29 results about "Boost regression tree" patented technology

The invention relates to the technical field of equipment management, in particular to a dynamic evaluation method and system for the health condition of a refrigerating unit, and the method comprises the steps: constructing a refrigerating capacity and energy efficiency ratio prediction model through a gradient lifting regression tree algorithm, and training a high-precision reference through the full utilization of historical data; a stable and reliable prediction reference and a data analysis basis are provided for subsequent dynamic evaluation; a static model is converted into a dynamic model responding to a real-time working condition, real-time evaluation on the performance of the refrigerating unit is achieved, and meanwhile, the sensitivity and accuracy of state monitoring are effectively improved by combining degradation degree cooperative calculation; the real-time score of the refrigerating unit is compared with the preset judgment threshold value, so that the health state of the refrigerating unit can be quickly and accurately judged; when it is detected that the refrigerating unit is in the abnormal state, the model updating period and the sensor sampling frequency can be dynamically adjusted, and it is ensured that more accurate and timely data are obtained at the critical moment.

The invention discloses a satellite precipitation two-stage error correction method based on an actual measurement site, and the method comprises the steps: extracting drainage basin DEM data, obtaining a mask file of a drainage basin, and extracting the satellite precipitation of the drainage basin; performing inverse distance weighted interpolation on the rainfall data of the actual measurement site to a corresponding resolution to obtain grid data of the actual measurement site; error correction data are obtained through a deep learning model or dynamic quantile mapping; calculating a residual error between the rainfall data after error correction and grid rainfall of the actual measurement site, and learning a nonlinear relationship between the rainfall data and the residual error by using a gradient lifting regression tree model; and adding the precipitation residual error of the nonlinear residual error correction model to the error-corrected satellite precipitation error to obtain a residual error-corrected precipitation product. According to the method, the satellite precipitation product is corrected based on the actual measurement site data, the high-precision precipitation product is obtained, point-to-surface conversion of the precipitation data can be realized, and data support is provided for input of a refined hydrological model.

The invention relates to the technical field of index structure recombination, in particular to a nursing process management database index optimization method, which comprises the following steps of: performing cost calculation on multiple indexes by adopting a gradient lifting regression tree, comparing the cost calculation with a threshold value, and forming a recombination candidate priority list under the support of a numerical result; a potential recombination object has a quantifiable judgment standard, predicted consumed time is calculated in combination with the low-ebb IO capacity and compared with the time slice length, the screening process can correspond to the system resource bearing condition, a B + tree algorithm is adopted for calculating a loading factor and the continuous scanning length, and through merging, splitting and redistribution operation, the screening efficiency is improved. A node organization structure is finely adjusted, it is guaranteed that an index structure and an access mode are kept coordinated, information entropy is utilized to sort and screen a field sequence, the field arrangement with high distinction degree is preferentially used for constructing a new index, atomic switching operation is executed to replace an old index, and the field sequence has adaptability in a dynamic data environment.

This application discloses an intelligent orchestration method for Spark tasks based on gradient boosting regression trees, relating to the field of resource scheduling technology. The method includes: acquiring multiple Spark tasks and generating multiple task orchestration schemes; for any given task orchestration scheme, inputting the number of executors for each Spark task into a preset gradient boosting regression tree model, and outputting the predicted runtime of each Spark task; determining the fitness score of each task orchestration scheme based on the number of executors, startup time, and predicted runtime of each Spark task in each orchestration scheme; adjusting each task orchestration scheme and iteratively executing the prediction and scoring steps until a stopping condition is met; and determining the task orchestration scheme with the lowest fitness score as the target orchestration scheme for each Spark task. This application combines the prediction and evaluation stages to achieve a closed-loop decision-making process, improving the overall utilization of cluster resources.

