231 results about "Confidence interval" patented technology

An approach to time-series data point anomaly detection may be presented. Data point anomalies in time-series data can cause a cascade of incorrect predictions in a time-series data prediction model. Presented herein may be an approach to decompose a time-series training data set into elementary components, such as seasonal, trend and residual. The approach may determine one or more confidence intervals for elementary components of data points including level shift, variance, and outlier. From these confidence intervals, new data points can be analyzed and identified as anomaly data points. The approach may also prevent anomaly data points from being incorporated into a time series data prediction model, reducing prediction error in the prediction model.

The invention relates to the related technical field of risk indicator construction management, in particular to an evidence-based migraine risk indicator construction system, which comprises a feature extraction module for acquiring physiological data and coupling abnormal features, an error correction module for mapping features and correcting errors, a dynamic adjustment module for adjusting a probability confidence interval, and an evidence-based migraine risk indicator management module for establishing an evidence-based migraine risk indicator. The technical problems that generation of the risk report is fixed to a preset migraine risk index, the risk report cannot adapt to individual differences, and the accuracy of risk assessment is limited are solved, physiological feature data are extracted, coupling abnormal features are analyzed, and the accuracy of risk assessment is improved. The method has the advantages that physiological changes before migraine attack are reflected more comprehensively, reliability of risk assessment is improved, historical intervention cases are introduced, confidence intervals of risk probability values are dynamically adjusted, dynamic adjustment is performed according to individual differences of patients and intervention response conditions, and adaptability of migraine risk reports is improved.

The invention provides a cigarette abnormal flow standard risk prediction method and system. The method specifically comprises the steps of collecting multi-source data to construct a causal feature set, generating a causal graph according to domain knowledge and algorithm mining, quantifying an average processing effect through a dual machine learning method, constructing a risk model to position a root cause, simulating risk change after intervention through anti-fact inference, and outputting a result. Through the dual machine learning and anti-fact inference technology, the causal effect is accurately quantified, the intervention effect is simulated, the confidence interval is dynamically adjusted, and the risk prediction and decision support capability can be remarkably improved.

The invention relates to the technical field of information retrieval, in particular to an artificial intelligence digital science and technology platform based on big data analysis, which comprises a data uncertainty modeling module for receiving a big data stream, determining a confidence interval boundary for each data item to represent the uncertainty of the data item, generating a data item uncertainty parameter set, and combining each data item with the corresponding data item uncertainty parameter set, and then storing the data item and the corresponding data item uncertainty parameter set to form a structured probability data record. According to the method, by independently establishing an uncertain parameter structure for each data item in the data stream, tight binding of the data items and the corresponding uncertain parameters is realized, a structured probability data record is generated, the precision and reliability of the data in the query and calling process are guaranteed, and retrieval errors caused by isolated storage of the data items are avoided; through probability attribute grouping of a preset space and a numerical range and establishment of a multi-dimensional probability index structure, the search space and the filtering difficulty of probability data are reduced.

The invention relates to a fault early warning method, device and equipment for an energy storage system. The method comprises the following steps: acquiring target data of a target parameter; the target data is generated by preprocessing real-time operation data and real-time environment data of the energy storage system; extracting a target feature corresponding to each target parameter based on the target data; calculating a feature confidence interval of each target parameter based on the historical data of the target parameters, and marking suspected abnormal features based on the feature confidence intervals; identifying at least two fault types based on the suspected abnormal features; based on the current environment data, the load power of the energy storage system and the historical operation and maintenance data of the energy storage system, carrying out fuzzy reasoning on the risk membership degree corresponding to each fault type and dynamically adjusting the basic weight coefficient corresponding to each fault type; and determining a comprehensive risk index based on the risk membership degree corresponding to each fault type and the dynamically adjusted dynamic weight coefficient, and performing fault early warning analysis processing based on the comprehensive risk index. The method can improve the accuracy of fault early warning.

