43 results about "Heteroscedasticity" patented technology

The invention discloses a joint probability density prediction method of short-term output power of a plurality of wind power plants. The method comprises the following steps: carrying out single point value prediction on output power of each wind power plant by using a support vector machine regression prediction model; building a sparse bayesian learning model as to a prediction error to carry out probability density prediction of the error, so as to obtain an expected value and a variance of marginal probability density function prediction of the output power of a single wind power plant; carrying out statistic analysis on the prediction error characteristics of the output power of the plurality of wind power plants, building a dynamic conditional correlation-multivariate generalized autoregressive condition heteroscedasticity model, and integrating a marginal probability density prediction result of the output power of the single wind power plant and a correlation coefficient matrix to obtain a joint probability density function of the output power of the plurality of wind power plants; forming a multidimensional scene including space-time correlation characteristics by using a sampling technique. By adopting the joint probability density prediction method, a mean prediction value and prediction uncertainty information of the output power of the single wind power plant can be provided; the dynamic space-time correlation characteristics between output power prediction of the plurality of wind power plants also can be quantitatively described.

The invention relates to a robot walking control method based on a foothold compensator. The robot walking control method comprises the following steps of firstly, establishing a constraint dynamic model of a robot; secondly, designing the foothold compensator based on the constraint dynamic model according to the constraint dynamic model; thirdly, establishing a heteroscedastic sparse Gauss process regression model and realizing the mapping calculation from input to output of the foothold compensator; fourthly, locally updating the heteroscedastic sparse Gauss process regression model; fifthly, establishing the foothold compensator based on the heteroscedastic sparse Gauss process regression model; sixthly, performing prediction control over the walking of the robot according to the foothold compensator based on the heteroscedastic sparse Gauss process regression model. Compared with the prior art, the robot walking control method disclosed by the invention has the advantages of accurate prediction, high learning speed and the like.

The embodiment of the invention provides a satellite long-period heteroscedasticity degradation prediction and evaluation method based on GRU and GARCH, and the design idea comprises: firstly carryingout preprocessing and time sequence decomposition on an original parameter for the long-term degradation and heteroscedasticity characteristics of telemetry parameter data collected by a sensor, andpredicting a trend term through a GRU model. In order to solve the problem of long-term degradation, the GARCH model predicts residual terms to solve the problem of heteroscedasticity; and a satelliteparameter prediction result is obtained in combination with a satellite seasonal period rule. Meanwhile, normal fluctuation threshold information of the satellite is extracted from the residual termand is combined with the season term and the trend term, so that construction of the satellite stability consistency adaptive threshold is realized, and the satellite stability consistency health assessment method is provided based on the adaptive threshold. The method can accurately predict the telemetry data which is greatly influenced and fluctuated by the satellite environment and multiple tasks, and the threshold value can be updated on line and is more effective and accurate compared with a traditional method.

The invention discloses a real-time overbounding method for civil aviation satellite navigation integrity augmentation system errors. According to the method, a time sequence model is utilized to analyze the temporal correlation of pseudo-range errors, fat-tail error samples without the temporal correlation are obtained, and overbounding calculation is performed after normalization processing is performed on the fat-tail error samples through a generalized autoregressive conditional heteroscedasticity (GARCH) model, so that a stationarity model of temporal correlation errors is established; and a confidence upper limit of a standard deviation of the pseudo-range errors is determined, optimal pseudo-range error sequence length is determined, and therefore processing of the uncertainty and error non-stationarity of the model is realized. Through the method, an extra integrity risk cannot be introduced; and due to tighter overbounding, the protection level calculated in real time is lowerthan that calculated through a traditional method, and consequently the availability of a satellite navigation augmentation system is improved.

The invention discloses a joint probability density prediction method of short-term output power of a plurality of wind power plants. The method comprises the following steps: carrying out single point value prediction on output power of each wind power plant by using a support vector machine regression prediction model; building a sparse bayesian learning model as to a prediction error to carry out probability density prediction of the error, so as to obtain an expected value and a variance of marginal probability density function prediction of the output power of a single wind power plant; carrying out statistic analysis on the prediction error characteristics of the output power of the plurality of wind power plants, building a dynamic conditional correlation-multivariate generalized autoregressive condition heteroscedasticity model, and integrating a marginal probability density prediction result of the output power of the single wind power plant and a correlation coefficient matrix to obtain a joint probability density function of the output power of the plurality of wind power plants; forming a multidimensional scene including space-time correlation characteristics by using a sampling technique. By adopting the joint probability density prediction method, a mean prediction value and prediction uncertainty information of the output power of the single wind power plant can be provided; the dynamic space-time correlation characteristics between output power prediction of the plurality of wind power plants also can be quantitatively described.

A computer-implemented method includes employing a dynamic Boltzmann machine (DyBM) to predict a higher-order moment of time-series datasets. The method further includes acquiring the time-series datasets transmitted from a source node to a destination node of a neural network including a plurality of nodes, learning, by the processor, a time-series generative model based on the DyBM with eligibility traces, and obtaining, by the processor, parameters of a generalized auto-regressive heteroscedasticity (GARCH) model to predict a time-varying second-order moment of the times-series datasets.

A computer-implemented method includes employing a dynamic Boltzmann machine (DyBM) to predict a higher-order moment of time-series datasets. The method further includes acquiring the time-series datasets transmitted from a source node to a destination node of a neural network including a plurality of nodes, learning, by the processor, a time-series generative model based on the DyBM with eligibility traces, and obtaining, by the processor, parameters of a generalized auto-regressive heteroscedasticity (GARCH) model to predict a time-varying second-order moment of the times-series datasets.

The invention discloses a water level flow relation curve deducing method based on the Copula function. The method comprises the following steps: collecting a section water level and flow data materials, constructing a joint probability distribution function of the water level and flow by using the Copula function based on determining an marginal probability distribution function, solving a condition probability distribution function of the flow at a specified water level, and deducing a water level flow relation curve according to the mathematical statistical principle, and performing uncertainty analysis. The water level flow relation curve deducing method has relatively high statistical theory foundation, allows the edge distribution of the water level and the flow in any form, and canalso accurately describe the nonlinear and heteroscedasticity correlation structure of the water level and the flow. In addition, a point estimation value of the flow can be obtained, the uncertaintyof model parameters and model structures can be considered more comprehensively, and a comprehensive uncertainty interval of the flow is obtained

The invention discloses a method for extracting a reservoir dispatch function based on joint probability distribution. By collecting basic reservoir data and long series of measured runoff data, the reservoir is deterministically optimized and dispatched to obtain an optimal dispatch process sample series, and decision variables and state variables are determined. Then use the Copula function to construct the joint probability distribution function of the decision variable and the state variable, and then infer the conditional probability distribution function of the decision variable when the state variable is given, and perform the extraction and uncertainty analysis of the reservoir dispatching function. The reservoir dispatching function extracted by the invention can accurately capture the inherent nonlinear and heteroscedastic correlation structure between decision variables and state variables, and the calculated confidence interval estimation value can quantitatively evaluate the uncertainty of reservoir dispatching, which provides useful information for decision-making risk analysis. reference basis.