46 results about "Predictive regression" patented technology

The invention discloses an agent action prediction method based on multi-scale spatial perception, which comprises the steps of map spatial structure modeling, historical trajectory feature extraction, feature fusion and multi-modal action prediction. According to the map space structure, map feature extraction is conducted on two-dimensional vector map data in an application scene map through a multi-scale map convolutional neural network, and high-dimensional map feature information is obtained; the historical trajectory feature extraction adopts a convolutional neural network and a feature pyramid network to extract high-dimensional trajectory data features of all agents; according to feature fusion, correlation modeling fusion is carried out on high-dimensional map feature information and high-dimensional trajectory data feature information through a self-attention mechanism, and agent trajectory fusion features with direction information are obtained; according to the multi-modal trajectory prediction, intelligent agent trajectory fusion features are used for prediction regression and confidence scoring, multi-modal complete trajectory coordinates of action prediction and corresponding confidence scores are output, and a reasonable auxiliary decision is provided for intelligent agent actions.

The invention relates to the field of natural language processing technology and law decision prediction (LJP), and provides a method based on a BERT model, keyword extraction and a cross attention mechanism. Specifically, the method comprises the steps that firstly, case fact description is input into a keyword extraction model based on Transform, legal fact keywords are extracted, and the legal fact keywords are input into a BERT model to obtain vector representation; thirdly, fusing the forensic semantics into the model by utilizing a cross attention mechanism, so as to predict criminal name prediction, law article prediction (classification task) and criminal period prediction (regression task); according to the method, the keywords related to the context are extracted and embedded into the BERT model, and meanwhile, a cross attention mechanism is adopted to be brought into the forensic semantics, so that the problem that key event information is difficult to accurately capture in a traditional method is effectively solved, and the prediction accuracy of each subtask of the LJP is improved.

The invention relates to an adaptive energy consumption control method, system and device based on a computing power chip and a medium. The method comprises the following steps: constructing an optimal prediction regression model based on multi-channel hardware state indexes of the computing power chip at different frequencies to predict performance data of the computing power chip at different frequencies; acquiring to-be-predicted data of the computing power chip under the current working load, predicting the to-be-predicted data based on the optimal prediction regression model, and acquiring a target optimal frequency meeting a preset performance index condition, the target optimal frequency corresponding to optimal energy consumption; and performing adaptive adjustment on the working frequency of the computing power chip under the current working load based on the target optimal frequency. According to the method and the device, the target optimal frequency which meets the performance requirement and is lowest in chip energy consumption can be accurately selected, the energy consumption is reduced while the chip performance is ensured, and the problems of low adjustment precision and high energy consumption of frequency adjustment of a computing chip in different working load performances are solved.

The invention relates to the technical field of cave three-dimensional modeling, in particular to a cave three-dimensional modeling method and system based on multi-sensor fusion, and the method comprises the steps: recognizing a simulation feature network point set, calculating the simulation precision according to the simulation feature network point set and an actual verification network point set, carrying out the point drawing according to the area of a unit test grid and the simulation precision, and carrying out the calculation of the simulation precision. Performing regression analysis on the area precision point set to obtain an area precision regression line; constructing a triangular plane current grid according to a current to-be-measured cave region; performing inner wall secondary sensing sampling on the three-dimensional unit current grid according to predicted regression precision to obtain a secondary three-dimensional sampling point set; and performing three-dimensional correction on the current grid of the three-dimensional unit by using the secondary three-dimensional sampling point set to obtain an initial cave three-dimensional model, and fusing the cave sensing image and the initial cave three-dimensional model to obtain a target cave three-dimensional model. According to the invention, the laser sampling efficiency and the cave three-dimensional modeling precision can be improved.

The invention discloses a solid waste water content prediction system and method based on an infrared thermal imager. The temperature field of the solid waste on the conveyor belt in the natural convection and forced convection moisture evaporation process is monitored and compared in real time through the thermal infrared imager, and the moisture content of the solid waste can be accurately predicted in combination with data processing, environmental parameter correction and prediction of a regression equation. Compared with a traditional sampling detection method, the method has the advantages of being high in real-time performance, non-contact, simple in equipment, good in economical efficiency and the like, can adapt to a complex industrial field environment, and provides a new technical means for intelligent and refined control over waste incineration. The system is simple in structure, high in reliability, good in detection precision and capable of achieving real-time prediction of the water content of the solid waste entering the furnace.

