Stepwise Regression Algorithm Research Articles

Apolipoprotein L-I is positively associated with hyperglycemia and plasma triglycerides in CAD patients with low HDL

Apolipoprotein L-I (apoL-I) is present on a subset of HDL particles and is positively correlated with plasma triglycerides (TGs). We measured plasma apoL-I levels in coronary artery disease (CAD) subjects with low HDL who were enrolled in an angiographic CAD prevention trial. At baseline, apoL-I levels (n = 136; range, 2.2-64.1 mug/ml) were right skewed with a large degree of variability. Multivariate analysis for biological determinants of apoL-I revealed that the log of VLDL-TG (+0.17; P < 0.05) and hyperglycemia (HG; +0.26; P < 0.005) independently predicted apoL-I level. Hyperglycemic patients (n = 24) had mean apoL-I levels >50% higher than normoglycemic subjects (n = 112; 13.2 vs. 8.3 mug/ml, respectively; P < 0.001). No relationship between apoL-I level and change in CAD was found (r = 0.06, P = 0.49). Simvastatin-niacin therapy did not alter apoL-I levels (n = 34; P = 0.27), whereas antioxidant vitamins alone increased apoL-I by >50% (n = 36; P < 0.01). Genotyping of a known apoL-I polymorphism (Lys166Glu) did not independently account for any of the variability in apoL-I levels. In conclusion, we found TG and HG to be the strongest predictors of apoL-I within a dyslipidemic CAD population. These data provide further characterization of the novel HDL-associated apoL-I.

Building neutron cross-section dependencies for few-group reactor calculations using stepwise regression

Approximation of few-group neutron cross-sections by functions of burnup and thermal-hydraulics parameters of a fuel cell is considered. The cross-section is written as a sum of two terms: the base cross-section, which depends only on burnup and is computed under the nominal reactor core conditions, and the deviation, which depends on burnup and thermal-hydraulics variables of the cell. A one-dimensional dependence of the base cross-section is interpolated by a cubic spline. Multi-dimensional dependencies of the deviation are approximated by a polynomial. Construction of the polynomial is performed by a best-fitting selection of the polynomial terms using the stepwise regression algorithm. The number of terms to satisfy a user-given accuracy of approximation is minimized. As an example, approximation of a set of two-group macro and micro cross-sections as functions of burnup, coolant and fuel temperature, coolant density and boron concentration is considered for a fuel pin cell of a VVER reactor. The constructed five-dimensional polynomial approximating cross-sections within 0.05% tolerance has about 20 terms for fast group cross-sections and 50 terms for thermal group cross-sections. The error of approximation is verified on the two data sets: the initial data used for approximation and the test data being computed on randomly selected points. Mean square and maximum errors are comparable for all the cross-sections for both sets of data. These results show that the initial data can be applied to control the approximation error.

The Impact of Radar Data on Short-Term Forecasts of Surface Temperature, Dewpoint Depression, and Wind Speed

Abstract A statistical system that uses surface observations and radar data to provide 1-, 3-, and 6-h forecasts of temperature, dewpoint depression, and wind speed is developed. Application of the system to independent data demonstrates that the radar data result in a 3%–6% reduction in the mean-squared error (MSE) of the temperature forecasts when compared with that when only surface observations are used as predictors. The addition of the radar data similarly provides a 2%–4% reduction in MSE of the dewpoint depression forecasts. However, there is no difference in skill when the radar data are included in forecasts of wind speed. The majority of the radar-data predictors selected by the stepwise regression algorithm for inclusion in the forecast system are from areas that are poorly sampled by surface observations. This result supports the argument that more uniform sampling (better coverage) of the environment is playing a role in the improvement of the forecasts. Moreover, analysis of individual even...

Modeling categorical variables by logistic regression.

To demonstrate the use of logistic regression in health care research. Forward and backward stepwise logistic regression algorithms were systematically applied to a real-world data set comprising 301 cancer patients and a set of explanatory variables. Four variables were identified as effective predictors of pain reporting by cancer patients during chemotherapy: fatigue, depression, severity of colds or viral infections, and insomnia. The 4-predictor model was validated by (a) significance tests of regression coefficients at p<0.05, (b) significant improvement of this model over competing models, and (c) goodness of fit indices. Logistic regression is useful for health-related research in which outcomes of interest are often categorical.

