50 results about "Meta-regression" patented technology

A method of grading an essay using an automated essay scoring system is provided. The method comprises the steps of deriving a set of predetermined features from the essay, wherein the predetermined feature set comprises one or more features that are independent from the test prompt, scoring the feature set with a scoring equation, wherein a multiple regression analysis with graded essay data produces weights for the scoring equation, generating a raw score for the essay; and processing the raw score for the essay into a score category based on an adaptive cutoff algorithm. Also provided is a method of generating a model in which to grade essays, wherein the data used to generate the model is independent from the test prompt or essay topic.

Information presented to a user via an information access system is ranked according to a prediction of the likely degree of relevance to the user's interests. A profile of interests is stored for each user having access to the system. Items of information to be presented to a user are ranked according to their likely degree of relevance to that user and displayed in order of ranking. The prediction of relevance is carried out by combining data pertaining to the content of each item of information with other data regarding correlations of interests between users. A value indicative of the content of a document can be added to another value which defines user correlation, to produce a ranking score for a document. Alternatively, multiple regression analysis or evolutionary programming can be carried out with respect to various factors pertaining to document content and user correlation, to generate a prediction of relevance. The user correlation data is obtained from feedback information provided by users when they retrieve items of information. Preferably, the user provides an indication of interest in each document which he or she retrieves from the system.

The invention discloses a method for optimizing a unit regression test case set based on a control flow diagram. The method comprises the following steps: A, finding out modified points corresponding to a modified part from the control flow diagram of a tested unit being tested unit before and after; B, screening out test cases, of which the execution paths pass through the modified points, to serve as one part of the regression test case set, namely a selected test case set (Tselected), and running all the test cases in the set; C, calculating all the reachable successor nodes of the modified points and selecting a node subset N which is not covered in the running process of the step B; D, if the N is null, executing the step F, and if the N is not null, executing the step E; E, selecting one node from the N to serve as a coverage target generation test case, adding the coverage target generation test case into a new test case set (Tnew) and then completely running, updating the N, deleting the nodes covered in the running process and returning to the step D; and F, finishing construction of the regression test case set (TR). By application of the method, the efficiency of the regression test can be improved and the effectiveness and the sufficiency of the regression test are guaranteed.

The invention discloses a comprehensive electric energy meter verification method and system based on an improved least square method. The method comprises the steps: generating a scatter diagram of original data, deleting an abnormal value, and obtaining sample data; carrying out Pearson correlation analysis and VIF inspection on independent variables in the sample data; determining a multi-colinearity existence range between the independent variables; checking the multiple collinearity by fitting a sample error average regression line and a median regression line; performing multivariate regression analysis according to an inspection result, and preliminarily determining a regression equation; checking the credibility of the regression equation, and determining a data regression model; correcting the data regression model through residual analysis; and normalizing the weight of each variable, calculating an influence weight of each variable on the error, and substituting the influence weight into the data regression model to carry out comprehensive verification on the electric energy meter. According to the invention, whether the metering error of the electric energy metering device exceeds a standard specified range can be effectively verified, and the reliability and stability of the electric energy metering device are ensured.

The invention relates to an industrial process soft measuring method based on an integrated type independent element regression model, and is used for non-Gaussianity industrial process data. A conventional non-Gaussianity soft measuring regression modeling method requires selection of a non-quadratic function to measure non-Gaussianity, but different industrial process data or objects can cause a fact that enough experiential knowledge for guiding the selection of the non-quadratic function is hard to acquire in an actual application. The method provided by the invention is advantageous in that different soft measuring models are acquired by comprehensively and fully using different non-quadratic functions for training, and the problem of the selection of the non-quadratic function is effectively prevented; a final prediction result is acquired by accumulating weighting coefficients, and then prediction precision of a corresponding soft measuring model is not affected by the selection of the non-quadratic function. The prediction effect of the soft measuring model is greatly improved, and therefore key indexes or quality indexes during the process can be accurately and reliably predicted.

An improved method for forecasting and modeling product demand for a product. The forecasting methodology employs a causal methodology, based on multiple regression techniques, to model the effects of various factors on product demand, and hence better forecast future patterns and trends, improving the efficiency and reliability of the inventory management systems. A product demand forecast is generated by blending forecast or expected values of the non-redundant causal factors together with corresponding regression coefficients determined through the analysis of historical product demand and factor information. The improved method provides for the saving and updating of previously calculated intermediate regression analysis results and regression coefficients, significantly reducing data transfer time and computational efforts required for additional regression analysis and coefficient determination.