The invention discloses a sensitivity data acquisition method and system based on environmental factors, and the method comprises the steps: obtaining a data set in a preset time period, the data set comprising meteorological data, environmental protection data, pollen concentration data and doctor seeing data; performing model training by using the data set to construct a prediction regression model; calculating each index coefficient and intercept based on the prediction regression model, wherein each index coefficient at least comprises a relative humidity coefficient, an air pressure coefficient, a temperature coefficient, a pollutant concentration coefficient, a pollen concentration coefficient, a precipitation coefficient and a wind speed coefficient; obtaining current environment data and future predicted environment data, wherein the current environment data and the future predicted environment data are in one-to-one correspondence with the index coefficients; and calculating the sensitivity data based on the current environmental data, the future predicted environmental data, the index coefficients and the intercept, so that the sensitivity data can be graded, and protection early warning can be provided for users in time according to the grading condition.

The invention discloses a soft soil area vertical shaft deep foundation pit ground subsidence prediction method and device, and belongs to the field of deep foundation pit engineering. The method comprises the following steps: establishing a two-dimensional finite element model for ground surface settlement analysis under the coupling action of a seepage field and a stress field; the working condition of the model is adjusted to determine main factors affecting ground surface settlement; establishing a regression equation for predicting the maximum wall deflection caused by support excavation and an empirical equation for predicting a deformation ratio based on the main factors; training and predicting parameters in the regression equation and the empirical equation by adopting a machine learning algorithm on the basis of accumulated ground surface settlement data obtained by numerical analysis of the typical construction process of the vertical shaft; determining the maximum surface settlement based on the deflection equation and the deformation ratio equation; and based on the maximum surface subsidence, according to the outside-pit subsidence influence range and the subsidence curve form, the surface subsidence influence area around the foundation pit is partitioned, and surface subsidence is predicted. According to the method and device, the ground subsidence induced by shaft foundation pit construction can be rapidly calculated.

The application provides a driver fatigue prediction method and device, electronic equipment and storage medium, which obtains a monitoring video; obtains facial state features and a fatigue score according to the monitoring video; trains a fatigue prediction regression model by taking the facial state features as an input vector and the fatigue score as a target result; extracts facial state features of a to-be-detected driver from a to-be-detected image or video, inputs the facial state features of the to-be-detected driver into the trained fatigue prediction regression model, obtains a fatigue score corresponding to each time in a future period of time, and then can infer the time required to reach different fatigue levels according to the current time and the time corresponding to different fatigue levels, so as to achieve the purpose of predicting the fatigue condition of the driver in advance, help early warning, avoid vehicle accidents, and the same model can obtain corresponding prediction time according to different fatigue level requirements, and has high expansibility.

The application discloses an agent action prediction method based on multi-scale space perception, comprising map space structure modeling, historical trajectory feature extraction, feature fusion and multi-modal action prediction, wherein: the map space structure modeling uses a multi-scale graph convolutional neural network to extract map features from two-dimensional vector map data in an application scenario map, to obtain high-dimensional map feature information; the historical trajectory feature extraction uses a convolutional neural network and a feature pyramid network to extract high-dimensional trajectory data features of all agents; the feature fusion models and fuses the correlation of high-dimensional map feature information and high-dimensional trajectory data feature information through a self-attention mechanism, to obtain agent trajectory fusion features with direction information; and the multi-modal trajectory prediction uses the agent trajectory fusion features for prediction regression and confidence scoring, to output multi-modal complete trajectory coordinates and corresponding confidence scores of action prediction, so as to provide reasonable auxiliary decision-making for agent action.