The invention belongs to the technical field of wind power weather forecasting, and provides an all-region three-dimensional wind speed correction method and system, and the method comprises the steps: constructing a weather numerical forecasting model, and obtaining wind field forecasting data; fusing the preprocessed multi-source data by adopting an optimal interpolation method to obtain three-dimensional space-time continuous wind field analysis data; based on the wind field forecast data and the wind field analysis data, features are extracted and fused, then a historical forecast error sample set is constructed, a wind speed correction model is constructed, and the historical forecast error sample set is utilized to train the wind speed correction model; introducing an initial value, a physical parameter and boundary condition disturbance, calculating a mean value and a standard deviation of each set result, extracting a probability distribution feature of a wind speed, constructing a confidence interval, estimating a probability density function, and quantifying an occurrence probability of an extreme wind speed event; and the prediction result of the wind speed correction model and the multi-source wind field observation data are fused to generate final three-dimensional wind field data, so that the actual requirements of wind power prediction and power grid dispatching can be met.

The invention relates to a power capacity evaluation method based on a dynamic confidence interval and double-layer optimization, which solves the problems of inappropriate single-side modeling, static evaluation, source-load coupling modeling data processing mode and the like in the prior art, and adopts the technical scheme that a prediction error of wind power output is obtained based on historical data and a prediction value of the wind power output; based on the prediction error, determining an influence parameter generating the error; creating a power supply side model based on the prediction error and the influence parameter; obtaining a demand response sensitivity coefficient and a time-varying adjustment parameter based on the historical load data, the demand response execution record and the external influence factor; creating a load side model based on the demand response sensitivity coefficient and the time-varying adjustment parameter; creating a source-load joint scene model based on the power supply side model and the load side model; optimizing the source-load joint scene model through a double-layer robust optimization model; an assessment risk probability index and a capacity vacancy expectation are determined based on the optimized source-load joint scenario. The method has the advantages that a source-load joint scene model is created, and the coverage range and precision of an evaluation result in time and space dimensions can be remarkably improved; meanwhile, by means of a dynamic confidence interval adjustment mechanism and a double-layer optimization strategy, effective balance between the electric power market economy and the operation safety is achieved.

The invention discloses a power equipment health evaluation method and system based on time sequence analysis and probability modeling, equipment and a medium, and relates to the technical field of power equipment state monitoring and fault prediction, and the method comprises the steps: obtaining and preprocessing multi-source operation data of power equipment, outputting a predicted value and a confidence interval of a future parameter through a time sequence prediction model, calculating a residual sequence of an actual observation value and a predicted value, fitting distribution through a probability distribution model, establishing a statistical characteristic model of a normal operation state of the equipment, performing anomaly judgment, calculating a health degree index of the equipment based on a deviation degree and a dynamic weight of a monitoring parameter and weighted accumulation, and dividing equipment state grades according to the index. Quantitative evaluation of the health state of the equipment is realized. According to the method, accurate quantification and early abnormity identification of the health state of the power equipment are realized, a reliable basis is provided for predictive maintenance, and the intelligent level and the safety guarantee capability of power grid operation and maintenance are remarkably improved.

The invention discloses a water level time sequence data cleaning method based on regression statistics, which comprises the following steps: S1, sorting flood data of a historical session of a monitoring station to form a complete flood sample set; s2, establishing a function model according to the nonlinear change characteristics of the flood water level; s3, performing regression function fitting on the historical flood sample data, and calculating linear correlation between each fitting function and actual data; s4, based on a fitting result of the optimal regression function, establishing a Gaussian distribution model of a difference value; s5, determining a confidence interval of Gaussian distribution through the verification sample, and judging whether the data is abnormal data or not according to the confidence interval; s6, the detected abnormal data are corrected, and the cleaned data are obtained; and S7, processing the first N time sequence points in the flood time sequence data by adopting a moving average algorithm, and eliminating the interference of accidental fluctuation on regression fitting. The method has the advantages of being high in cleaning precision, high in adaptability and good in flood nonlinear change characteristic processing effect.