The invention discloses a method for predicting the single flight noise of an airport, which comprises the following steps of: giving corresponding weight to single flight noise data of the airport at first; discriminatively establishing the learning ability of a sample through the weight; training a plurality of different learning models; selecting better learning models therein to be integrated; and abstracting the noise prediction problem of the airport as the multiple regression analysis problem in mathematics. A system for realizing the prediction method disclosed by the invention comprises an airport single flight noise data acquisition module, an airport noise data storage module and an airport single flight noise prediction module. According to the method disclosed by the invention, the instability of the current prediction method based on the single learning algorithm is overcome; the prediction precision and the prediction stability are effectively increased; compared with the general optimization and all integrations of the single learning algorithm, the prediction speed is increased; and the airport noise prediction practicability is increased.

The invention discloses a construction method and prediction method of an industry electric quantity demand prediction model. The construction method of the industry electric quantity demand prediction model comprises the steps that: the first leading factor of industry electric quantity increase is obtained according to a set industry electric quantity correlation analysis program, wherein the industry electric quantity correlation analysis program is a combination of a correlation rule mining program and a principal component analysis program; according to a pre-established historical industry electric quantity correlation analysis data set, the second dominant factor of the industry electric quantity increase is extracted; and according to the first dominant factor and the second dominant factor, an ARIMA model and the multiple regression model are corrected and coupled, so that the industry electric quantity demand prediction model is obtained. Compared with a traditional method, the method has the advantages of greatly improved prediction precision, high applicability and feasibility, can be popularized to other industries.

A method of building a patent valuation model includes acquiring patent information, processing the patent information and separately performing a plurality of multiple regression analyses in which aplurality of key valuation elements preset for a valuation index are dependent variables, calculating multiple representative values, calculating a representative value of a plurality of regression coefficients for each independent variable of a plurality of multiple regression models calculated through the plurality of multiple regression analyses, and generating a valuation model for the valuation index by building a valuation model in which the calculated representative values are coefficients for the respective independent variables.

The invention discloses a regional credit prediction method and system based on multiple regression and a time sequence. The method comprises the following steps: obtaining influence factor data; obtaining factor processing data; according to the factor processing data, establishing a multiple regression model, and obtaining a first output result of the multiple regression model; determining whether the first output result meets a first preset threshold value or not, and when the first output result does not meet the first preset threshold value, obtaining a step-by-step forward screening instruction; sending the screening output result to an expert screening unit to obtain an expert screening result; substituting the expert screening result into the multiple regression model to obtain a regional factor screening result; establishing a time sequence model; obtaining a second output result of the time sequence model; and obtaining a regional credit prediction result according to the regional factor screening result and the second output result. The technical problems that in the prior art, a regional credit prediction model is unscientific in prediction, low in accuracy, complex in model composition, high in cost, not beneficial to maintenance and lack of quantitative analysis on influence factors are solved.

The invention discloses an electric energy meter comprehensive verification method and system based on a stepwise regression algorithm, and the method comprises the following steps: carrying out the visualization of an original data distribution scatter diagram, deleting an abnormal value, and generating sample data; drawing a normal probability graph according to the sample data and carrying outnormal distribution inspection processing; fitting a sample error average regression line and a median regression line according to the sample data after inspection processing, and screening the sample data; performing multiple regression analysis on the screened sample data by using the stepwise regression algorithm, and determining a regression equation; checking the credibility of the regression equation, and determining a data regression model; performing independent variable weight normalization processing on the regression equation, and calculating the influence weight of the independentvariable on the error based on the relevancy; and substituting the influence weight of the independent variable on the error into the data regression model and then carrying out comprehensive verification on the electric energy meter. According to the invention method, the electric energy meter is comprehensively calibrated, and the reliability and stability of the electric energy metering deviceare ensured.

The invention discloses a data calibration method for a portable air quality monitor. Independent variables and dependent variables are set; an independent variable monitoring data matrix is constructed; a partial least squares regression model is creatively used for calibrating the air quality monitoring data; in view of diversity and cross influence of error factors generated by monitoring data,the advantages of principal component analysis, typical correlation analysis and a linear regression analysis method in multiple regression analysis are integrated; the calibrated data is compared with real data, the calibrated data has the advantages that absolute errors and overall relative errors are obviously reduced, the simulation is repeatable, the precision of the monitoring data is improved, calibrated data information can be obtained only by importing the monitoring data into a program of a computer through a USB interface, and the method is simple and easy to operate.