The invention discloses a non-invasive pulmonary arterial hypertension hemodynamic monitoring method based on machine learning, and belongs to the technical field of pulmonary arterial hypertension hemodynamic monitoring. The non-invasive pulmonary arterial hypertension hemodynamic monitoring method based on machine learning comprises the following steps: collecting BCG signal data through static data collection equipment; eCG signal data are acquired through dynamic data acquisition equipment; pPG signal data are collected through a photoelectric finger clip; static feature extraction is carried out on static data composed of the BCG signal data and the PPG signal data, and dynamic feature extraction is carried out on dynamic data composed of the ECG signal data and the PPG signal data; and inputting the static characteristics and the dynamic characteristics into a regression model, and training the regression model by taking the hemodynamic parameter CO measured by the right cardiac catheter at the same time as a target to obtain a cardiac displacement prediction regression model. By adopting the non-invasive pulmonary arterial hypertension hemodynamics monitoring method based on machine learning, the problems that an existing pulmonary arterial hypertension hemodynamics monitoring method is complex in operation and cannot meet daily rehabilitation training monitoring use are solved, and the prediction precision is improved.

The present application relates to a kind of tool wear prediction methods based on part geometric feature and cutting process parameters, first complex profile parts are split into simple part geometric feature, for different part geometric feature, establish about part feature parameters, tool parameters and cutting process parameters Tool wear prediction regression model;According to the parameter range set up orthogonal experiment, measure the tool wear under different parameters, then each group tool wear, parameter logarithm is taken and is converted into the form of matrix, input into MATLAB software, the coefficient and exponential part of tool prediction regression model parameter are solved, the coefficient and exponential part logarithm is taken anti-function, obtain the tool wear prediction regression model under different characteristics;Finally, the significance test and residual analysis are carried out on tool wear prediction regression model, and the final tool wear prediction model is obtained.The parameters referred to in the present application are more, the application range is wider, and the tool wear prediction accuracy is higher.

The embodiment of the present invention discloses a method for predicting the strength of drug inhibitory efficacy based on a small data set, which uses adaptive data expansion, reconstructs feature space, and relabels labels as relative size relationships of inhibitory efficacy, thereby converting the predicted accurate numerical value of inhibitory efficacy into a direct prediction of the relative size of inhibitory efficacy, and converting the accurate value prediction regression task into a simpler classification task. At the same time, the adaptive feature space reconstruction preserves the relative size relationship of the physicochemical properties between compounds, and more importantly, significantly weakens the inherent similarity between compounds with similar inhibitory efficacy. Using the data set processed by the present invention, the machine learning model can more easily distinguish similar compounds, and thereby screen out samples with better inhibitory efficacy.

The present invention discloses a twin network target tracking algorithm with enhanced positioning quality information, including: determining the target, obtaining the first frame image, determining the annotation box using known information, cropping the target area, and obtaining the target template; similarity comparison, obtaining the current frame, using the feature extraction subnetwork to obtain the feature map of the target and the search area and perform similarity comparison to obtain a similarity matrix; target positioning, the fusion classification branch performs the classification task based on the similarity matrix, and outputs the similarity prediction score; boundary prediction, the regression branch predicts the target boundary based on the similarity matrix, and selects the best prediction box in combination with the similarity prediction score; outputs the tracking result, outputs the target boundary box, and performs target tracking for the next frame. The present invention can improve the classification and boundary prediction process, introduces the CNN network to make the model more fully utilize the semantic information in the feature map, and the prediction result has a higher credibility.

The invention provides a two-passenger one-dangerous vehicle tracking method and system, and the method comprises the steps: determining whether a vehicle is a two-passenger one-dangerous vehicle or not based on an entrance portal, and issuing a vehicle mark for the vehicle if the vehicle is the two-passenger one-dangerous vehicle; in the vehicle driving process, identifying a two-passenger one-danger vehicle based on the roadside equipment, and recording the driving position of the corresponding vehicle based on the vehicle mark; a plurality of continuous driving positions of the same vehicle are constructed into prediction vectors, the prediction vectors are input into a pre-trained prediction regression model, and the prediction regression model outputs the predicted driving position of the corresponding vehicle at the prediction time point; and judging based on the actual driving position of the vehicle at the predicted time point and the predicted driving position of the predicted time point, and determining whether to give an alarm to the vehicle or not. According to the scheme, road section vehicle passing data is analyzed based on ETC portal vehicle transaction flow data, two-passenger one-danger vehicle whole-course tracking is achieved, an alarm can be given in time, and the supervision safety is improved.

