98 results about "Predictor variable" patented technology

The invention discloses a wastewater treatment plant modeling method and system based on a physical information neural network. The method comprises the following steps: acquiring and inputting working condition parameters and inlet water quality data to a neural network model, and performing forward calculation to obtain predicted outlet water state variables and key kinetic parameters; determining a first deviation based on a difference value between the predicted effluent and an actual measurement value, and substituting the predicted variable and the parameter into an activated sludge mechanism differential equation set to calculate a residual error so as to obtain a second deviation; performing weighted summation on the two deviations to form a total optimization target, and performing iterative training according to the total optimization target to obtain a final model; different working condition parameters are substituted into the model for multiple times of forward deduction, and finally in combination with energy consumption data, the working condition with the lowest energy consumption meeting the effluent standard is determined as the optimal process control parameter. By implementing the technical scheme provided by the invention, the actual operation energy consumption meeting the effluent quality standard is reduced.

The invention relates to the technical field of automobile design, in particular to an intelligent generation type design method of an anti-collision beam, which comprises the following steps: firstly, performing experimental analysis on the anti-collision beam to obtain a section image comprising the anti-collision beam and performance response data corresponding to the section image; performing image recognition and text extraction on the section image; fusing the recognized image and the extracted text by using a multi-modal multi-layer fusion model to obtain a multi-modal design variable; training the constructed conditional variation network by using the multi-modal design variables and the performance response data corresponding to the multi-modal design variables to obtain a conditional variation model; generating design variables by using the conditional variation model, and predicting performance response data corresponding to the variables; optimal performance response data are screened out from the obtained performance response data, and then the optimal design scheme of the anti-collision beam to be designed is obtained. According to the method, the optimal design scheme is determined by using the conditional variation network, and the efficiency and the precision of optimization design are effectively improved.

The invention relates to a training method of a respiratory tract infection disease progress and prognosis prediction model. The training method comprises the following steps: extracting mRNA from peripheral blood of a target patient, and carrying out transcriptome sequencing to obtain a sequencing result; based on the ferroptosis related gene set, comparing ferroptosis score differences of two groups of patients with community-acquired pneumonia and sepsis, and screening corresponding ferroptosis related genes with statistical significance from a sequencing result; screening out genes meeting preset conditions from the ferroptosis related genes based on LASSO regression; and establishing an RTI clinical outcome prediction model through logistic regression by taking whether the patient is sepsis or not as an outcome dichotomy variable and taking the screened gene expression quantity as a prediction variable. According to the invention, after the prediction model is subjected to machine learning screening such as LASSO and the like, the core feature with the highest prediction value is reserved, so that the risk of over-fitting of the model on training data is reduced.

The invention discloses a multivariate time sequence missing variable prediction method, program and system based on a denoising diffusion model, and a storage medium, and belongs to the field of deep learning. The invention provides a novel model named DiffMissing based on DDPM. The novel model comprises a forward diffusion module, a context condition encoder module, a denoising network module and a prediction module. According to the model, two view angles of variables and time are fused, and the consistency of the interiors of the variables and the time can be more accurately ensured, so that information in data is more comprehensively understood, and the capability of predicting missing variables is remarkably improved. The DiffMissing prediction model provided by the invention has a remarkable technical effect, and can effectively improve the function of reconstructing the missing variable and the function of predicting the variable of the time sequence data.

The invention relates to a multi-resolution geological data conversion method and system based on ensemble learning, and belongs to the technical field of geological information processing, and the conversion method comprises the steps: obtaining a low-resolution geological data set containing a target element, and a high-resolution geological data set not containing the target element; performing spatial grid aggregation processing on the high-resolution geological data set to generate a predictive variable matrix which is spatially aligned with the low-resolution geological data set; training a Stacking integrated regression model by taking the predictive variable matrix as input and a target element value in the low-resolution geological data set as an output target; inputting the high-resolution geological data set into the fusion prediction model, and outputting a preliminary prediction value of a target element; and performing spatial error correction on the target element preliminary prediction value based on the target element actual value of the low-resolution geological data set to generate corrected high-resolution target element data. According to the invention, multi-scale and multi-source heterogeneous geological data can be effectively integrated.