The vascularized iliac crest and bone graft anatomy, indications and results

This paper introduces new CFD-based correlations for ribbed roughened solar air heater. An absorber plate of SAH (solar air heater) attached with circular ribs is analyzed in this article. ANSYS Fluent v16 is utilized to analyze the turbulent airflow for different arrangements of rib. All results are based on CFD (computational fluid dynamics) simulations. The analysis is carried out within the limits of Re from 3800 to 18,000 and P from 10 to 25. The resulting data is reduced into correlations using a linear stepwise regression algorithm. The developed equations for artificially roughened heater of solar air predict all the data for friction factor and Nusselt number within ±5% relative absolute deviation.

The Convergence of Efroymson's Stepwise Regression Algorithm

Abstract The stepwise regression algorithm that is widely used is due to Efroymson. He stated that the F-to-remove value had to be not greater than the F-to-enter value, but did not show that the algorithm could not cycle. Until now nobody appears to have shown this. To prove that the algorithm does converge, an objective function is introduced. It is shown that this objective function decreases or can occasionally remain constant at each step in the algorithm, and hence the algorithm cannot cycle provided that Efroymson's condition is satisfied.

A fundamental equation for dichlorodifluoromethane (R-12)

A new fundamental equation for dichlorodifluoromethane (R-12) is presented. The equation is valid for temperatures from 175 K to 600 K with pressures to 70 MPa. Temperatures are given on the new International Temperature Scale of 1990. Newly correlated ancillary functions representing the vapor pressure, saturated liquid density and saturated vapor density were used to insure proper behavior of the equation of state at the phase equilibrium boundary. A multi-property stepwise linear least squares regression algorithm was used to determine the coefficients of the equation of state. Data forms used in the fit include experimental P- ϱ -T, coexistence densities, isobaric heat capacity and velocity of sound. Extended corresponding states methods were used to augment the experimental database. Comparisons are made to the experimental data to verify the accuracy of the formulation. It is suggested that this equation be used in comparative studies of systems using the new ozone-safe refrigerants as working fluids. The equation can also serve as a reference equation in the extended corresponding states prediction of the thermodynamic properties of the ozone-safe refrigerants.

Identification of non-linear processes using reciprocal multiquadric functions

In this paper radial basis function (RBF) networks are used to model general non-linear discrete-time systems. In particular, reciprocal multiquadric functions are used as activation functions for the RBF networks. A stepwise regression algorithm based on orthogonalization and a series of statistical tests is employed for designing and training of the network. The identification method yields non-linear models, which are stable and linear in the model parameters. The advantages of the proposed method compared to other radial basis function methods and backpropagation neural networks are described. Finally, the effectiveness of the identification method is demonstrated by the identification of two non-linear chemical processes, a simulated continuous stirred tank reactor and an experimental pH neutralization process.

Evaluation of success probability in multiple seam room-and-pillar mining

A complete evaluation of multiple seam mining potential for any coal property requires the analysis of historical data and the controlling mining conditions. The major problem affecting multiple seam mining is interaction between workings in the contiguous seams. The success or failure of multiple scam mining can be evaluated using statistical techniques. A computer-based stepwise multiple regression algorithm is used to develop an equation for predicting the probability of success in two seam-mining situations. The model provides a measure of risk involved in mining multiple seams. The data used to develop the model are from a historical record of attempted multiple seam mining in the Eastern United States. The variables significantly influencing success or failure of multiple seam mining are percent sandstone in the interburden, ratio of interburden thickness to lower seam thickness, time delay, interburden thickness, thickness of both seams, depth of cover to upper seam, columnated pillars, columnated entries and the leading seam. The model developed was found to be quite satisfactory as a predictor of the probability of success within the range for which it was developed.

The stepwise regression algorithm seen from the statistician’s point of view

abstract (introduction; abridged): the purpose of this paper is to explain in statistical terms how the algorithm works which is used in computer programmes for stepwise regression. stated more precisely, the purpose is to do this for one of severalpossible, closely related variants of the algorithm. the paper will contain nothing substatially new but may yet perhaps shed some new light on well-known facts. there are two reasons why the author has thought it worthwile writing this paper. one is that stepwise regression programme descriptions are usually written by computer people who tend to use flow-chart terminology rather than statistical terminology. an example in case is the standard reference, efroymson (1960). the other reason for writing the paper is the author's belief that the computer algorithm may serve as a vehicle for teaching students some important aspects of multiple regression. section 7 below is written with this latter point of view in mind.;