The invention discloses a multivariate linear regression analysis method and a multivariate linear regression analysis system based on big data, belongs to the field of multivariate regression analysis algorithms, and aims to solve the problems that influences caused by user portrait feature changes cannot be known, selection cannot be carried out according to the user portrait features adopted bythe influences, and comprehensive analysis cannot be carried out on the big data. Based on big data, a user portrait is obtained in real time, and user portrait characteristics are obtained; Based onthe user portraits in the two adjacent stages, the conversion rate of the user portraits is calculated from the previous stage to the next stage; a multivariate linear regression model is establishedbased on a multivariate regression analysis algorithm, all user portrait features obtained in real time and corresponding conversion rates; And based on the established multivariate linear regressionmodel, the conversion rate of the to-be-calculated user portrait features is calculated in the previous stage to the next stage. The method is used for carrying out multivariate linear regression analysis on big data to obtain a current analysis result.

The invention provides a lie judgment method based on an eye movement technology. The lie judgment method is characterized by comprising the following steps of: step S1, collecting the eye movement data of a subject; S2, preprocessing the eye movement data; S3, carrying out equalization processing; S4, carrying out row normalization processing; S5, performing Pearson correlation analysis on eye movement data evaluation indexes, and screening out a normalized average pupil diameter and normalized average fixation duration with the maximum correlation as influence indexes; S6, establishing a multiple regression model; S7, collecting the eye movement data of a to-be-tested person, calculating the normalized average pupil diameter and the normalized average fixation duration of the to-be-tested person, and inputting the normalized average pupil diameter and the normalized average fixation duration into the multiple regression model; and S8, outputting the evaluation value of a lie to be measured. According to the lie judgment method, the multiple regression model is constructed by using the eye movement technology, the eye movement data of the to-be-tested person is collected for lie judgment, and the lie judgment method has the advantages of being simple in operation, high in measurement efficiency, high in objectivity and the like.

The invention relates to a network delay prediction method based on correlation analysis. The method comprises the following steps of integrating history measurement data in a network control system, calculating a correlation coefficient of each independent variable for a dependent variable, and carrying out arrangement according to absolute values in a descending order; taking the independent variable and the dependent variable with the maximum correlation coefficient absolute value to establish a unitary regression equation; and importing the independent variables into the regression equation one by one according to the correlation coefficient absolute values, when a significance test result indicates that obvious correlation exists between the independent variables and the dependent variables, importing the variables into the regression equation, and updating the regression equation. Through correlation analysis in data mining, a great deal of control network data is integrated and analyzed, the correlation existing between the network delay and other parameters in the network is found, the correlation is applied to online prediction, so the network delay can be calculated according to the monitorable data in the current network, and the stability of the control network is ensured.

The invention discloses a label big data-based talent recommendation algorithm under a system framework, which comprises the following steps: step 1, establishing a multiple regression mathematical model based on statistical analysis of data; step 2, establishing an expected multi-dimensional data table B containing multiple factors, deriving a distribution condition of a candidate content and achievement index IB, a corresponding weight WB and a random error space vector, and taking the distribution condition as a quantifiable index reference for evaluating and judging the candidate content and achievement; step 3, performing deep mining on behavior keyword data, and performing denoising operation on keyword data without behavior factors; and step 4, updating and constructing a multi-dimensional comprehensive evaluation table S, and analyzing variables by taking data in the basic table as metadata and quantitative dimensions, namely attributes, contents, achievements and behaviors, of the basic table as factors. Login simulation is realized by adopting a simulation technical means, an expected effect hotspot distribution series mathematical model is established through key discrete data such as attributes, contents, achievements and behaviors of candidates, and the mathematical model is abstracted into an algorithm.

The invention provides a metabolic channel analysis modeling method based on mass spectrum metabonomics. The method comprises the following steps: S1, collecting biological samples of normal organisms and diseased organisms, converting mass spectra obtained by collecting the samples into mzML files, and carrying out centralization, denoising and alignment treatment on the mass spectrum files to obtain a two-dimensional data matrix of metabolites of the samples; s2, performing centralization and Unite Variance normalization processing on the two-dimensional data matrix of the metabolites of the samples and a sample category vector matrix; s3, acquiring a metabolite-channel mapping relation, and optimizing a channel weight coefficient; and 4, sorting the channels, adjusting a penalty factor, determining selected frequencies of the channels, and sorting the channels by using the selected frequencies of the channels. According to the method provided by the invention, the metabolites are grouped into mutually overlapped channels, a partial least square method and Group Lasso are combined to establish a regression model, and the channel weight coefficient and the penalty factor are introduced, so that multiple regression based on 'grouping sparse' is realized.

