20 results about "Local regression" patented technology

The invention relates to the technical field of action recognition, in particular to a pattern rope skipping action recognition system and method, and the system comprises a limb angle collection module, a stability recognition module, an action triggering and screening module, a rope body direction analysis module and a cross mapping judgment module. According to the method, by introducing three-dimensional coordinate extraction and included angle change trend analysis of limb key joints in the jumping period, the dynamic characteristics of limbs in the action process can be accurately reflected, and the action amplitude and stability change can be effectively expressed through an angle data sequence constructed by multi-dimensional indexes such as the maximum value, the minimum value and the average value of the included angle in the period; a stability identification mechanism is constructed to carry out local regression fitting and residual error statistics on a periodic angle trend, accurate labeling of a stable state of an action posture is realized, and then a key frame sequence with an actual jump intention and a continuous posture in the jump action is effectively identified through fusion screening of labeled frames and a jump-empty state.

The invention relates to an ecological system health space-time driven attribution analysis method based on IGTWR-XGBoost, and the method comprises the steps: obtaining an ecological system health index of a target region, calculating the ecological system health index of the target region according to the ecological system health index, and collecting a multi-source driving factor; constructing an interactive geographic space-time weighted regression model according to the multi-source driving factor, the ecological system health index and a geographic-time weighted regression model; acquiring a multi-dimensional local regression coefficient matrix according to the interactive geographic space-time weighted regression model; correcting a multi-source driving factor according to the multi-dimensional local regression coefficient matrix to form a space-time enhanced feature matrix; and inputting the space-time enhanced feature matrix into an XGBoost model, and obtaining a prediction result of the ecological system health index and global and local contribution values of each driving factor. According to the invention, the accuracy and reliability of ecosystem health driving mechanism analysis can be significantly improved.

The invention provides an ocean internal wave parameter inversion method and system based on temperature chain buoy data. The method comprises the following steps: preprocessing an acquired multi-depth temperature sequence and then constructing a temperature-time profile; detrending and self-adaptive smoothing are carried out through empirical mode decomposition, and a temperature background field and a pulsating field are separated; extracting a vertical gradient of a temperature background profile by using LOESS local regression, and introducing Tikhonov regularization and threshold gating to guarantee gradient robustness; mapping temperature pulsation into vertical displacement based on isothermal surface conservation, and screening effective data in combination with a credibility index; finally, the amplitude, the period and the wave height of an internal wave are inversed through multi-window wavelet spectrum and Hilbert envelope conjoint analysis, and the reliability of a result is improved through vertical modal consistency verification. According to the method, consistent inversion of ocean internal wave parameters can be completed only by means of a single temperature chain buoy, and the inversion accuracy under the background of weak gradient, weak signal and noise is remarkably improved.

The invention provides a traffic time-space diagram reconstruction method and system based on neighborhood adaptive regression, and relates to the technical field of computers, and the method comprises the steps: obtaining historical traffic data; performing space-time diagram construction processing according to the historical traffic data to obtain a historical data space-time diagram; performing sample construction according to the historical data space-time diagram, and dividing the down-sampling low-resolution space-time diagram into image blocks with predefined sizes in an overlapping and cutting mode to obtain a training sample pair set; performing neighborhood search on the set based on the training sample to construct a neighborhood set; performing neighborhood regression processing according to the neighborhood set, and fitting to obtain local regression parameters; and performing high-resolution reconstruction processing according to the local regression parameters, predicting high-resolution sub-block values by applying the regression parameters to linear transformation of the to-be-reconstructed image blocks, and integrating all the sub-blocks to obtain a reconstructed traffic space-time diagram. According to the method, efficient reconstruction precision improvement can be realized under a small sample training condition.

The invention discloses a city street four-season green vision rate estimation method based on an improved local regression model, and the method comprises the steps: generating street scene sampling points according to road network data, collecting street scene images, carrying out the semantic segmentation of the street scene images through a deep learning model, and extracting a vegetation region in the street scene images. The method comprises the following steps of: calculating the green vision rate of a street scene sampling point in each season, extracting a multi-season normalized vegetation index of the sampling point through remote sensing data, establishing a regression model for the normalized vegetation index and the green vision rate of a specific season by utilizing a local regression model fused with a season weight, and estimating the missing green vision rate by utilizing the normalized vegetation index. According to the method, annual street view data are accurately collected and processed, a remote sensing technology, a deep learning technology and an image segmentation technology are combined, and an effective means is provided for scientific evaluation of urban street greening in different seasons. Meanwhile, the method provided by the invention is high in automation degree, and is suitable for practical application of large-scale urban landscaping evaluation.