The invention discloses a method and device for predicting the remaining service life of a lithium ion battery, and belongs to the technical field of lithium ion batteries, and the method comprises the steps: obtaining the remaining service life of a target lithium ion battery at a historical moment; taking the remaining service life within a preset time period before any historical moment as model input, taking the remaining service life at the historical moment as model output, and constructing a prediction regression model; carrying out model parameter estimation on the prediction regression model by adopting a variational Bayesian method, and carrying out dynamic optimization adjustment on model parameters by combining a particle filtering method; obtaining the remaining service life in a preset time period before the current moment as model input, and obtaining the remaining service life at the current moment through the constructed prediction regression model; according to the method, reasoning of the prediction regression model is carried out by adopting a variational Bayesian method, dynamic optimization is carried out by adopting a particle filtering method, and high-accuracy prediction of the remaining service life of the lithium ion battery is realized.

This disclosure relates to the field of bridge engineering, specifically to a design method, system, and medium for HPC-RC combined eccentrically compressed columns. The method includes: defining a design space; selecting optimal design data combinations based on engineering specification constraints, bearing capacity constraints, and cost-effectiveness values ​​to construct a design data sample set; constructing a two-branch heterogeneous neural network model, the model including an input layer, a shared feature extraction layer, a diameter prediction classification branch, and a reinforcement area prediction regression branch; constructing a loss function composed of focal loss and mean square error loss; training the two-branch heterogeneous neural network model using the design data sample set; and using the trained two-branch heterogeneous neural network model to predict the diameter and total reinforcement area of ​​the eccentrically compressed column. This disclosure improves the stability and accuracy of the design, reduces the computational burden, increases computational speed, and reduces computational complexity.

The present application discloses a method, system, device, and storage medium for predicting ADHD pathogenic subcutaneous nuclei, wherein the method includes the following steps: obtaining a first magnetic resonance image dataset of each subtype of ADHD case, a second magnetic resonance image dataset of subcutaneous nuclei of normal children, and first scale data of each subtype case; performing structural covariation analysis of the subcutaneous nuclei on the first magnetic resonance image dataset and the second magnetic resonance image dataset to obtain a first abnormal subcutaneous nucleus; obtaining a first contribution score based on the volumes of two subcutaneous brain regions of the first abnormal subcutaneous nucleus; obtaining a trained prediction regression model based on the first contribution score and the first scale data; inputting the first magnetic resonance image of the ADHD case to be predicted into the prediction regression model to obtain a prediction result for the subcutaneous nucleus. This method can provide a scientific basis for the effectiveness of clinical treatment and rehabilitation decisions. This application can be widely used in the field of medical technology.

A system and method evaluate a communication tower for modification. The system includes a prediction module, a scoring module, and an output device. The prediction module receives asset data associated with the communication tower, and implements a prediction model to generate, from the asset data, a model probability score associated with the communication tower. The scoring module generates a tower modification recommendation from the model probability score. The output device outputs the tower modification recommendation. The system receives model training data for training the prediction model to implement a predictive regression model. The method implements the system.

The invention provides a shield tunnel segment floating prediction method, device and equipment, and relates to the technical field of shield tunnel construction, and the method comprises the following steps: S1, constructing a segment floating prediction model database; s2, establishing a prediction regression model based on the CPO-LSSVM, determining a hyper-parameter optimization range in the LSSVM and parameters of a CPO optimization algorithm, and performing hyper-parameter optimization by adopting five-fold cross validation; s3, performing integration of a plurality of weak learners on the prediction regression model by using an Adaboost integrated learner to obtain an optimization model; s4, training and testing the optimization model to obtain corresponding evaluation indexes; and S5, inputting newly collected parameters into the optimization model to obtain a predicted value of the segment floating amount. The method can be used for predicting the maximum floating amount of the duct piece after shield tunnel construction, the prediction result is high in precision and high in efficiency, and the method can be used for guiding on-site construction.