The invention discloses a regression analysis-based news competitiveness analysis method and a visualization device. The method comprises the following steps: obtaining a news event development trend growth rate-based competitiveness model through a zero-sum game and news event development trend growth rate; carrying out multivariate regression analysis on the competitiveness model and expanding into a multivariate regression model; evaluating the competitiveness between news events through a semipartial correlation coefficient and quantifying the competitiveness; and estimating the competitiveness model by the multivariate regression model through calculating the fitting degree of the competitiveness model. The device comprises an obtaining module, an expansion module, an evaluation and quantification module and an estimation module. According to the regression analysis-based news competitiveness analysis method and the visualization device, analysis of news data is achieved; and an experiment proves that the regression analysis-based news competitiveness analysis method and the visualization device have relatively high fitting degree; and the regression analysis-based news competitiveness analysis method and the visualization device are applicable to visualization analysis of the news event competitiveness on news media.

The invention discloses a human body composition and visceral fat content prediction method, which comprises the following steps: determining a segmented impedance model, and collecting an impedance value; constructing input data and output data according to the standard body fat rate gold standard; selecting characteristic values; determining an eigenvalue vector combination; increasing constraint conditions, and increasing the result credibility of the regression model; and obtaining a human body composition and visceral fat content prediction model. The method provided by the invention is compared with results of KLPS regression, a traditional cross validation SVR regression algorithm and a multiple regression method in visceral fat area prediction, and the effectiveness and superiority of the method are verified through correlation and mean square error analysis.

Disclosed are a modeling method and device for distributed machine learning, and equipment. By setting a corresponding target modeling unit for an acquired training data set, the target modeling unit is one of the following: a classification modeling unit, a regression modeling unit or a clustering modeling unit; a parameter selection mode and a verification mode are configured for each modeling algorithm included in the target modeling unit to obtain a plurality of initial modeling algorithms and a plurality of groups of training subsets and verification subsets; the training subsets in each group are respectively input into each initial modeling algorithm, and a prediction model of each initial modeling algorithm is obtained according to a distributed task scheduling strategy; and each prediction model is evaluated according to the evaluation parameters to obtain a target initial modeling algorithm meeting a preset condition. The target initial modeling algorithm is trained according to the training data set to obtain the target prediction model, and the to-be-predicted data is predicted through the target prediction model, so that the time of automatic modeling is shortened, the skill requirements on analysts are reduced, and machine learning is more intelligent.

The invention discloses a medical data missing processing method, device and equipment based on a multiple regression model, which can solve the technical problems of poor data filling quality and insufficient accuracy when the existing data filling method is used for filling data at present. The method comprises the steps of obtaining a missing tuple corresponding to medical data, determining a complete tuple matched with the medical data type corresponding to the missing tuple, wherein the missing tuple is composed of missing attributes and partial complete attributes, and the complete tuple is composed of complete attributes; generating a preset number of multiple regression models by utilizing the complete attributes contained in the complete tuple; determining a candidate filling attribute combination about the missing attribute in the missing tuple; and screening out a target candidate filling attribute combination with the minimum total fitting error on the multiple regression model from the candidate filling attribute combinations, and filling the missing tuple by using the target candidate filling attribute combination. The method and the device are suitable for filling the missing medical data.

The invention provides an intelligent disease cognition system based on a uric acid value domain range. The system comprises a data acquisition module used for acquiring uric acid data to be diagnosed, obtaining a uric acid standard index value range, processing the uric acid data to be diagnosed according to the uric acid standard index value range, and obtaining index data; a regression model module used for establishing a multivariate logistics regression algorithm, acquiring body data of the user to be diagnosed and calculating a probability value between the body data of the user to be diagnosed and the index data; a matching module used for establishing a regular expression and matching the index data, the probability value and the to-be-diagnosed user body data with the uric acid standard index value range and the corresponding disease information by utilizing the regular expression; and a diagnosis module used for generating a corresponding diagnosis report according to the matching result. According to the method, the multivariate logistic regression algorithm and the regular expression are established, so that the recognition of uric acid diseases is realized, and the cognitive rate and accuracy are improved.

The embodiment of the invention relates to a method for analyzing influence factors of doctor seeing of remote patients. The method comprises the following steps: acquiring case data, geographic element data and social economic data of the remote patients seeing a doctor in a target city; based on the case data, the geographic element data and the social economic data, determining a plurality of first influence factors of medical seeing of the remote patient; screening the plurality of first influence factors by using a multiple regression model to obtain a plurality of second influence factors; taking the doctor seeing rate of the patient flowing out of the prefecture-level city to the target city as a dependent variable, taking multiple second influence factors as independent variables, performing fitting analysis on the dependent variable and the independent variables by utilizing an OLS regression model, a GWR regression model and an MGWR regression model, and determining a target regression model with the highest fitting degree; and determining a target influence factor from the second influence factors based on the significance test variable of the target regression model so as to analyze the target influence factor. According to the technical scheme provided by the invention, the influence factor condition of medical treatment of the remote patient can be accurately analyzed.