The invention discloses a water quality probability forecasting method based on Bayesian multi-time-sequence deep learning. The method comprises the following steps: S1, determining a forecasted water environment water ecological index, a driving index and a forecasting day number; s2, collecting time sequence data monitored by the forecasting indexes and the driving indexes, and after data preprocessing, constructing a data set required by model construction; s3, carrying out data division on the time sequence data, constructing a driving index forecasting model by adopting a multi-time sequence deep learning method, and carrying out parameter learning by selecting a Bayesian random discarding method; s4, performing effect evaluation on the accuracy and precision of the model, and adopting a hyper-parameter optimization method to improve the simulation forecast effect; s5, carrying out model training by adopting all data without segmenting the training set and the test set, carrying out water quality probability forecasting by utilizing the trained model, and outputting a forecasting mean value and a confidence interval; according to the method, the confidence interval is output while high-precision prediction is provided, and the scientificity and stability of prediction are improved.

The invention discloses a complex water quality parameter prediction method and system based on deep learning, and the method comprises the steps: S1, collecting and preprocessing original water quality parameter data, and obtaining the preprocessed water quality parameter data; s2, extracting a water quality parameter data point cluster from the preprocessed water quality parameter data by adopting an improved DBSCAN algorithm; s3, extracting a feature fusion vector, and extracting a coupling feature vector in combination with the water quality parameter data point cluster; s4, calculating self-adaptive scale and multi-scale water quality parameter characteristics of the gating circulation unit, and then performing water quality parameter prediction; and S5, adding Gaussian noise into the preprocessed water quality parameter data, carrying out simulation calculation, and calculating a pH value predicted value confidence interval, an ammonia nitrogen concentration predicted value confidence interval and a chemical oxygen demand predicted value confidence interval according to a simulation calculation result. The method can solve the problem that a traditional water quality parameter prediction method is difficult to deal with the poor prediction effect caused by coupling association of various water quality parameter types.

A method for real time determination as to whether a prediction model output should be included or withheld from a real time data series based on a combination of a confidence interval associated with the prediction model output and a classification of the prediction model output.

The invention discloses an interpretable multi-dimensional CI index dynamic scoring and risk assessment method, and belongs to the technical field of compliance risk assessment. According to the method, firstly, multi-source heterogeneous compliance data are collected, cleaned, subjected to caliber alignment and subjected to quality evaluation, and standardized data are obtained; a hierarchical index semantic system is constructed based on the standardized data, and a core index set is obtained through multi-dimensional screening; then, constructing a multi-dimensional dynamic threshold, calculating a weighted score in combination with the core index set, quantifying a score confidence interval through a Bayesian confidence quantification model, and mapping a risk level to form complete score data; and finally, carrying out interpretability analysis on the complete scoring data, and generating hierarchical interpretation texts and visual display contents. According to the method, systematization, precision and transparency of compliance risk assessment are realized, the suitability and credibility of a scoring result are improved, and reliable support is provided for compliance decision making.

A system and methods for multivariant learning and optimization repeatedly generate self-organized experimental units (SOEUs) based on the one or more assumptions for a randomized multivariate comparison of process decisions to be provided to users of a system. The SOEUs are injected into the system to generate quantified inferences about the process decisions. Responsive to injecting the SOEUs, at least one confidence interval is identified within the quantified inferences, and the SOEUs are iteratively modified based on the at least one confidence interval to identify at least one causal interaction of the process decisions within the system. The causal interaction can be used for testing, diagnosis, and optimization of the system performance.

The invention provides an abnormal index determination method and device, equipment and a storage medium, and the method comprises the steps: carrying out the evaluation of target parameter data based on a deep learning model, and obtaining a first evaluation result; wherein the target parameter data is any one kind of parameter data in various kinds of parameter data of the target equipment; evaluating the target parameter data based on a statistical model to obtain a second evaluation result; evaluating the target parameter data based on the feature extraction model to obtain a third evaluation result; fusing the first evaluation result, the second evaluation result and the third evaluation result according to a preset weight to obtain a target evaluation result; and if the target evaluation result exceeds the target confidence interval, determining the parameter item corresponding to the target parameter data as an abnormal index. By applying the technical scheme of the invention, the problems that the representativeness of data features extracted by a single feature extraction model is insufficient, and the monitoring result of abnormal indexes is inaccurate can be solved.