The invention discloses a lithium ore lithium extraction rotary kiln stepped air duct heat transfer structure optimization method and system, and relates to the technical field of intelligent control. Multi-source sensing data in a rotary kiln stepped air duct are collected in real time, geometric parameters of the stepped air duct are combined, and coupling characteristics are extracted from the multi-source sensing data by using a principal component analysis method; constructing a dynamic heat transfer resistance coefficient matrix on the basis of the coupling characteristics and the multi-source sensing data, inputting the coupling characteristics and the dynamic heat transfer resistance coefficient matrix into a pre-trained gradient lifting regression tree model, outputting a real-time effective heat transfer area proportion of each air duct layer, and calculating the real-time effective heat transfer area proportion of each air duct layer on the basis of the real-time effective heat transfer area proportion. Step air duct parameters are optimized by adopting a self-adaptive particle swarm algorithm, and the optimized step air duct parameters are used for adjusting the step air duct parameters in the rotary kiln in real time through a PID (Proportion Integration Differentiation) controller to form closed-loop control; and the heat transfer efficiency in the calcining process can be greatly improved.

The application discloses a kind of based on DNA methylation's construction method of biological age prediction model, the application is downloaded from GEO data website, derived from Chinese population, contains calendar age data, whole blood sample 450k methylation chip data, by the method of elastic network combined bootstrap, 31 methylation sites of modeling candidate are screened, multiple linear regression, support vector machine, random forest and gradient boosting regression tree are used to carry out preliminary construction and evaluation of model, then further filter methylation sites using full subset regression, obtain methylation age prediction model based on 18 methylation sites.Finally, using any provincial team natural population data, methylation age prediction model is optimized, and finally the biological age prediction model based on 18 methylation sites is obtained.The model is suitable for Chinese population, the number of methylation sites contained is less, is not affected by blood cell components, and the prediction accuracy is good.

The application discloses a kind of TBM rock breaking efficiency intelligent prediction methods based on rock slag size, which comprises the following steps: install industrial camera on TBM slag belt, and real-time acquisition rock slag image information in TBM tunneling process;The rock slag image information obtained is automatically identified, and the shape and geometric characteristics of the rock slag are obtained;Collect TBM's on-site tunneling parameters, rock slag geometric information characteristics and rock breaking efficiency data, and form sample data set;A gradient boosting regression tree model optimized by particle swarm algorithm is established using the sample data set, and the intelligent prediction of TBM rock breaking efficiency is realized.The application uses particle swarm algorithm to optimize the gradient boosting regression tree model, so as to find the best parameters of the gradient boosting regression tree model, improve the prediction accuracy of the model, and the accurate prediction of rock breaking efficiency can effectively provide reference for TBM tunneling parameter setting and efficient rock breaking.

This invention relates to the field of ecological environment assessment and spatial simulation technology, and discloses a method for grassland degradation ecological risk assessment based on a multi-agent model. The method includes a spatial risk assessment module that acquires multi-source basic data and divides it into discrete grid cells, calculating a landscape ecological risk index to generate spatial distribution results; a driving mechanism identification module that extracts driving factors to construct an improved regression tree model and outputs nonlinear response rules; and a multi-agent dynamic simulation module that constructs grassland environmental entities based on grid cells, injects risk indices and response rules, generates herder behavior entities, and inputs management policy parameters. Iterative calculations are performed using these two types of entities to simulate the dynamic process of herder relocation and resource consumption. After the iteration reaches the target year, the grid landscape attributes are reclassified according to the remaining grass cover, and the data is re-input into the assessment module for aggregation calculations, outputting a spatial distribution map to complete the evolution assessment. This invention can intuitively test the intervention effect of management policies on grassland degradation.

The invention relates to the technical field of soil quality evaluation and improvement, in particular to a shajiang black soil structural obstacle factor and threshold value calculation method, which comprises the following steps: firstly, screening 22 structural obstacle factors covering soil physicochemical and biological attributes, and constructing a combined weight by adopting an entropy weight method, a principal component analysis method and an average weight method; the method comprises the following steps: calculating the membership of each factor by combining ring-type, peak-type and S-type membership functions, obtaining a soil structural obstacle comprehensive index through an accumulation method, dividing obstacle grades, determining a threshold value of each factor by taking a gradient boosting regression tree as a core and combining T test and random forest verification, and finally constructing a minimum data set based on principal component analysis and clustering analysis; the problems that an existing method is one-sided in obstacle factor screening, high in weight determination subjectivity, single in threshold calculation method and difficult to popularize are solved.