The invention provides application of multiple regression analysis in tax decision. Factors affecting tax decision are analyzed; related data is collected and sorted; decisions or recommendations are made using a regression analysis model; the tax department is assisted in solving decision process problems, such as high emotional influence and lack of data support. The factors affecting tax decision feature complexity, dynamism, finiteness and the like, so that the problems such as variety and high upgrade frequency occur in the acquisition process of basic data. The application of the multiple regression analysis has the advantages that no special requirements on basic data are required, and both discrete variables and continuous variable are available; an analysis result is an event probability more scientific and reasonable, perdition precision is high, and prediction results are stable.

The invention discloses a gas reservoir recovery ratio prediction method based on multiple regression. The method comprises the following steps: selecting target reservoir rock, processing the reservoir rock into parallel samples, and preprocessing the parallel samples; according to the characteristics of the selected target reservoir, determining a plurality of single factors influencing reservoir gas reservoir recovery in advance, and acquiring target parameters of each single factor through corresponding experiments; analyzing the influence of each single factor on the recovery ratio of the gas reservoir, and screening out a plurality of single factors which mainly influence the recovery ratio of the gas reservoir; and calculating a prediction value of the recovery ratio based on a multiple regression model. According to the method, the output process of reservoir gas is restored more truly, multiple main control factors are substituted through multiple regression to obtain the recovery ratio, and the calculation error of the recovery ratio is reduced; compared with a traditional recovery ratio prediction method, dependence on field production data is greatly reduced in the method; and meanwhile, the method is suitable for gas reservoir prediction of various gas reservoirs, different development modes and different development stages, and an application range is wide.

The invention provides a stock market valuation method based on a gray prediction algorithm and a multiple regression analysis model. The method comprises the following steps: S1, carrying out the first-order accumulation of data, forming a data sequence, and obtaining a corresponding gray differential equation; s2, solving parameters of a grey differential equation; s3, establishing a generated data sequence model to solve a differential equation to obtain a prediction model; s4, establishing an original data sequence model, namely, generating a simulation sequence value of the original datasequence through accumulation and subtraction; s5, solving valuation indexes. According to the method, the grey prediction algorithm and the multiple regression model are combined, in the field of absolute valuation, the improved ground algorithm can make up for the defects of a DCF stock right free cash flow model, and the stock market valuation problem is solved more accurately and quickly.

The present invention relates to a method and system for measuring and correcting human body temperature based on multiple regression. The method includes the following steps: measuring the body surface temperature of the test personnel entering the temperature measuring area to obtain the body surface temperature, and obtaining the internal environment temperature at this time , the external environment temperature and the body temperature of the tester; based on big data and multiple regression analysis methods, under the conditions of the corresponding external environment temperature and the internal environment temperature, the body surface temperature of the tester and the body temperature of the human body The corresponding relationship between body temperatures is analyzed to obtain a human body temperature measurement correction model; based on the human body temperature measurement correction model, the body temperature of the person to be measured is corrected to calculate the body temperature of the person to be measured. The present invention has a good improvement effect on the problems of missing and false alarms caused by inaccurate conversion of human body temperature from body surface temperature measured by an infrared thermometer, and can realize accurate measurement of pedestrian body temperature in a daily detection environment and ensure rapid passage.

The present invention relates to a method of diagnosing and further determining the course of the disease in an individual with adult Still's disease. The method includes the following steps: (a) providing a blood sample and peripheral blood mononuclear cells; (b) detecting the expression level of the protein biomarker IL-18 in the blood sample; (c) detecting the expression of long-chain non-coding RNA biomarkers MIAT or THRIL in the peripheral blood mononuclear cells; (d) calculating the A-score by using the predicted function obtained by multiple regression analysis with the expression levels of IL-18, MIAT and THRIL detected in steps (b) and (c); (d) calculating the A-score by using the predicted function obtained by multiple regression analysis with the expression levels of IL-18, MIAT and THRIL detected in steps (b) and (c); and (e) comparing the A-score calculated in step (d) with a critical value, obtained by the receiver operating characteristic curve method, corresponding to the area under the ROC curve reaching a maximum, the A-score above the critical value value indicates that the individual suffers from adult Still's disease. The sequence of step (b) and step (c) can be reversed, which does not affect the judgment of the embodiment of the present invention and the final result.