The invention discloses a five-directional stressometer monitoring data anomaly identification method based on PCA confidence interval analysis, and relates to the technical field of hydropower engineering. According to the method, circle centers (zt1, zt2) and radiuses r1r2 of two-dimensional coordinate data points in each time period are obtained, all the circle centers are subjected to arithmetic average to obtain a final circle center, all the radiuses are subjected to arithmetic average to obtain a final radius, and therefore a global confidence circle is formed; and when new data appear, performing the same PCA transformation and coordinate mapping on the new data, and judging whether a new point is in the final confidence circle so as to realize anomaly detection. According to the method, data dimensionality reduction is performed on data of each monitoring point of the five stress meters by adopting a principal component analysis method, and a final confidence circle is formed by performing arithmetic averaging on two-dimensional circle center coordinates and radiuses of monitoring data in multiple periods; judging whether the data of each monitoring point is normal or not by judging whether the circle center coordinate of the new confidence circle is in the final confidence circle or not; the whole method is suitable for hydropower engineering dam monitoring, equipment diagnosis and data analysis scenes of rapid anomaly detection.

Devices, systems, methods, and processes for managing network devices through generated predictions and associated confidence levels are described herein. Networks within a floorplan can be operated at full capacity all day in an inefficient way when not adjusted due to traffic patterns and seasonality changes. Data related to the topology of the network, along with historical data can be utilized to generate predictions of various network needs. For example, the overall network throughput capacity needs may be predicted for a series of points in the future. An associated confidence level can be generated as well including one or more confidence intervals. These can be utilized to select a future need for the network and generate a corresponding sustainable network configuration for the network devices and / or their transceivers that can provide sufficient network needs while minimizing the overall power used. This can be automated over time once trust has been established.

The power system inertia prediction method based on the variational Bayesian attention normalization stream comprises the following steps: constructing a system inertia data set; preprocessing the system inertia data set, and dividing the preprocessed data set into a prediction set and residual data according to a time range; respectively generating Q, K and V by adopting variational Bayes, carrying out relative position coding on the generated vectors, and carrying out weighted fusion on the Q, K and V after position coding by utilizing a multi-head attention mechanism; inputting the output of the multi-head attention mechanism into an attention normalization flow model, and capturing complex probability distribution of data through reversible transformation; carrying out model training by adopting a self-defined mixed loss function; and performing multiple Monte Carlo sampling on the trained model to obtain a sampling prediction set, calculating a prediction mean value and a standard deviation based on the sampling prediction set to obtain a significance level, and constructing an interval prediction result under a corresponding confidence interval. According to the prediction method, accurate probability prediction of the inertia of the power system is realized.

The invention discloses a new energy uncertainty interval fine modeling method based on ELM-Copula. Comprising the following steps: 1) data reading and preprocessing: obtaining a power prediction value and an actual output value by reading historical operation data of a new energy electric field, performing kernel density estimation on the per-unit power prediction value and the actual output value, and calculating an error absolute value according to the per-unit power prediction value and the actual output value; 2) constructing a dynamic Copula function model; 3) measuring goodness of fit of the model; 4) calculating a confidence interval of a prediction error, analyzing uncertainty of power generation power prediction, and giving a confidence interval of a prediction value; according to the method, a multi-type dynamic Copula model is introduced to construct a dynamic dependency structure between a prediction error and power, the ELM is applied to a post-processing stage of a dynamic Copula prediction interval, a correction coefficient is generated by learning historical deviation characteristics, and an original interval is shrunk, so that the prediction precision and practicability are improved on the premise that the coverage rate is not reduced.

