristics, and temporal characteristics of each grid cell. After fusing these acquired features, it predicts the occurrence information of out-of-hospital emergency events corresponding to each grid cell based on the fused features. This allows for comprehensive consideration of multiple dimensions such as time, space, weather, traffic, and society for event prediction. Furthermore, by combining this with an emergency medical equipment coverage model, it determines the deployment information of emergency medical equipment in the target area, enabling more reasonable and accurate planning of emergency medical equipment.

The invention discloses a network resource scheduling method and device, equipment and a storage medium, and relates to the technical field of wireless communication. The method comprises the following steps: acquiring network state sensing data of a target wireless network; according to the network state sensing data, network environment cognitive information corresponding to the target wireless network is determined, and the network environment cognitive information comprises physical environment information, equipment portrait information and network event prediction information; and determining a network parameter configuration action according to the network environment cognitive information, and allocating schedulable network resources of the target wireless network according to the network parameter configuration action. By using the method, comprehensive, deep and dynamic cognition of the target wireless network environment is realized by determining the network environment cognition information, prospective and active network resource scheduling is realized by determining the network parameter configuration action according to the network state perception data, the accuracy and effectiveness of resource allocation are improved, and the resource allocation efficiency is improved. And thus, the requirements of high-quality and low-delay network connection are met.

The invention belongs to the technical field of meteorological prediction, and discloses an abnormal air temperature event prediction method, which realizes early recognition and early warning of an abnormal air temperature event by introducing a recurrent neural network and fusing historical air temperature observation data and geographical location information. The method comprises the following steps: firstly, automatically identifying and extracting an abnormal temperature event from large-scale meteorological observation data, and constructing a high-quality data set suitable for prediction modeling; secondly, performing dynamic feature extraction on the temperature time sequence by using a recurrent neural network, and capturing a time sequence evolution rule of the abnormal temperature; and finally, in combination with the geographic position and climate relevance, screening key places having relevance with the to-be-predicted place, establishing a spatial enhancement mechanism, and optimizing model input features. According to the method, effective fusion of the time dynamic features and the space correlation features is realized, and the accuracy of abnormal temperature event prediction and the model robustness are remarkably improved.

The application provides a high-concurrency working condition processing system and method based on a distributed system, and relates to the technical field of cloud computing. The method comprises the following steps: acquiring system running state parameter information of the distributed system; determining a current concurrency working condition level and a current high-concurrency response mode of the distributed system within a preset period according to the system running state parameter information and a high-concurrency identification threshold; inputting the system running state parameter information into a pre-trained concurrency event prediction model to obtain a high-concurrency working condition level change sensitivity of the distributed system, wherein the concurrency event prediction model is obtained by training historical concurrency event data of the distributed system; and determining a corresponding high-concurrency response mode of the distributed system within the preset period according to the high-concurrency working condition level change sensitivity, the current concurrency working condition level and the current high-concurrency response mode.