The invention discloses a time sequence reconstruction and machine learning coupled chlorophyll monitoring data breakpoint restoration method, and belongs to the technical field of water quality monitoring. The invention discloses a chlorophyll monitoring data breakpoint restoration method based on coupling of time sequence reconstruction and machine learning, and the method comprises the following steps: S1, collecting water quality monitoring data, and cleaning the monitoring data to obtain preprocessed data; s2, performing time sequence reconstruction on the preprocessed data to obtain a weekly average 1 data set; s3, respectively constructing a radial basis function neural network model and a back propagation neural network model by taking the chlorophyll concentration as a response variable and the conventional water quality parameter as a predictive variable; s4, performing performance evaluation on each model by taking a root mean square error, an average absolute percentage error, goodness of fit and relative error distribution statistics as evaluation indexes, and screening out an optimal model; and S5, applying the conventional water quality parameters in the breakpoint interval of the chlorophyll monitoring data in the water body to the optimal model, and outputting the restored chlorophyll concentration value to complete the dynamic restoration of the breakpoint.

The invention relates to the technical field of liver disease prediction, in particular to a metabolism-related fatty liver disease intelligent prediction method and system and a storage medium. The method comprises the following steps: collecting multi-source data, respectively obtaining basic demographic information, laboratory indexes and prediction indexes, and extracting quantitative and qualitative tongue picture parameters; performing variable screening on the tongue picture parameters and the clinical indexes, and determining key prediction variables; obtaining a key variable value according to the key prediction variable, obtaining a prediction result of the occurrence risk of the metabolism-related fatty liver disease, and outputting the prediction result; the system comprises a multi-source data acquisition module, a variable screening module and a prediction result acquisition module. By means of the mode, the intelligent tongue picture parameters and the clinical indexes are fused, and the effect of early prediction of the metabolism-related fatty liver diseases is achieved.

The invention discloses a soil humidity inversion method based on an ADF model, and belongs to the field of soil humidity inversion. The method comprises the following steps: S1, selecting a soil humidity inversion area, and obtaining an in-situ SSM data set of the corresponding area; s2, selecting a prediction variable, extracting the prediction variable from the corresponding prediction data set, and combining the prediction variable with the in-situ SSM data set to form a sample data set; s3, constructing an ADF model, and performing training and verification based on the sample data set to estimate SSM according to the prediction variable; s4, when the ADF model outputs the SSM estimation value, determining a prediction variable which has the maximum influence on SSM estimation; and S5, evaluating the generalization ability of the ADF model based on the sample data set. According to the ADF model, prediction accuracy, generalization ability and interpretability are realized at the same time.

The invention discloses a metallogenic prediction method and device based on fluid parameters and fluid field modeling, a medium and a product, and relates to the field of metallogenic prediction. The method comprises the following steps: firstly, carrying out lithofacies research on a fluid inclusion sample in a target area to determine a metallogenic stage; analyzing and testing the fluid inclusion samples at different mineralization stages to obtain a plurality of fluid parameters; exploratory data analysis is conducted on the multiple fluid parameters, and key fluid parameters closely related to the content of the main metallogenic elements are screened out; modeling the numerical relationship between the key fluid parameters and the main metallogenic elements by adopting a machine learning or deep learning algorithm, and determining an optimal numerical model; processing the key fluid parameters by adopting different spatial interpolation algorithms, and establishing a three-dimensional fluid field model; according to the three-dimensional fluid field model and the optimal numerical model, three-dimensional metallogenic prediction is carried out in combination with the metallogenic geological elements, and a prospecting target area is determined; the fluid parameters are introduced as predictive variables, so that the precision of metallogenic prediction is improved.