The invention relates to a method for obtaining credible blasting vibration dominant frequency, which solves the problem of dominant frequency prediction under complex geology through EMD-HHT multi-scale decomposition and statistical confidence interval analysis and is difficult to adapt to dynamically changing construction conditions. EMD decomposition adaptively eliminates low-frequency noise caused by nonlinear response of the rock mass, and the signal-to-noise ratio is increased; a main frequency determination strategy is automatically switched through confidence interval width (RW) to adapt to complex geological conditions; 95% confidence coefficient ensures that a dominant frequency predicted value accords with actual energy distribution; compared with a traditional Samsu formula, the main frequency prediction error is reduced by more than or equal to 40%.

The invention discloses a carbon emission dynamic accounting method and system. The method comprises the following steps: collecting production equipment direct emission, logistics carbon emission, social media / satellite remote sensing and other abnormal event data; extracting unit yield carbon intensity and supply chain path sensitivity through principal component analysis; constructing a QUBO model and solving an optimal transportation path by using quantum annealing; training a prediction model in combination with the sudden emission risk index to output a carbon emission value; dynamically updating an emission factor library according to the cleanliness of the power grid, and accelerating generation of a confidence interval by adopting a quantum random number generator; and when the predicted value exceeds the threshold value and the lower limit of the confidence interval is higher than the threshold value, triggering carbon capture, carbon credit purchase or production regulation compensation according to the priority. The system correspondingly comprises a multi-mode acquisition module, a calculation optimization module, a dynamic accounting module and a carbon compensation module. Full-chain real-time carbon tracking and intelligent compensation are achieved, the problems of traditional accounting data isolation and compensation lag are solved, and the industrial carbon management accuracy is improved.

The invention discloses an EMB clamping force prediction method considering multi-source uncertainty, and belongs to the technical field of automobile industry, and the method comprises the steps: obtaining a historical mapping relation between clamping force and control input, building prior distribution, and employing normal distribution as the prior distribution to describe an expected range of the clamping force; the method comprises the following steps: acquiring observation data of clamping force historically collected by a sensor, introducing the observation data into a likelihood function, and setting a relationship between the observation data and real clamping force to be influenced by multi-source noise to generate an error model so as to describe probability distribution of the observation data under specific clamping force by maximizing the likelihood function; on the basis of a Bayesian inference framework, updating the trust degree of the clamping force predicted by the EMB through prior distribution and a likelihood function, and calculating posterior distribution by using known prior information and real-time observation data; and extracting 95% confidence interval and uncertainty index in the posterior distribution, and when the confidence interval or the system entropy value is greater than the uncertainty index, rejecting or optimizing the prediction result of the clamping force predicted by the EMB.

The invention discloses an electromagnetic wave rain attenuation model selection method based on a confidence interval upper bound, and belongs to the technical field of electromagnetic wave propagation loss. According to the method, an improved confidence interval upper bound reinforcement learning algorithm is adopted, an action space is defined as a plurality of different rain attenuation models, and the system can dynamically select an optimal model from a plurality of candidate models by inputting rainfall rate information; the reward function is based on an absolute error between a unit distance electromagnetic wave rain attenuation value predicted by each model and an actual observation value, the smaller the error is, the higher the reward is, otherwise, the lower the reward is. Through continuous iterative training, the method can adaptively select the rain attenuation model with the highest prediction precision in different environments, thereby improving the accuracy of electromagnetic wave propagation prediction.

The invention provides a short-term power load prediction method and device, a storage medium and electronic equipment, and the method comprises the steps: determining key feature factors according to historical data and historical fluctuation conditions; corresponding initial data are collected and preprocessed, and preprocessed data are obtained; carrying out parallel training on an initial Prophet algorithm and an initial LightGBM model by adopting the preprocessed data, adjusting operation parameters of the two models based on prediction precision to obtain two corresponding models, and finally outputting a corresponding confidence value and a confidence interval; according to the method, two loss function penalty terms are added, the two models are integrated, a final load prediction result is obtained according to confidence values and confidence intervals of the two models, the two models are trained in parallel, a prediction value optimization module is added, penalty terms of a LightGBM algorithm are introduced, the training effect of each model is guaranteed, and the load prediction efficiency is improved. And the prediction value has higher precision on the whole.

