93 results about "Quantile regression" patented technology

The invention relates to the technical field of gait analysis in biomedical engineering and provides a human gait analyzing method and system based on multi-sensor fusion. The method comprises the steps of filtering sensor signals to eliminate the signal noise error according to human motion features, and eliminating the integral error by means of the improved zero velocity updating algorithm, so that the method and system can be adapted to different walking scenes; fusing multiple sensor data with the Denavit-Hartenberg method, and reducing the leg position calculation error; calculating the step speed, step length, step frequency, walking period and walking track of a tested person accurately during walking after error correction is conducted; establishing a gait database, and conducting statistic analysis on gait data of different tested persons with the quantile regression analysis method. By the adoption of the method and system, gait parameter measurement precision can be improved, and gait parameters of different tested persons are comparable through standardization.

The invention discloses a power system short-term load probability forecasting method, a device and a system. The short-term load probability density forecasting model of Gaussian process quantile regression is established by selecting an optimal input variable set affecting the load. Firstly, the importance score of input variables is given by stochastic forest algorithm, and the influence degreeof each input variable is sorted. Secondly, particle swarm optimization algorithm is used to search the super-parameters of the model to form the optimal Gaussian process quantile regression prediction model, avoiding the adverse effect of artificial experience setting initial parameters on the prediction performance of the model. The invention can avoid the shortcomings of manual experience selection, the load forecasting model established in the optimal input variable set has low error, which further reduces the forecasting error, and overcomes the problems that the common conjugate gradient method is easy to fall into the local optimal solution, the iterative number is difficult to determine, and the optimization performance is greatly affected by the initial value selection, so that the self-searching and group cognitive ability can be brought into full play.

The invention discloses a data index exception monitoring method and system, a storage medium and electronic equipment. The data index exception monitoring method comprises the following steps that: obtaining a data index which needs to be monitored and the historical data of an associated data index corresponding to the data index which needs to be monitored; according to the data index which needs to be monitored and the historical data of the associated data index, generating a scatter diagram; according to the data index which needs to be monitored, the historical data of the associated data index and the scatter diagram, through a quantile regression algorithm, selecting an upper tail, a lower tail and a median, and taking the upper tail, the lower tail and the median as parameters togenerate a learnt monitoring sample model through machine learning; and comparing a data difference between the data index which is obtained in real time and needs to be monitored and the monitoringsample data, and carrying out exception detection on the data index which needs to be monitored, wherein the exception detection on the data index which needs to be monitored comprises exceptional fluctuation point detection and exceptional fluctuation tendency detection.

The invention relates to a harmonic contribution division method and a harmonic contribution division system. The harmonic contribution division method comprises the following steps of acquiring harmonic voltage data of a bus and harmonic current data of a harmonic source to be calculated on a feeder line; calculating background harmonic impedance by using a leading fluctuation quantity method according to the harmonic voltage data and the harmonic current data; and dividing harmonic contributions of the harmonic source to be calculated by using a quantile regression method according to the harmonic voltage data, the harmonic current data and the background harmonic impedance. The background harmonic impedance is estimated by the leading fluctuation quantity method, fluctuation quantity with a leading function is screened out to calculate the background harmonic impedance, influences of background harmonic and measurement noise fluctuation on a background harmonic impedance estimation result are restrained effectively, and the background harmonic impedance is calculated accurately; and the background harmonic current is calculated according to the background harmonic impedance, and quantile regression is performed to obtain the harmonic contributions of the harmonic source. Calculation deviation caused by background harmonic fluctuation can be reduced, division accuracy is improved, and the stability and the data utilization rate are high.

The invention discloses a method for predicting the short-term wind power probability density based on the EWT quantile regression forest. The method comprises the steps of 1) decomposing an originalwind power sequence into a series of mutually different feature empirical modes by using the empirical wavelet transform (EWT); 2) recombining the empirical modes according to a frequency range to form high frequency, intermediate frequency and low frequency components; 3) select an input variable for each component by using the Pearson correlation coefficient; 4) establishing a quantile regression forest prediction model for each component, and obtaining regression prediction results of different quantile points; 5) superposing the prediction results of the components to obtain a wind power prediction value; and 6) obtaining the prediction of the wind power probability density by nuclear density estimation. The method provided by the invention effectively improves the prediction precisionof the wind power, obtains the prediction of the wind power probability density at any moment, and can well solve the wind power prediction problem in a power system.