The invention provides a water soluble organic nitrogen property prediction method based on spectrum fingerprint information and machine learning and application. The method specifically comprises the following steps: research area selection, sample collection, pretreatment and related index detection. Preliminarily verifying the data set, screening representative spectrum indexes, and constructing a spectrum fingerprint information matrix; dividing a training set and a test set, using a machine learning method to construct prediction models by using the training set, and using the test set to perform performance evaluation on the prediction models; a prediction model with better performance is selected, a Shapley method is used for calculating the SHAP value of each characteristic value, and the key characteristic value with the dominant model influence degree is selected as a prediction variable to indicate the property characteristics of the soluble organic nitrogen in the research area. According to the method provided by the invention, the properties of the complex soluble organic nitrogen can be conveniently and efficiently preliminarily judged, the understanding of the current situation of the water ecological environment is further improved, and a certain technical support is provided for formulating water ecological environment protection measures.

The present disclosure relates to a server for calculating an order quantity on the basis of material usage and demand prediction of an unmanned store, and a control method therefor, wherein a machine learning model may be trained on the basis of material usage data of the unmanned store, and the order quantity may be calculated by calculating, on the basis of a plurality of first prediction variables and second prediction variables, a result value including the order quantity and an available sales period for each product of the unmanned store, by using the trained machine learning model.

A machine learning system and method for optical critical dimension measurement. From a training set of spectra and references, features are extracted and subjected to regression analysis to generate predictor variables. Using feature functions, inverse feature functions, a machine-learning predictor component and masks, a machine-learning optical critical dimension explainer is generated. A wafer is analyzed by metrology tools and the machine-learning predictor component calculates a critical dimension inference from measured spectra. Theoretical spectra are then generated by the predictor component based upon a modification of the critical dimension inference. The measured spectra are compared to the theoretical spectra and the fit of the measured spectra to the theoretical spectra is evaluated for acceptance. The results of the comparison and analysis is output in human readable form.

The invention relates to the technical field of Internet of Things, and provides an industrial production prediction model construction method and system. A standard industrial data set is subjected to periodic decomposition and trend smoothing processing according to a time period to generate a time sequence change sequence, and the time sequence change sequence is subjected to power load segmentation adjustment and feature extraction to obtain a production power load capacity sequence and a capacity feature set. Carrying out industrial state type identification and state probability calculation according to the production power load sequence and the capacity feature set to generate a state probability sequence, and integrating the production power load sequence and the capacity feature set according to the state probability sequence to obtain a prediction variable set of each industrial state type; and constructing an industrial production prediction model according to the state probability sequence and the prediction variable set. According to the method, the state condition regression model is constructed through temperature and calendar stripping, mixing fusion and state recognition, and the precision, robustness and interpretability of industrial output proximity prediction are improved.

This invention discloses a lightweight modeling method, apparatus, system, and storage medium for high-dimensional data meteorological forecasting tasks, belonging to the technical field of computation, extrapolation, or counting. The method constructs prediction vectors and target variables from the original high-dimensional data required for the target meteorological forecasting task. Through an improved variational autoencoder strategy, it reconstructs latent predictor variables that influence or potentially influence the target variables from the prediction vectors. It then selects latent variables highly correlated with the original predictor variables to construct a candidate predictor variable dataset. After removing candidate variables that are significantly uncorrelated with the target variables and those with strong linear correlations from the candidate predictor variable dataset, it performs key predictor variable selection and constructs a lightweight prediction model based on the key predictor variables. This invention significantly reduces the dimensionality of the model input data while maintaining prediction accuracy, and enhances the model's prediction accuracy by constructing predictor variables that influence the target variables and have clear physical meaning.

The application discloses a kind of wind turbine blade icing mass prediction methods based on data driving, method includes: based on preset value range, obtain sampling data;Simulation is carried out by CFD simulation software, and the corresponding icing mass data set under different parameter conditions is obtained;Icing mass data set obtained by simulation is compared with the data set of laboratory environment simulation result and is corrected;According to the icing mass data set obtained after correction, respectively to each variable Application polynomial fitting obtains the order relationship of the variable and icing mass after no-parameter system identification, obtains the function relationship of each parameter and icing mass;The variable value to be predicted is obtained by actual environmental information, and the variable value to be predicted is predicted according to the function relationship, and the icing prediction result is obtained.The application has the advantages of low cost, strong real-time, high precision, small data processing amount when using, etc., and can be widely applied in data processing technical field.