The invention discloses a multi-step time sequence prediction method and system based on double-segmentation conformal prediction. The method comprises the following steps: acquiring new input data; distributing the new input data to a corresponding clustering cluster to obtain a clustering result; extracting information from the recorder according to a clustering result to construct a prediction interval; adjusting the corresponding content according to the prediction interval; wherein a similar trend sequence is vertically classified and clustered, errors of adjacent time steps are horizontally and dynamically combined to optimize window division, over-estimation and under-estimation errors are asymmetrically processed to construct a precise confidence interval, the precise confidence interval is stored in the recorder, and an error set is dynamically updated. By implementing the method provided by the invention, the defects in the prior art can be overcome through a two-dimensional segmentation mechanism, more accurate uncertainty quantization is realized, and the adaptability and accuracy of multi-step time sequence prediction are improved.

Selecting an optimal model by acquiring, via a predictive analytics engine of a learning machine, an input dataset, and receiving a number of possible regression models for selection, the input dataset includes a plurality of labeled cases. Each candidate regression which is generated through a forward selection procedure is fitted to the input dataset to describe a relationship between one or more explanatory variable values and response variable values of the input dataset. A predictive power of the possible regression model is measured by computing a usual square coefficient of multiple determination, and either a point estimate of the square cross-validated correlation or a two-sided confidence interval of the square cross-validated correlation associated with the given regression sample. Based on the predictive power, the possible regression model that meets a predictive power threshold is selected as an optimal regression model.

The invention discloses a data-mechanism-knowledge combined driving dam safety state self-diagnosis method and system, and the method comprises the steps: obtaining dam operation state data, extracting a state quantity based on a pre-constructed computable evidence graph, and generating a monitoring evidence; performing inversion correction on mechanism model parameters under the priori constraint of the atlas, and generating mechanism evidence; the monitoring evidence and the mechanism evidence are accessed to a probability graph model for joint inference, and arbitration correction is carried out when a conflict triggering condition is met; and carrying out credibility calibration on the diagnosis result by adopting a conformal prediction method based on the historical calibration sample set, and outputting a security state diagnosis conclusion containing a confidence interval. According to the method, the problems of multi-source evidence conflict, mechanism parameter distortion and low diagnosis result credibility are effectively solved, and credible and automatic diagnosis of dam safety is realized.

A method can include accessing data for a number of reservoir sites, where the data include at least property data for reservoir properties; performing a determination, using the property data for the reservoir sites, as to whether the reservoir properties for the reservoir sites are independent; responsive to the determination, implementing a Sobol' indices technique or a Kucherenko indices technique to generate global sensitivity analysis results that include property indices for the reservoir properties; generating confidence intervals for the property indices; and generating a graphical user interface for rendering the property indices with the confidence intervals for a number of the reservoir properties.

The invention relates to a method for determining a confidence interval for an estimated position of an object and a method for training a physics-informed neural network for this purpose. The invention further relates to a computer program, a device, and a storage medium for this purpose.

A continuous asymmetric risk analysis system and method of operating the same. In one embodiment, the method includes receiving a definition of a risk event of a complex system based on a likelihood and consequence, and prevention and mitigation measures for the risk event, and receiving risk event data. The method also includes creating a model populated with the risk event data, and executing Monte Carlo simulations of the model to produce Monte Carlo results. The method also includes analyzing and aggregating the Monte Carlo results of the likelihood and consequences, and the prevention and mitigation measures for the risk event to create a nominal risk value of the risk event and asymmetric confidence intervals to produce a continuous gradient of outcomes. The method also includes organizing and presenting the outcomes for evaluation, and prescribing an action for the complex system based on a selected outcome.