The invention discloses a novel analytical method of deformation monitoring data of a dam slope. Dam slope displacement and a factor influencing the dam slope displacement are monitored, monitoring data are selected, a regression analysis based on proper orthogonal decomposition is carried out to obtain a feature orthogonal basis of proper orthogonal decomposition; quantile regression analyses are carried out on the feature orthogonal basis and the damp slope displacement to obtain regression equation parameters of all quantiles, and then original data substitution is carried out to obtain a quantile regression equation of the original data; and then changing situations of regression equation estimation values of independent variables under different quantiles as well as whether the regression equation estimation values pass the significance test successfully are analyzed and the influence degrees on the dam slope displacement by the independent variables are determined. According to the invention, the method has the great significance in early-warning model establishment for a slope of a dam, slope prediction and forecasting method providing, and slope sensitive influence factor analyses.

The embodiments herein disclose a system and method for particle swarm optimization and quantile regression-based rule mining for analyzing data sets involving only continuous explanatory variables. The system discloses an architecture for PSO based quantile regression rule mining for determining the prediction intervals (PIs). The system generates ‘if-then’ rules that yield PIs while solving a multiple regression problem having only continuous explanatory variables. The system performs an ensembling process to reduce the size of the rule base to a manageable number based on the quality metrics of prediction intervals. The system comprises a data set, and a rule miner designed to divide the data into deciles based on the descending order of the target attribute variable. PSO is invoked to derive a set of rules for each decile and capture the heteroscedasticity of the distribution of the data with the help of quantile regression, in a non-traditional way.

The invention discloses a dynamic heat setting value probability distribution predication method of an overhead power transmission line based on quantile regression. The dynamic heat setting value probability distribution predication method comprises the following steps: taking an uncertain factor of a dynamic heat setting value at the next moment and a historical heat setting value as input variables, and establishing a quantile regression function model; determining a mean value and a median of an input sample, and solving a parameter estimated value of the quantile regression function model; substituting the parameter estimated value into the quantile regression function model, and carrying out quantile regression fitting on sample data by adopting a linear condition quantile to solve different quantiles of the dynamic heat setting values of the power transmission line; and carrying out continuous evaluation on the dynamic heat setting values of the power transmission line in the future moment to obtain complete probability distribution of the dynamic heat setting values of the power transmission line in the future moment. With the adoption of the dynamic heat setting value probability distribution predication method, the dynamic heat setting values in the future moment are predicated and the uncertainty of the dynamic heat setting values is described, so that a good point predication value can be obtained, and a fluctuation interval of the dynamic heat setting value is analyzed; and finally, the whole probability distribution is obtained.

A prediction modeling system and computer program product for implementing forecasting models that involve numerous measurement locations, e.g., urban occupancy traffic data. The system a data volatility reduction technique based on computing a congestion threshold for each prediction location, and using that threshold in a filtering scheme. Through the use of calibration, and by obtaining an extremal or other specified solution (e.g., maximization) of empirical volume-occupancy curves as a function of the occupancy level, significant accuracy gains are achieved and at virtually no loss of important information to the end user. The calibration use quantile regression to deal with the asymmetry and scatter of the empirical data. The argmax of each empirical function is used in a unidimensional projection to essentially filter all fully congested occupancy level and treat them as a single state.