A method and system for detecting the three-dimensional morphology of aggregates based on two-dimensional particle parameters is disclosed, relating to the fields of aggregate detection and image processing technology. The method mainly involves acquiring two-dimensional morphological data of aggregates using an aggregate image measurement system (AIMS), acquiring three-dimensional morphological data of aggregates using a three-dimensional blue light scanner and visualization software (AVIZO), and then establishing a correlation mapping between two-dimensional and three-dimensional indicators using a gradient boosting regression tree method. This enables rapid and accurate prediction of the three-dimensional morphology of aggregates, facilitating refined management of mineral aggregate classification, identification, storage, and accurate retrieval, providing scientific support for engineering design, and promoting the intelligent and automated development of engineering management.

The invention relates to an immunoassay automation equipment parameter optimization method based on a gradient boosting regression tree model, and solves the technical problems of how to optimize an inter-plate CV value, an in-plate CV value and a% RE absolute value in a detection process of existing electrochemical luminescence immunoassay equipment and how to improve parameter suitability. According to the method, sample incubation time T, cleaning times N, magnetic bead adsorption time t1 and magnetic bead beating time t2 are predicted through a constructed gradient lifting regression tree model, and equipment executes detection work according to prediction results so as to obtain optimized CV values, S / N and% RE of each hole.

The invention relates to the field of machine learning and the field of soil remediation, and discloses a soil remediation parameter prediction method, device and equipment based on multi-modal data, and the method comprises the steps: determining a trained gradient boosting regression tree model, a trained extreme gradient boosting model and a trained random forest model; when the first test result, the second test result and the third test result meet predefined conditions, deploying a trained gradient boosting regression tree model, a trained extreme gradient boosting model and a trained random forest model; obtaining current features through the multi-modal data of the current biochar and the pollution environment parameters of the current soil, and obtaining predicted soil remediation parameters based on the trained gradient boosting regression tree model, the trained extreme gradient boosting model, the trained random forest model and the current features. The method is beneficial for improving the obtaining efficiency of the predicted soil remediation parameters.

The present application relates to the technical field of equipment management, and particularly relates to a refrigeration unit health condition dynamic evaluation method and system, comprising the following steps: a gradient boosting regression tree algorithm is used to construct a refrigeration capacity and energy efficiency ratio prediction model, historical data is fully utilized to train a high-precision benchmark, and a stable and reliable prediction benchmark and data analysis basis are provided for subsequent dynamic evaluation; a static model is converted into a dynamic model responding to real-time working conditions, real-time evaluation of the performance of the refrigeration unit is realized, and the sensitivity and accuracy of state monitoring are effectively improved in combination with degradation degree cooperative calculation; the real-time score of the refrigeration unit is compared with a preset judgment threshold, and the health state of the refrigeration unit can be quickly and accurately judged; when it is detected that the refrigeration unit is in an abnormal state, the model update period and the sensor sampling frequency can be dynamically adjusted, and more accurate and timely data can be obtained at a critical moment.

The application discloses a cross-border medical data sharing privacy protection method based on a blockchain and federated learning, and relates to the technical field of data security and medical informatization. The method comprises the following steps: configuring a multi-domain identifier and compliance priority for a medical node through multi-dimensional attribution calibration; and adopting dynamic privacy budget adjustment or channel separation processing for an overlapping domain scene. Privacy protection groups are divided according to countries / regions, trust domains and sensitive levels based on a rule grouping method, and a continuous flow sequence is marked. A privacy protection strategy prediction model is constructed, federated gradient boosting regression tree training is adopted, and model updating is realized by combining local encrypted gradient aggregation. The comprehensive privacy protection contribution degree is calculated through strategy-compliance vector matching and trust decay factor calculation, and the downstream grouping strategy configuration is dynamically controlled. The application optimizes the privacy protection strength and data sharing efficiency, avoids excessive protection, realizes the safe and efficient circulation of cross-border medical data, and is suitable for the medical data sharing scene of cooperation of multiple countries and institutions.