The invention discloses a road loss prediction method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring map information and road test data of a target area; according to the map information, extracting a geographic feature vector of the map information through a geographic pre-training model based on comparative learning training; according to the road test data and the geographic feature vectors, road loss prediction is carried out through a pre-trained road loss prediction regression model, and a road loss prediction value of the target area is obtained.According to the road loss prediction method and device, the road loss prediction is carried out through the geographic pre-training model based on comparative learning training, the direction of model feature extraction can be guided, and the accuracy of road loss prediction is improved. And the pressure of model fitting is reduced, so that the migration scene prediction accuracy is improved.

The invention relates to the technical field of disease risk intelligent prediction, in particular to a chronic kidney disease progress risk prediction system based on machine learning, which comprises an accumulated impact acquisition module used for inputting disease information and physiological index data of a current patient into a disease risk transmission network to acquire an accumulated impact condition; the predicted eGFR calculation module is used for inputting the accumulated impact condition into an eGFR prediction regression network to obtain a predicted eGFR curve; the similarity analysis module is used for calculating a similarity measurement value between the eGFR curves of the current patient and each target historical patient; the risk value calculation module is used for obtaining a renal function attenuation risk value based on the predicted eGFR curve, the similarity measurement value and the eGFR difference value between the current patient and each target historical patient; and the risk judgment module is used for judging whether the current patient has a renal function collapse risk based on the renal function attenuation risk value. According to the invention, early and accurate prediction can be carried out on the CKD progress risk.

The invention discloses a PCB product signal integrity prediction method and device and a storage medium, and belongs to the technical field of PCB manufacturing, and the method comprises the steps: inputting an obtained to-be-predicted PCB design parameter into a pre-trained and optimized prediction regression mathematical model, and outputting a signal integrity prediction result; the training and optimizing method of the predictive regression mathematical model comprises the following steps: determining dielectric loss according to the type of a PCB material; the copper surface roughness is determined according to the copper foil brownification type; training a prediction regression mathematical model by using the dielectric loss, the copper surface roughness and the frequency, the line width spacing, the dielectric thickness and the signal layer thickness which are input in advance; according to the method, big data processing, SI calculation and SI prediction are carried out on the production influence factors before and after the production process by utilizing the prediction regression mathematical model, and the SI risk is prevented, so that the high-risk material number is reduced, the cost loss is reduced, and meanwhile, the product quality is ensured.

The invention provides a pulverized coal bunker output calculation method based on a powder making cooperation system, and relates to the technical field of coal-fired unit powder supply, and the method comprises the steps: collecting the coal quality, CO, fly ash and conventional operation parameters of a unit for calculation, and obtaining the heat in a furnace under the current load; a boiler combustion prediction regression model is established, a parameter prediction value under the target load is obtained, and heat in the target load boiler is obtained; determining the heat distribution proportion of the operation modes of the coal mills under different loads by using CFD simulation, distributing the heat, and calculating the powder supply amount of each coal mill; forming a piecewise fitting calculation function through the output characteristics of the coal mill; and the output of the lower pulverized coal bunker inside and outside the starting delay time of the coal mill is calculated respectively, so that the technical problem of excessive or insufficient fuel supply caused by rough adjustment according to experience in the prior art is solved, and the technical effects of higher adjustment speed, refined pulverized coal supply and improvement of unit operation economy and safety are achieved.

The present invention discloses a brain age prediction method based on multimodal fusion of structural and functional MRI images, belonging to the field of medical image processing technology. The method first uses DenseNet121 to extract spatial structural features from structural magnetic resonance images; at the same time, a functional connectivity matrix is ​​constructed according to the time series of functions, and a graph structure is constructed based on the matrix, in which each node is characterized by the strength of its connection with other nodes, and the edge is converted from the absolute value of the connection strength to a sparse graph representation; then a graph attention network is used to extract functional features, and a cross-attention mechanism is used to fuse the structural and functional features; the gating mechanism fusion result is then used for the brain age prediction regression task. The brain age prediction method of the present invention makes full use of the complementary information of multimodal data and can accurately capture the biological markers of brain aging.