The invention discloses a short time electric power load prediction method and device. The short time electric power load prediction method predicts an electric power load through integrating quantile regression and robustness extreme learning and using a hybrid prediction model established after optimizing a hybrid particle swamp algorithm (PSOGSA). The quantile regression uses multiple quantiles of history electric power data influence factors to obtain a quantile equation corresponding to a conditional distribution of electric power load prediction data of sometime in the future; stochastic disturbance of electric power data inputted in quantile regression can describe statistic distribution of prediction load values in detail without the need of making any hypothesis on distribution, which makes the whole prediction model strong in robustness; the robustness extreme learning machine is more robust in an abnormal load value; and the electric power load can be predicted by combining the two methods and using a hybrid model formed after optimizing the PSOGSA. The invention also provides a short time electric power load prediction device having the same beneficial effects.

The invention provides a new missing database data completion method. The method is characterized by comprising steps: 1, missing detection is carried out on a given data set; 2, dimension reduction of an input variable is carried out, correlation between input dimensions is analyzed, pivoting (PCA) is adopted to select a correlated input dimension, and a new input data set is formed; 3, training set k partitioning is carried out, a cluster (Kmeans) is used for carrying out partitioning on the input training set, and k classes of training sets are obtained; 4, a k plane regression function is built, the optimal regression coefficient and the geometric center of each plane are solved, and a regression fitting function is given; and finally, data completion test is carried out. The experiment proves that the data completion method is extremely effective; in an allowable error range, a completed database with a use value is obtained; and the challenging technical problem brought to machine learning and data mining due to data incompletion can be solved to a certain degree; and the big data application technology progress is pushed.

The invention discloses a high-speed passive distance measuring method based on oxygen absorption and a multiple regression, mainly relating to a distance measuring method. The method comprises the following steps: receiving a target radiation signal which is subjected to atmospheric attenuation by using an infrared spectrometer; estimating the distance value of a target to be measured by using a Beer-Lambert law in combination with a multiple linear regression algorithm. The advantages of a Fourier transform multichannel are utilized fully, target distance information can be calculated by only measuring target radiation once without specifically assuming a target moving state or continuously tracking and repeatedly sampling the target, a the measuring process can be finished instantly, the detection distance is long, and high measuring reliability is realized. Moreover, selected oxygen adsorbs wave bands around 762 nm, thereby well eliminating the weather influences, ensuring the measuring accuracy. Furthermore, the high-speed passive distance measuring method has a significant application prospect on the aspects of infrared search tracking system and electro-optical countermeasure.

The invention relates to an infrared spectrum quantitative analysis method and an infrared spectrum quantitative analysis device based on multi-scale regression. The infrared spectrum quantitative analysis device comprises a spectrometer connected with a data signal wire, a preprocessor, a wavelet decomposition and reconfiguration processor, and a partial least-squares regression model integrator; and the infrared spectrum comprises a mid-infrared and near-infrared spectrum, with the wavelength range of 780nm to 5000nm. By wavelet decomposition and reconfiguration transformation, the multi-model establishment is realized; the difficulty of extracting the spectrum signal information by a single-model method is overcome; by independently determining factor quantity on different sub-models, the aim of sufficiently extracting effective information can be realized, and the prediction precision and stability of infrared spectrum analysis model can be improved.

The invention relates to a route sector traffic probability density prediction method. The method comprises the following steps of selecting the traffic flow of a route sector in preset time as a sample, and analyzing sample data; and according to sample data analysis, combining model parameter selection to probabilistically predicting a route sector traffic demand, and acquiring a first prediction result. The route sector traffic probability density prediction method is used to predict based on route sector traffic flow historical data which can be obtained in an existing system. Through combining a neural network and a quantile regression method, the several quantiles of the continuous traffic demand data of a certain day in the future are obtained. And then, the continuous conditional quantiles are used to acquire the probability density function and the probability density curve graph of the continuous traffic demands of the certain day in the future through a nuclear density estimation method. A specific point prediction value and a variation interval can be obtained, and the probability of each value of a route sector traffic demand prediction change interval can also be obtained. And the accurate point prediction value of the day is acquired.