The application belongs to the technical field of power consumption prediction, and discloses a fuzzy power consumption prediction method and system based on multi-scale double supervision driving, which comprises the following steps: multi-scale feature reconstruction is performed on data samples, reconstruction error is calculated, sample structure guide indexes are obtained by aggregating the reconstruction error, and the samples are screened in combination with structure guide indexes and entropy contribution; the screened samples are subjected to fuzzy division, the membership degrees of all samples are arranged in order to obtain a fuzzy membership degree matrix; structure supervision signals and target supervision signals are fused, and the membership degree matrix is adjusted by using the fused guide signals; the membership degrees in the new membership degree matrix are used as weights to perform weighted summation on the prediction results of each local regression model to obtain the final prediction result. The application designs a fuzzy division method combining a multi-scale structure adaptive mechanism with structure and target double supervision signals, and improves the modeling accuracy of the model for complex energy consumption data, the system adaptive ability and the sample utilization efficiency.

The invention relates to a semiconductor measurement calculation optimization method and system. The method comprises the steps that calculation partitions and users are provided, the users comprise a first user and a second user, and the calculation partitions comprise a first partition and a second partition; performing parallel global search by using the first partition to generate an initial point set; executing local regression fitting by using the second partition and taking the initial point set as a starting point to obtain a local regression result; and judging and iterating a local regression result. By integrating the cluster management system, different types of computing tasks are automatically scheduled to more efficient heterogeneous hardware to be executed, and the problems that in the prior art, the computing resource utilization rate is low, and the optimization process is prone to falling into local optimum are solved.

The invention discloses a subarea treatment method and device for soil erosion, a storage medium and electronic equipment, and the method comprises the steps: carrying out the preprocessing of collected multi-source data needed by landscape pattern research, and obtaining the preprocessed data; calculating an equation factor of the general soil loss equation by using the preprocessed data, and deducing a soil erosion modulus based on the equation factor inversion; determining a landscape index space layer of the landscape pattern, inputting the landscape index space layer and the soil erosion modulus into a space-time geographically weighted regression model, and outputting a local regression coefficient on each space-time unit; and dividing treatment stages based on the local regression coefficient and generating a partitioning scheme. By means of the method and device, the problem that in the related technology, due to the fact that correlation analysis between the landscape pattern and soil erosion is insufficient, the zoning treatment effect of the soil erosion is poor is solved.

The invention provides a power plant carbon dioxide emission quantification method and system based on satellite observation data. The method comprises the following specific steps: extracting and screening observation XCO2 data with an inversion condition in a set space range around a target power plant, obtaining background XCO2 optimal estimation by adopting robust local regression, further obtaining a background baseline, obtaining robust observation enhanced XCO2 data, and obtaining the optimal estimation of the background XCO2. And inputting the estimated value into an emission inversion model to realize emission rate inversion at the transit moment of the power plant, and repeatedly iteratively operating for at least 1000 times to output an estimated value set of emission intensity to judge that the transit inversion result is effective, so that a matched quality control mechanism is formed. According to the method, manual intervention and experience setting dependence are reduced through automatic screening and robust modeling, inversion stability, calculation efficiency and engineering applicability are improved, and the method has obvious advantages in a large-range and long-time-sequence power plant COemission monitoring and list verification scene.

The invention discloses a multi-school knowledge crack robust inference method based on generative quantile regression. The method comprises the following steps: carrying out tensor modeling and feature construction on multi-school academic data; constructing a local quantile regression function and a local loss function, and training and optimizing the local quantile regression function to obtain a local regression parameter vector of the ith campus; generating an agent sample of the central node; constructing a global quantile regression function, and estimating a global knowledge mastering quantile curved surface; constructing a knowledge crack tensor by adopting a local quantile regression function and a global quantile regression function, and performing comprehensive crack index calculation; performing structured smoothing processing on the comprehensive crack index based on a pre-repair relationship between the knowledge points, performing crack significance test, and forming a crack feature vector with a structural feature in any campus; clustering the crack feature vector of any campus; and a crack matrix is constructed, and a result set of multi-campus knowledge crack key inference is obtained.