An apparatus for computing a conditional conformal prediction interval for a machine learning point prediction regression model and calibration point predictions forming a distribution of an error around the point prediction regression model in an input space. The apparatus includes a conformal regions circuit configured to compute a quantile regression of the error to compute an approximation of a quantile of the error. The conformal regions circuit is further configured to identify a set of regions in the input space where the distribution within each region in the set of regions is interpretably constant. In one embodiment, the apparatus also includes a conformal prediction circuit configured to compute the conditional conformal prediction interval for the point prediction regression model conditioned on the identified set of regions and the corresponding computed quantile of the error for each region in the set of regions.

The invention discloses an unmanned aerial vehicle operation state multi-threshold evaluation method and system based on uncertainty quantization, and relates to the field of aircraft state evaluation. The problems that in an existing method, the evaluation capability of a key parameter prediction model reflecting the operation state of the unmanned aerial vehicle is insufficient, and evaluation result information given by a single-threshold abnormal state evaluation method is limited are solved. The method comprises the following steps: inserting a Dropout network layer in a prediction regression model, and obtaining a Monte Carlo Dropout prediction model; obtaining an uncertainty estimation result according to a prediction model of Monte Carlo Dropout; calculating an operation state evaluation quantitative score according to the uncertainty estimation result and the unmanned aerial vehicle operation state deviation degree; determining a multi-stage operation state evaluation threshold according to the operation state evaluation quantitative score; and evaluating the operation state of the unmanned aerial vehicle according to the multi-stage operation state evaluation threshold. The method is applied to the unmanned aerial vehicle flight field.

This invention relates to the field of cave 3D modeling technology, specifically a method and system for cave 3D modeling based on multi-sensor fusion. The method includes: identifying a set of simulated feature mesh points; calculating simulation accuracy based on the simulated feature mesh point set and the actual verification mesh point set; plotting points based on the unit test mesh area and simulation accuracy to obtain an area accuracy point set; performing regression analysis on the area accuracy point set to obtain an area accuracy regression line; constructing a triangular plane current mesh based on the current cave region to be measured; performing secondary sensing sampling on the inner wall of the current 3D unit mesh based on the predicted regression accuracy to obtain a secondary 3D sampling point set; using the secondary 3D sampling point set to perform 3D correction on the current 3D unit mesh to obtain an initial cave 3D model; and fusing the cave sensing image and the initial cave 3D model to obtain a target cave 3D model. This invention can improve laser sampling efficiency and cave 3D modeling accuracy.

The invention provides a model generation method and an orientation angle acquisition method. The model generation method comprises the following steps: acquiring a sample image containing a target object and annotation information corresponding to the sample image; obtaining a real observation angle of the target object according to the annotation information; a classification probability truth value and a regression truth value are generated for each angle interval in a plurality of preset angle intervals according to the real observation angle, and the angle intervals are set to have at least two angle intervals associated with any observation angle; inputting the sample image into a prediction model to obtain a prediction classification probability value and a prediction regression value about each angle interval output by the prediction model; obtaining a loss value according to the difference between a classification probability truth value and the predicted classification probability value and the difference between the predicted regression value and the regression truth value; and updating parameters of the prediction model according to the loss value so as to optimize the prediction precision of the observation angle of the target object by the prediction model.

This invention relates to the field of building structural performance evaluation, specifically a machine learning-based method for evaluating the seismic resistance of unreinforced masonry buildings. The method includes considering the spatial distribution and cumulative overall evaluation index of damage; pre-setting different levels of damage states; using an adaptive multi-scale progressive nonlinear dynamic response analysis method to obtain the median critical peak acceleration when the index reaches the limit value of each level under different working conditions; training a seismic resistance prediction regression model using the median values ​​of different levels; training a seismic resistance classification model using the median values ​​of each level; fitting adjustment parameters for service life; using the prediction model to obtain the median value prediction results corresponding to each level during evaluation; and inputting the prediction results optimized with the adjustment parameters into the classification model to obtain the building grade. This invention solves the technical problems of high computational cost of traditional evaluation methods, insufficient sample size and poor interpretability of machine learning methods, and difficulty in practical application of existing evaluation methods.