The invention discloses a wind power non-parametric interval prediction method based on self-adaptive double-layer optimization, and belongs to the field of renewable energy probability prediction. According to the method, an extreme learning machine and quantile regression are combined to carry out modeling on a prediction interval to form a lower-layer optimization problem; and the quantile level corresponding to the prediction interval is adaptively adjusted by taking the interval sharpness as a target to form an upper-layer optimization problem. Efficient and reliable training of the prediction model is realized by using a primal-dual interior point algorithm. The method does not need to depend on the priori hypothesis of wind power probability distribution, breaks through the centralsymmetry limitation of the traditional probability prediction on the interval quantile level, remarkably improves the reliability and sharpness of the prediction interval, and provides important reference for the operation and control of a high-proportion wind power system.

The present invention discloses a power system emergency material demand prediction method based on multiple regressions. The method comprises the steps consisting of modeling, normalization processing, multiple regression training, verification and the like. The power system emergency material demand prediction method based on multiple regression integrates factors such as local power grid scale, disaster event characteristics (disaster grade and the like), response grade and the like, builds a mathematics model according to the multiple regression method, obtains influence coefficients of emergency material prediction corresponding to each factor through adoption of historical data training, and rapidly calculates the demand for each emergency material after some disaster event happens through adoption of a prediction model to guide relevant departments to store or purchase materials according to the demand, so that it is ensured that the power network emergency maintenance is smoothly performed, and the fund waste caused by excessive storage or purchasing is avoided.

The embodiment of the invention discloses a power generation side electricity price quotation method and device based on artificial intelligence, a storage medium and electronic equipment, and belongs to the field of electric power measurement. First, artificial intelligence technology is used to predict node marginal electricity price of an integrated learning framework based on a decision tree. The power price distribution of different confidence intervals is obtained according to a quantile regression method, then a quotation decision-making model is constructed in combination with unit physical constraints and price scenes, optimal parameters of the quotation decision-making model are solved in combination with an intelligent technology, and the optimization speed is increased by introducing a momentum theory. Finally, according to different electricity price scenes, different quantity-cost-benefit schemes are established for comparison, so that an effective evaluation basis is provided for reasonable quotation making and risk management of a power generation side.

The invention belongs to the field of financial risk management, in particular to a time series probability distribution prediction method based on integrated empirical mode decomposition and nonlinear quantile regression. The method comprises the following steps: firstly, carrying out the integrated empirical mode decomposition on a financial price time series to obtain components with high regularity under different scales; secondly, independently predicting each component by 1-norm support vector machine quantile regression to obtain all quantile prediction results of each component; and thirdly, taking each quantile as a statistics target, independently adding the prediction result of each qnantile of each component, and integrating all prediction results to obtain the preduction result of each quantile so as to obtain the financial time series probability distribution prediction. The method provided by the invention can effectively predict a change probability of financial price, and can be applied to financial risk management and investment practice.

The invention discloses a wind power probability prediction method based on quantile regression. The method comprises the steps of step 1, conducting CEEMDAN decomposition on all original wind power sequences si(n); step 2, carrying out normalization processing on the wind power sequence data after CEEMDAN decomposition; 3, training the model to obtain predicted values of the wind power at different quantiles at each moment in a period of time in the future; and step 4, adopting a kernel density estimation method for the prediction value of each moment to obtain each probability density distribution so as to predict the future wind power complete probability distribution. According to the method, more useful information than point prediction can be obtained, and prediction of future wind power integrity probability distribution is realized.

The invention provides a short-term wind power prediction method based on a probability prediction model, and the method comprises the steps: 1) decomposing wind power historical data into a plurality of components through variational mode decomposition, building a leakage integral type echo state network model for each dimension of component for training prediction, reconstructing each prediction result, and obtaining a wind power point prediction value; 2) modeling the residual error of the point prediction by using an echo state quantile regression network to obtain residual error prediction values under different quantile conditions; and 3) integrating the point prediction value and the residual prediction value, and further improving the prediction precision by residual prediction on the basis of point prediction to obtain a probability prediction value of the wind power. According to the method, the wind power has the characteristics of randomness and volatility, the point prediction model and the residual prediction model are combined, the probability prediction model is obtained, the wind power is accurately predicted, and the method has great significance in ensuring safe, economical and stable operation of a power system.