The invention relates to the technical field of biology, in particular to application of an SNP (Single Nucleotide Polymorphism) molecular marker combination to identification of Shouguang chickens and non-Shouguang chickens. The SNP molecular marker combination is a genotype result of 16 loci, the 16 SNP molecular marker combination is used as a predictive variable, a Shouguang chicken or a non-Shouguang chicken is used as a predictive value, and 70% of individuals in a sample to be detected are used as a training group to carry out 20 times of random sampling to train a support vector machine model. Results show that the prediction precision of the identification of the Shouguang chicken is 99% when the 16 SNP molecular markers are combined for identification. And a more powerful technical support is provided for identification and genetic breeding of the Shouguang chicken.

The invention discloses a method for constructing a prediction model for subcutaneous fat hyperplasia caused by insulin injection. The method comprises the following steps: S1, selecting a plurality of participants for questionnaire survey according to inclusion standards; s2, performing ultrasonic inspection on all participants; s3, performing feature screening on the predicted variables by adopting LASSO regression; s4, developing a fat hyperplasia prediction model by using three algorithms of random forest, extreme gradient lifting and logistic regression, and selecting the fat hyperplasia prediction model with the optimal performance for research; and S5, explaining the influence of the predictive variables on the fat hyperplasia prediction model through an SHAP method. According to the method, a machine learning-based insulin injection-induced fat hyperplasia prediction model is created, the constructed extreme gradient lifting machine learning model shows relatively high efficiency in the aspect of predicting the occurrence of fat hyperplasia of the diabetic patient, and the diabetic patient with high-risk fat hyperplasia can be accurately identified.

The application discloses a kind of dabigatran bleeding risk prediction methods based on machine learning, it is related to computer-aided drug risk management technical field, the method is by obtaining patient baseline and follow-up data, multiple imputation method is handled missing value;With HAS-BLED score as the basis, candidate variables are screened from potential risk factors using LassoCV algorithm, and anemia, insulin use and antifungal agent use are determined as new prediction variables through clinical correlation analysis to construct a set of prediction variables;Finally, a random forest or gradient boosting algorithm is used to train the model to output the bleeding risk assessment results of the individual to be tested.The application combines machine learning algorithm, improves the prediction accuracy of dabigatran bleeding risk in Chinese non-valvular atrial fibrillation patients, solves the problem of insufficient prediction accuracy of traditional HAS-BLED score, and provides a more reliable basis for clinical individualized anticoagulant therapy decision.

A computer implemented system for predicting a property or classification associated with document data. The system has a data extraction module configured to receive input document data and extract from the input document data a plurality of data sets, each data set having data of one of a plurality of data types. The system also has processing pathways each configured to process one of the plurality of data sets to generate a vector output representative of the data set processed by the processing pathway. The system has a vector concatenation layer configured to concatenate the vector outputs of each processing pathway to generate a concatenated vector, and a plurality of predictions heads. Each prediction head is configured to process the concatenated vector to generate a prediction variable indicative of a property or classification predicted to be associated with the input document data.

The application discloses a prediction model for pulmonary arterial hypertension diagnosis and severity judgment based on pulmonary arteriography, a construction method and application. By collecting and analyzing pulmonary arteriography and test information of pulmonary arterial hypertension patients and controls, the application constructs a clinical prediction model for pulmonary arterial hypertension diagnosis and severity judgment. The diagnosis model and the severity judgment model both contain five independent prediction variables. After the model is constructed, the model is visualized through a nomogram, and the performance is evaluated through an ROC curve, a DCA curve and a calibration curve. The model is tested through a machine learning method, and the results show that the model has good performance. In addition, through correlation analysis of continuous variables, the application further constructs a linear regression equation with mean pulmonary arterial pressure as the dependent variable. The application provides a new clue for non-invasive screening of pulmonary arterial hypertension and has important significance for expanding the disease screening population and early diagnosis.