The invention relates to the technical field of hydrology and water resource prediction, and aims to provide an incoming water prediction method based on segmented regression and local kernel weighted integration in order to solve the problems that the global structure and local characteristics of hydrology data are difficult to consider and the weight distribution in real-time prediction is inaccurate in an existing prediction method. The method is expected to effectively improve the accuracy and robustness of incoming water prediction by performing adaptive partitioning on historical hydrological states and constructing a local optimal model and combining real-time feature similarity dynamic weighted integration. According to the technical scheme, the method comprises the steps of obtaining historical hydrological state data of a watershed, and constructing a standardized multi-dimensional feature vector; dividing the multi-dimensional feature vector into K hydrological state intervals by adopting an unsupervised clustering algorithm; independently training a corresponding local regression prediction model by using data of each interval; obtaining a new to-be-predicted sample, calculating the feature similarity, and determining the weight belonging to each interval; and calling each local regression model to calculate a predicted value, and carrying out weighted integration on the predicted value according to the weight to obtain an incoming water prediction result.

The application relates to the technical field of equipment calibration, in particular to a calibration method and system of a multi-channel pressure sensor, which comprises the following steps: constructing a flexible pressure touch response platform, performing channel calibration according to the flexible pressure touch response platform, and performing preliminary deviation modeling; generating a nonlinear response correction factor according to the output result of the preliminary deviation modeling, and performing local regression adjustment on the channels according to the nonlinear response correction factor; in response to the completion of the local regression adjustment of each channel, performing unified response surface fusion processing; in response to the completion of the unified response surface fusion processing, constructing a stability comprehensive criterion function, and generating a calibration result according to the stability comprehensive criterion function. The application overcomes the limitations of traditional methods which depend on linear models and static single-point calibration, and solves the error accumulation problem caused by the interaction interference and delayed response among channels.

The application relates to the technical field of road maintenance, and provides a double-line self-updating method and system for asphalt pavement performance estimation, which comprises the following steps: constructing a general prediction model based on historical data and a local regression model based on measured data; through a short-term adaptive prediction path, model fusion strategy is adopted to fuse the prediction results of the two models, the weight is dynamically adjusted according to the data accumulation condition, and a final performance prediction value is output; through a long-term gradual updating path executed in parallel, new knowledge samples are detected based on a clustering model, when the samples reach a threshold value, incremental learning of the general prediction model is triggered, and self-updating of the model is realized. The application overcomes the limitations of existing static models, can dynamically adapt to different road section characteristics and data changes, and improves the accuracy, self-adaptability and long-term reliability of asphalt pavement performance estimation.

The application relates to the field of heavy metal pollution cause analysis, and discloses a method and device for analyzing source-sink relationship of historical mine and surrounding farmland soil. The method comprises the following steps: collecting heavy metal pollution data, wherein the heavy metal pollution data comprises mine production related data and historical pollution condition information data; collecting samples for analysis to obtain sample analysis data, wherein the samples comprise mine solid waste samples, acid wastewater samples, irrigation water samples, sediment samples and farmland soil samples; collecting and summarizing the heavy metal pollution data and the sample analysis data to form an achievement database; extracting sample input data from the achievement database; inputting the sample input data into a geographic weighted regression model to obtain a correlation coefficient of heavy metal content of the surrounding farmland soil and heavy metal content of the mine solid waste sample. The application can realize quantitative and accurate source analysis of heavy metals in the soil around the mine by calculating the correlation coefficient through the geographic weighted regression model based on the local regression analysis model and the weighted geographic position.

This invention provides a geographically weighted infectious disease risk analysis method and system, comprising: rasterizing multi-source data and constructing a unified grid dataset and analysis indicator library; calculating the Pearson correlation coefficient between grid cases and various indicators in the unified grid dataset, and outputting the final set of influencing indicators; determining the neighborhood range using an adaptive bandwidth method and determining the spatial weights corresponding to neighborhood samples based on a specified kernel function; constructing exposure sequences corresponding to the current analysis period and multiple historical lag periods for each key indicator; forming a cross-basis local sample set based on the cross-basis features constructed based on the target grid cell and neighboring grid cells, constructing a geographically weighted distributed lag nonlinear model based on the cross-basis local sample set, and performing local regression fitting; comparing the effect differences of different exposure levels over all lag periods to obtain the cumulative effect value, and determining the lag period that contributes the most to the infectious disease risk of the target grid cell as the optimal lag period.