The invention discloses an urban public service facility construction decision-making method influencing population density, and the method comprises the steps: dividing an urban space into square geographic grid units with the same side length; performing unary linear regression analysis among different indexes by combining three parameter indexes, namely the density of interest points of all types, the density of interest points of different function types and the mixing degree of the function types of the interest points, which reflect the current development level of urban public service facilities; and further performing quantile regression analysis on the parameter indexes passing the regression model test, constructing a correlation and an influence process between population distribution data in the geographic grid units and public service facility indexes, and comprehensively judging the influence strength of the public service facilities on regions in different population development stages. According to the method, the public service facility types which should be preferentially built in regions with different population development levels in a city are accurately reflected, so that the population density improvement efficiency is improved, and the use efficiency of public service facility configuration is optimized.

The invention relates to a system for predicting the traffic probability density of a route sector. The system comprises a sample acquisition module suitable for acquiring the traffic flow of the route sector within a preset time as a sample; the sample analysis module is suitable for sample data analysis; the first prediction result prediction module carries out probabilistic prediction on the traffic demand of the route sector according to sample data analysis in combination with model parameter selection, and obtains a first prediction result. A neural network is combined with a quantile regression method, so that a plurality of quantiles of continuous traffic demand data of a certain day in the future are obtained. Then, the continuous conditional quantiles are utilized, and a probability density function and a probability density curve graph of traffic demand continuity in a certain day in the future are obtained through a kernel density estimation method. Therefore, the specificpoint prediction value and the change interval thereof can be obtained, the probability of each value of the traffic demand prediction change interval of the route sector can be obtained, and the accurate point prediction value of the day can be obtained.

The invention discloses a method for evaluating airport publishing capacity by adopting historical operation data envelope lines. The method comprises the following steps: (1) collecting actual historical flight data of an airport to be evaluated; (2) counting airport operation sorts and making a scatter diagram; (3) carrying out airport operation stage division based on clustering analysis; (4) estimating the saturation departure capacity of the airport based on Loss regression analysis; (5) analyzing the airport entering and leaving sortie coupling relation based on quantile regression; (6) evaluating the airport preferential approach capacity based on the confidence interval; and (7) evaluating the airport publishing capacity based on the airport operation data envelope line. According to the historical operation data of the airport, the approach capacity, the departure capacity and the approach and departure coordination capacity of the peak operation period of the airport are quantitatively evaluated through technical means of clustering analysis, regression analysis and mathematical statistics, , and the evaluation result is accurate.

The invention discloses a short-term power load probability prediction method based on CNN and quantile regression, and the method comprises the following steps: 1, collecting power load data, and determining a key influence factor according to the correlation between the power load data and an external influence factor; 2, preprocessing the data, and segmenting the input data into training set data and test set data; 3, performing short-term load probability density prediction model training based on the convolutional neural network and quantile regression by using the training set data in the step 2 to obtain a trained short-term load probability density prediction model based on the convolutional neural network and quantile regression; 4, inputting test data into the trained QRCNN modelto obtain predicted values under different quantiles; and 5, taking predicted values under different quantiles as input, and carrying out load probability density prediction by using a kernel densityestimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.

The invention relates to the technical field of track analysis, in particular to a track quality index threshold reasonability analysis method based on quantile regression, which comprises: 1, a step of determining a reasonable standard value; 2, a step of preprocessing the data of a track inspection vehicle, comprising detection data selection, mileage correction and time sequence-based abnormal peak value processing; 3, a step of analysis of detection data: acquiring a multiple relationship between a half peak value and a standard deviation corresponding to peak value management and mean value management through linear regression of different quantiles. Accordingly, the reasonability of TQI management values and various standard deviations in mean value management of lines where different operation speeds per hour and different track board types are located is analyzed, and the analysis specifically comprises quantile regression statistical analysis, index fitting analysis and mean value suggestion management value analysis. According to the method, the track quality state can be better analyzed.