The present invention discloses a calculation method and device, electronic device, and storage medium for time downscaling of river material flux. The method comprises: collecting water quality data, hydrological data, and meteorological data of a river; calculating the daily scale flux of river material based on the data and setting it as a response variable; accumulating and calculating the monthly scale flux based on the daily scale flux, and forming a prediction variable of a model with the hydrological data and meteorological data; constructing a random forest based on the response variable and the prediction variable, and calculating the feature importance of the prediction variable; optimizing the prediction variable based on the feature importance result of the prediction variable; retraining the random forest model based on the optimized prediction variable, fitting the daily scale material flux, and evaluating the overall performance based on the linear correlation result with the response variable value; inputting the new monthly scale flux of the river and the hydrological data and meteorological data into the random forest prediction model, estimating the predicted value of the daily scale flux of the river material, and completing the material flux downscaling.

The invention discloses a sewage effluent ammonia nitrogen prediction method based on a local enhanced optimization echo state network, and the method comprises the steps: obtaining technological parameter data and effluent ammonia nitrogen data in a sewage treatment process, and carrying out the preprocessing, thereby obtaining a prediction variable set and a candidate auxiliary variable set; forming a primary auxiliary variable set; calculating a covariance matrix of variables in the primary auxiliary variable set, and extracting principal components of which eigenvalues are greater than 1 through eigenvalue decomposition to obtain a core auxiliary variable set containing 4-10 variables; constructing a local enhancement optimization echo state network based on the core auxiliary variable set; inputting a core auxiliary variable in a test sample into the local enhanced optimization echo state network, and outputting an effluent ammonia nitrogen predicted value; and evaluating the prediction precision by adopting a mean square error to complete the ammonia nitrogen prediction of the effluent of the sewage. According to the method, a model capable of adaptively cutting redundant neurons and optimizing sewage quality prediction is constructed according to nonlinear dynamic change characteristics of ammonia nitrogen concentration, so that high-precision and low-delay prediction of ammonia nitrogen in urban sewage treatment effluent is realized.

The invention relates to atmospheric chemical reaction calculation substitution and acceleration in an atmospheric chemical model, in particular to an atmospheric chemical reaction calculation substitution model and a construction method thereof, the model has a high-precision reappearance chemical reaction process, and in 19-day off-line prediction of 279 predictive variables of a case, the proportion of the predictive variable decision coefficient exceeding 0.99 is 96%; the model can realize continuous stable prediction for at least 10 days; compared with an original GEOS-Chem chemical integrator, the ChemKNet has the advantages that the time consumed by calculation on a single CPU can be saved by about 79%, only about 1% of the original calculation time is needed on a single GPU, and the calculation efficiency of large-scale simulation is greatly improved; the model supports coupling and continuous prediction; the model is provided with a full-chema chemical mechanism covering GEOS-Chem, and is provided with global simulation and a full-chema chemical mechanism covering GEOS-Chem.

In some embodiments, a video decoder decodes a video from a bitstream. The video decoder accesses a bin string representing a partition of the video and processes each coding tree unit (CTU) in the partition to generate decoded values in the CTU. The process includes initializing context variables for context adaptive binary arithmetic coding (CABAC), Rice parameter variables, and palette predictor variables only when the CTU is the first CTU in a tile, or the CTU is the first CTU in a slice, or parallel coding is enabled and the CTU is the first CTU in a CTU row of a tile. No other initialization is performed for these variables. The video decoder decodes the CTU based on the initialized context variables, Rice parameter variables, and palette predictor variables.

The invention belongs to the technical field of artificial intelligence and power system crossing, and discloses an electricity price prediction method based on automatic hyper-parameter and feature optimization and a related device. The electricity price prediction method comprises the steps of predicting input data based on a selected time period, performing electricity price prediction by using an electricity price prediction model, and obtaining an electricity price prediction result of the selected time period; the electricity price prediction model is defined, constructed and trained according to the optimal configuration scheme; the obtaining process of the optimal configuration scheme comprises the steps of obtaining a target prediction variable used for model construction and selected input features useful for target prediction to form a structured data set, defining a mathematical optimization space based on the structured data set, and searching in the mathematical optimization space to obtain the optimal configuration scheme. According to the technical scheme disclosed by the invention, the technical problems of low efficiency, high dependence on artificial experience, suboptimal configuration combination, fragmentation of an optimization process, difficulty in maximization of prediction precision and the like in an existing electricity price prediction model construction process are solved.