The present application relates to the technical field of HDI printed circuit board processing, in particular to a processing method for solving the problems of copper surface residual glue and small size of binding PAD in cavity product position, after laser cavity processing, the characteristic vector is input into the trained gradient boosting regression tree (GBRT) model, and the adaptive plasma parameter is output, and the cavity copper surface is treated by adopting the adaptive plasma parameter; the present application realizes the dynamic matching of residual glue characteristics and treatment parameters by adopting the GBRT model to adaptively output the plasma parameter and combining with the sand blasting path planned by the A* algorithm; wherein the prediction error of the GBRT model to the plasma parameter is less than or equal to 5%, the priority coverage of the region to be optimized is ensured, the path overlap rate is less than or equal to 5%, the final residual glue removal rate is improved to more than 98%, and the problem of incomplete treatment of traditional fixed parameters is solved.

The application provides an optimization design method of additive manufacturing process parameters based on machine learning, which can realize rapid and accurate prediction of product forming quality under any process parameters for a wide range of material systems by establishing an additive manufacturing process parameter-material performance gradient boosting regression tree (GBDT) model and double optimizing the model hyperparameters by using random search (RS) and K-fold cross validation (K-CV) algorithms, so as to quickly and accurately determine the best process parameters, and solve the problems of high calculation and test cost and long cycle in the optimization of the process parameter window of laser additive manufacturing.

The present application relates to the technical field of material design, and particularly relates to a prediction model construction method for T 50 and T 90 in a platinum-based bimetallic catalytic material propane oxidation reaction, the present application performs weighted operation on intrinsic physicochemical parameters of an active metal content ratio as a weight to construct weighted features, removes redundant features to form a feature pool through a Pearson correlation coefficient, performs systematic evaluation on seven kinds of mechanism different regression models under a leave-one-out cross-validation framework, and selects a gradient boosting regression tree as an optimal model, performs exhaustive search screening on remaining variable features after fixing calcination temperature, calcination time and reaction atmosphere as key process features, determines an optimal feature combination composed of fixed features, active metal melting point, active metal conductivity, carrier formation energy and carrier band gap, and establishes a prediction model according to the optimal feature combination. The model of the present application is subjected to gridding virtual screening and closed-loop verification of experiments, the prediction value is highly consistent with the experimental value, and the model can replace a large number of preliminary screening experiments.

The invention discloses a core adding and code adding method and system based on warehouse management equipment. The method comprises the following steps: S1, collecting equipment basic information; s2, calling a built-in rule engine to load and preset, generating a complete equipment number and outputting a two-dimensional code image; s3, decoding the equipment number information in the two-dimensional code image, and executing structured verification; s4, constructing a multi-level interaction architecture, wherein the multi-level interaction architecture comprises a business logic layer, an interaction middle layer and a data storage and interface layer; s5, when the equipment numbering module completes number generation, two-dimensional code generation and RFID writing are triggered; s6, reading equipment RFID tag data in batches through an integrated UHF RFID module, and predicting multi-point RSSI data acquired in real time based on a regional gradient lifting regression tree model; and S7, when the checking module completes the checking operation, carrying out data synchronization updating and recording an operation state. According to the invention, high-efficiency, accurate and intelligent management of warehouse equipment is realized.

This invention discloses a target recognition method based on adaptive multinomial kernel function and multimodal fusion, belonging to the field of target recognition technology. The method includes: acquiring the original signal of the target to be identified; preprocessing the original signal using a wavelet threshold denoising algorithm to generate a signal to be analyzed; performing multimodal time-frequency analysis on the signal to be analyzed, outputting three types of time-frequency feature maps; dynamically weighting and fusing the three types of time-frequency feature maps to generate a fused time-frequency feature tensor, and then processing it through a convolutional neural network to generate a time-series feature vector; inputting the time-series feature vector into a convolutional neural network and a gradient boosting regression tree (GBRT) model respectively for processing, and then outputting preliminary recognition results and feature verification results based on the target category to complete target recognition. This invention achieves accurate target recognition by combining the advantages of multimodal analysis technology with short-time Fourier transform and wavelet transform.

The invention relates to the technical field of HDI printed circuit board processing, in particular to a processing method for solving the problems of residual glue on a copper surface at the position of a Cavity product and small size of a bound PAD, after laser Cavity processing, feature vectors are input into a trained gradient boosting regression tree (GBRT) model, adaptive Plasma parameters are output, and the adaptive Plasma parameters are adopted to carry out Plasma processing on the Cavity copper surface; according to the method, the Plasma parameters are output in a self-adaptive mode through the GBRT model, and dynamic matching of the residual glue characteristics and the processing parameters is achieved in combination with the sand blasting path planned through the A * algorithm; wherein the prediction error of the GBRT model on the Plasma parameter is less than or equal to 5%, the preferential coverage of a to-be-optimized area is ensured, the path overlapping rate is less than or equal to 5%, the final residual glue removal rate is improved to 98% or above, and the problem that the traditional fixed parameter processing is not thorough is solved.