This invention provides a method for reconstructing complex surfaces based on white light interferometry data: The method acquires surface topography measurement data of the complex surface to be reconstructed using white light interferometry; constructs a sliding local region centered on each point to be estimated within the measurement area; processes the measurement data within each local region to obtain the height value of the corresponding point to be estimated; and outputs the global reconstruction result of the complex surface after traversing all points to be estimated across the entire region. The process of processing the measurement data within each local region includes the following steps: S1, using MM estimation to perform robust pre-estimation of the measurement data within the current local region; S2, using the absolute value of the residual as input, and through mixing... t Distribution clustering adaptively classifies the measurement data to obtain the retained valid measurement data; S3, repeat S1 to S2 until the outlier removal iteration is completed to obtain the final valid measurement dataset; S4, use weighted Jacobi basis functions to perform local regression to obtain the height value of the point to be estimated.

The invention discloses a cancer prediction model construction method based on a causal network and adaptive feature selection, and the construction process comprises the steps: 1) obtaining multi-omics data, and generating an omics feature matrix through similarity network iterative fusion, chi-square test dimension reduction and diffusion enhancement; 2) introducing a causal diffusion Do-calculus algorithm to construct a gene causal consensus network, constructing a sample specific network in combination with local regression residual analysis, and extracting topological features to obtain a network feature matrix; 3) calculating a modal weight based on balance accuracy, and performing weighted splicing to generate an integrated multi-modal feature; and 4) inputting the integrated multi-modal features into a full connection layer embedded with a feature gating mechanism and sparse group Lasso regularization, obtaining a cancer prediction model through end-to-end training, and screening core markers. The problems that an existing prediction model is poor in multi-omics data integration effect, lacks a causal mechanism and is difficult to explain deep nonlinear interaction of genes are solved, and the precision and generalization ability of the cancer prediction model are improved.

The invention relates to the technical field of water resource management, and discloses a multi-objective collaborative dynamic optimal configuration method and system for water resources of Mauwu sand, and the method comprises the following steps: building a spatial grid database; calculating a local regression coefficient of each network, constructing a topological structure of rainfall, underground water, vegetation and lakes, and outputting an association probability between elements; constructing a multi-objective optimization model and a coupling dynamic optimization model; and solving the coupling dynamic optimization model to generate a Pareto optimal solution set, calculating an ecological option value of each allocation scheme, and performing optimization adjustment on the water resource allocation schemes under different scenes to generate an optimal water resource allocation scheme. Through deep fusion of multi-target dynamic collaboration, spatial fine modeling, hysteresis effect quantification and long-term value evaluation, a water resource management closed loop adaptive to the Maowu sand land is formed, and the stability and toughness of an ecological system are greatly enhanced while economic and social development is guaranteed.

The invention relates to the technical field of road maintenance, and provides a double-line self-updating method and system for asphalt pavement performance estimation, and the method comprises the steps: constructing a general prediction model based on historical data and a local regression model based on actual measurement data; fusing prediction results of the two models by adopting a model fusion strategy through a short-term self-adaptive prediction path, dynamically adjusting the weight according to a data accumulation condition, and outputting a final performance prediction value; a new knowledge sample is detected based on a clustering model through a parallel execution long-term progressive update path, and when the sample reaches a threshold value, incremental learning of a general prediction model is triggered to realize autonomous update of the model. The method overcomes the limitation of an existing static model, can dynamically adapt to different road section characteristics and data changes, and improves the precision, self-adaptability and long-term reliability of asphalt pavement performance estimation.

The invention discloses a near-surface nitrogen dioxide concentration inversion method based on two-stage ensemble learning. The method comprises the following steps: acquiring ground station nitrogen dioxide, satellite nitrogen dioxide troposphere vertical column concentration and other auxiliary variable data; performing Kriging interpolation processing on the satellite data and other geographical auxiliary variable data; in the first stage, a base model is constructed by adopting ET, RF, GBRT and AdaBoost algorithms, and split node feature weights are dynamically adjusted through a feature-level attention mechanism; in the second stage, prediction results of all the models in the first stage serve as secondary features to be input into the GWR model, a spatial heterogeneity weighting matrix is constructed in combination with attention weight, and dynamic optimization of local regression parameters is achieved; and inverting the near-surface nitrogen dioxide concentration by using a two-stage integrated learning method. According to the method, the accuracy and reliability of near-surface nitrogen dioxide concentration inversion are improved through the two-stage ensemble learning model fusing the attention mechanism.