The invention discloses a day-ahead photovoltaic power non-parametric probability prediction method based on QRA-LSTM. Photovoltaic historical data and numerical weather forecast data (NSW) are adopted to train a group of mutually independent long-term and short-term memory (LSTM) deterministic prediction models, and a quantile regression average algorithm (QRA) is adopted to integrate each independent LSTM prediction model to generate a non-parametric probability prediction model of photovoltaic output. Non-parametric probability prediction can describe the uncertainty problem which is difficult to reflect by simple deterministic prediction, and the result has higher credibility. The method can effectively avoid the problem that deterministic prediction and probability prediction are separately considered, provides an important basis for decision scheduling of scheduling personnel, and is huge in application value and prospect.

The invention discloses a power load probability prediction method based on constrained parallel LSTM quantile regression, and the method comprises the steps: collecting the load power and impact factor data of a plurality of sample days, and forming a data set; setting model hyper-parameters; establishing a constrained parallel LSTM model, and pre-training each quantile LSTM in the constrained parallel LSTM model to obtain a weight and offset parameter set; performing overall training on the constrained parallel LSTM model, performing fine adjustment on the weight and offset parameters in thetraining process, and determining the optimal weight and offset parameters of the constrained parallel LSTM model; inputting the verification set into the trained constraint parallel LSTM model, andselecting an optimal hyper-parameter of the model according to the verification error; and inputting the test sample into the constrained parallel LSTM model with the optimal hyper-parameter, and carrying out inverse normalization on a prediction result output by the constrained parallel LSTM model. According to the method, quantile regression prediction of the power load is carried out by adopting the constrained parallel LSTM model, so that the predicted load probability distribution is more reasonable, and intersection between quantile prediction values is avoided.

A prediction modeling system, method and computer program product for implementing forecasting models that involve numerous measurement locations, e.g., urban occupancy traffic data. The method invokes a data volatility reduction technique based on computing a congestion threshold for each prediction location, and using that threshold in a filtering scheme. Through the use of calibration, and by obtaining an extremal or other specified solution (e.g., maximization) of empirical volume-occupancy curves as a function of the occupancy level, significant accuracy gains are achieved and at virtually no loss of important information to the end user. The calibration use quantile regression to deal with the asymmetry and scatter of the empirical data. The argmax of each empirical function is used in a unidimensional projection to essentially filter all fully congested occupancy level and treat them as a single state.

Disclosed is an aircraft parameter mapping system and method based on supporting vector machine and multiple regression. The system comprises a data input module, a parameter mapping check module, a parameter mapping verification module, a parameter mapping confirmation module and a mapping display module. Through few aircraft-scheme-related response parameters and self-varying parameter data sample pairs, a parameter mapping relation between the response parameters and the self-varying parameter data sample pairs is established; data samples of part of the response parameters in the data sample pairs are allowed to contain errors or noise, a high-precision calculation model is replaced with parameter mapping to engage in optima design of aircraft overall schemes, the requirement for quickly acquiring optimal aircraft overall schemes is met, and optimizing efficiency is improved.

The invention belongs to the technical field of economic index monitoring analysis, and discloses a computer-based real-time economic index monitoring analysis method. According to the method, a computer-based real-time economic index monitoring analysis system is provided. According to the method, fluctuations and tendencies of data are observed through a tendency chart drawn by selecting upper digits, lower digits and medians as parameters through a quantile regression method, so as to obtain economic fluctuation laws of the recent years and summarize corresponding economic conditions, so that the data index monitoring accuracy can be effectively improved, the abnormal fluctuation tendencies of data indexes can be effectively judged, problems which are difficult to discover can be discovered while erroneous judgement is avoided, and medium and long-term tendency abnormal conditions can be monitored in real time; and newest data indexes are adapted along with the change of network environment, so that the operation and maintenance cost at ordinary times are reduced, foreshadowing is carried out for data analysis and report obtaining, errors of economic index results are decreasedand the efficiency and correctness of monitoring analysis work are improved.