The invention provides a grillage structure underwater explosion response prediction method and system based on parametric modeling, and relates to the technical field of ship and ocean engineering structure safety assessment, and the method comprises the steps: constructing a test parameter set; inputting the test parameters into finite element simulation software in batches to obtain a response data set, wherein the response data set comprises multiple groups of response data; constructing each group of response data and the corresponding test parameters into a group of training data to obtain a training data set; training the random forest model through test parameters and rupture labels in the training data set to obtain a rupture prediction model; and training the gradient lifting regression tree model through test parameters, the maximum equivalent stress value, the maximum equivalent plastic strain value, the maximum deformation displacement value and the crevasse area in the training data set to obtain a physical response prediction model. The response prediction speed can be greatly increased through parametric modeling and model prediction, and the trained prediction model does not depend on a grid structure and has very high generalization ability.

The invention relates to the technical field of atmospheric environment monitoring, in particular to a water surface atmospheric environment monitoring and tracing system based on multi-source data fusion. According to the system, a monitoring layer adopts a dynamic sampling frequency adaptive algorithm to realize multi-source data synchronous acquisition, and the data precision of a high-pollution area is improved; the data transmission and fusion layer solves the problems of time-space dislocation and standardization of multi-source data through a time attention mechanism and an inverse distance weighting algorithm; the analysis layer improves a forward matrix decomposition model, realizes automatic pollution source identification by introducing a Bayesian optimization factor number and combining with random forest classification, and quantifies secondary pollution contribution by matching with a gradient lifting regression tree model; and the application layer realizes pollution source quantitative traceability and high-value area positioning through a geographically weighted regression model and hotspot analysis, and outputs visual products and prevention and control suggestions. The system effectively solves the problems of low water surface atmosphere monitoring data quality and insufficient traceability precision, and provides technical support for cross-regional joint defense and joint control of the Yangtze river basin.

The application discloses a kind of tunneling roadway dust concentration prediction method and system, including obtaining the actual dust data of tunneling roadway tunneling process, physical experiment data and simulation experiment data and based on single class support vector machine scheme to preprocess to construct training data set;Based on deep neural network and gradient boosting regression tree scheme constructs tunneling roadway dust concentration prediction initial model and trains to obtain tunneling roadway dust concentration prediction model;Actual tunneling roadway dust concentration prediction is carried out using the obtained tunneling roadway dust concentration prediction model.The application is based on deep neural network and gradient boosting regression tree scheme constructs tunneling roadway dust concentration prediction model and trains, through the interaction and joint training between two networks, not only realizes the prediction of tunneling roadway dust concentration, but also higher reliability, better accuracy.

This application relates to the fields of machine learning and soil remediation. It discloses a method, apparatus, and device for predicting soil remediation parameters based on multimodal data. The method includes: determining a trained gradient boosting regression tree model, a trained extreme gradient boosting model, and a trained random forest model; deploying the trained gradient boosting regression tree model, the trained extreme gradient boosting model, and the trained random forest model when the first test result, the second test result, and the third test result meet predefined conditions; obtaining current features using current biochar multimodal data and current soil pollution environmental parameters; and obtaining predicted soil remediation parameters based on the trained gradient boosting regression tree model, the trained extreme gradient boosting model, the trained random forest model, and the current features. This application is beneficial for improving the efficiency of obtaining predicted soil remediation parameters.

The invention discloses an intelligent Spark task arrangement method based on a gradient boosting regression tree, and relates to the technical field of resource scheduling, and the method comprises the steps: obtaining a plurality of Spark tasks, and generating a plurality of task arrangement schemes; for any task arrangement scheme, inputting the number of actuators of each Spark task into a preset gradient lifting regression tree model, and outputting the predicted running time of each Spark task; according to the number of actuators, the starting time and the predicted operation duration of each Spark task in each task arrangement scheme, determining a fitness score of each task arrangement scheme; adjusting each task arrangement scheme, and iteratively executing the prediction and scoring steps until a stop condition is met; and determining the task arrangement scheme with the minimum fitness score as a target arrangement scheme of each Spark task. According to the method, the prediction link and the evaluation link are combined, so that a decision closed loop is realized, and the overall utilization rate of cluster resources is improved.