A method for improving multi-target error equalization degree by using GAT-BILSTM&CNN-LSTM model, comprising: obtaining load data and air temperature data of a target area; data cleaning is performed on the load data and air temperature data; feature engineering processing is performed on the data cleaned data to obtain historical data; a GAT-BILSTM algorithm network is set and a CNN-LSTM algorithm network is set; model combination: using GAT-BILSTM to train the historical data, calculating the prediction error of each prediction variable of the load rate prediction variable of the target area, then calculating the average error, then combining the load rate prediction variable exceeding the average error with the result of CNN-LSTM, the combination method is weighted by using error reciprocal method, and the remaining load rate prediction variable remains unchanged to output the result. The present application can solve the problem that the prediction errors of each target in the multi-target prediction model are too different, and at the same time, the overall error is reduced as much as possible.

The invention discloses a primary metastatic breast cancer postoperative BCSS column diagram prediction model and a construction method thereof, and belongs to the technical field of biomedicine. The construction method comprises the following steps: collecting clinical pathological characteristics of a patient after a preliminary diagnosis metastatic breast cancer operation; a Cox regression model is adopted to carry out regression analysis on clinical pathological features of a patient without radiotherapy, and independent BCSS predictive variables are screened out; and determining an index for establishing a column graph prediction model from the BCSS prediction variable, and establishing the prediction model. The prediction efficiency and clinical practicability of the model are evaluated through the C index, the ROC curve, the calibration curve and the clinical decision curve, the patients are divided into low-risk groups, medium-risk groups and high-risk groups according to the total score of the column diagram, and finally the benefit conditions of each risk group receiving postoperative radiotherapy are observed. The column graph model constructed by the method has good distinction degree and calibration degree, provides a basis for a clinician to formulate an individualized postoperative radiotherapy strategy, and has important clinical significance for improving long-term survival of a patient with metastatic breast cancer initially diagnosed.

Systems and methods for automatically determining and indicating specialty medications are disclosed. In some embodiments, a disclosed method includes: obtaining a request from a user for specialty determination of a medication; extracting, based on predictor variables of a machine learning model, relevant data of the medication from at least one database; computing, using the machine learning model, a probability score for the medication based on the relevant data, wherein the probability score indicates a probability that the medication will be determined as a specialty medication; generating, based on the probability score, a specialty indicator indicating whether the medication will be determined as a specialty medication; and transmitting at least one of the specialty indicator or the probability score to the user.

A computer program product, a monitoring method for a production process in a production system, a production system provided with a monitoring device and a method for configuring the monitoring device, which is configured to monitor the production process in the production system for quality forecasting, wherein a process variable is detected and, based on this, a plurality of aggregation variables are determined, where a quality parameter is also detected, multiple aggregation variables and the quality parameter are combined to form a workpiece data set, a respective interdependence between a respective aggregation variable and the quality parameter is additionally determined and a forecasting variable is determined from among the aggregation variables based on the determined interdependence, the forecasting variable is specified as the variable of the production process to be monitored for the operation of the monitoring device.

Abstract In some aspects, a computing system can generate and optimize a neural network for risk assessment. The neural network can be trained to enforce a monotonic relationship between each of the input predictor variables and an output risk indicator. The training of the neural network can involve solving an optimization problem under a monotonic constraint. This constrained optimization problem can be converted to an unconstrained problem by introducing a Lagrangian expression and by introducing a term approximating the monotonic constraint. Additional regularization terms can also be introduced into the optimization problem. The optimized neural network can be used both for accurately determining risk indicators for target entities using predictor variables and determining explanation codes for the predictor variables. Further, the risk indicators can be utilized to control the access by a target entity to an interactive computing environment for accessing services provided by one or more institutions. Abstract 20 26 20 53 15 06 J ul 2 02 6 A b s t r a c t 2 0 2 6 2 0 5 3 1 5 0 6 J u l 2 0 2 6