The invention discloses a wheat-corn rotation future yield prediction method and system based on a process model and machine learning, and belongs to the crossing field of agricultural information technology, crop model and machine learning, and the method comprises the steps: constructing a time-space continuous county-level scale region data set; performing parameter calibration on the process model by using point location data, and simulating wheat and corn yields of a site scale; generating a sample set through a fertilization scene experiment and an extreme climate combination scene experiment, and dividing the sample set into a training set and a verification set; a gradient boosting regression tree model integrated with an attention mechanism is trained and verified, and a DNDC-GBRT-ATT prediction framework is constructed; and inputting the county-level scale region data set into the trained gradient boosting regression tree model, and outputting a county-level scale yield prediction result of the wheat-corn rotation mode under future climate change. According to the method, precise county-level prediction is realized, and reliable technical support is provided for regional agricultural production decision making.

The invention discloses a cross-border medical data sharing privacy protection method based on a block chain and federated learning, and relates to the technical field of data security and medical informatization crossing. The method comprises the steps of configuring multi-domain identifiers and compliance priorities for medical nodes through multi-dimensional affiliation calibration, and adopting dynamic privacy budget adjustment or channel separation processing for overlapped domain scenes. Privacy protection groups are divided according to countries / regions, trust domains and sensitivity levels based on a rule grouping method, and a continuous circulation sequence is marked. And constructing a privacy protection strategy prediction model, adopting federal gradient lifting regression tree training, and combining with local encryption gradient aggregation to realize model updating. And calculating a comprehensive privacy protection contribution degree through strategy-compliance vector matching and a trust attenuation factor, and dynamically regulating and controlling downstream grouping strategy configuration. According to the method, the privacy protection intensity and the data sharing efficiency are optimized, excessive protection is avoided, safe and efficient circulation of cross-border medical data is realized, and the method is suitable for a multi-country and multi-institution collaborative medical data sharing scene.

The invention discloses a target recognition method based on an adaptive polynomial kernel function and multi-modal fusion, and belongs to the technical field of target recognition, and the target recognition method comprises the steps: obtaining an original signal of a to-be-recognized target, and carrying out the preprocessing of the original signal based on a wavelet threshold denoising algorithm, and generating a to-be-analyzed signal; performing multi-modal time-frequency analysis on the to-be-analyzed signal, and outputting three types of time-frequency characteristic patterns; carrying out dynamic weighted fusion processing on the three types of time-frequency feature maps to generate a fused time-frequency feature tensor, and carrying out convolutional neural network processing to generate a time sequence feature vector; and respectively inputting the time sequence feature vector into a convolutional neural network and a gradient boosting regression tree GBRT model for processing, and then outputting a preliminary recognition result and a feature verification result based on a target category to complete target recognition. According to the method, through a multi-modal analysis technology, the advantages of short-time Fourier transform and wavelet transform are combined, and accurate recognition of the target is achieved.

The invention relates to the technical field of ecological environment assessment and space simulation, and discloses a grassland degradation ecological risk assessment method based on a multi-agent model, and the method comprises the steps: obtaining multi-source basic data through a space risk assessment module, dividing the data into discrete grid units, calculating a landscape ecological risk index, and generating a space distribution result; the driving mechanism identification module extracts driving factors to construct a lifting regression tree model, and outputs a nonlinear response rule; the multi-agent dynamic simulation module constructs a grassland environment entity based on a grid unit, injects a risk index and a response rule, generates a pasture behavior entity, inputs a management policy parameter, executes iterative operation by using the two entities, and simulates a pasture addressing transfer and resource consumption dynamic process; after iteration reaches the target year, grid landscape attributes are re-classified according to the residual grass quantity, the grid landscape attributes are re-input into the evaluation module for aggregation operation, and a spatial distribution map is output to complete evolution evaluation. The method can visually test the intervention effect of the management and control policy on the grassland degradation.