Improving the precision of the accumulated oxygen deficit using VO2-power regression points from below and above the lactate threshold

The accumulated oxygen deficit (AOD) method assumes a linear VO₂-power relationship for exercise intensities increasing from below the lactate threshold (BLT) to above the lactate threshold (ALT). Factors that were likely to effect the linearity of the VO₂-power regression and the precision of the estimated total energy demand (ETED) were investigated. These included the slow component of VO₂ kinetics (SC), a forced resting y-intercept and exercise intensities BLT and ALT. Criteria for linearity and precision included the Pearson correlation coefficient (PCC) of the VO₂-power relationship, the length of the 95% confidence interval (95% CI) of the ETED and the standard error of the predicted value (SEP), respectively. Eight trained male and one trained female triathlete completed the required cycling tests to establish the AOD when pedalling at 80 rev/min. The influence of the SC on the linear extrapolation of the ETED was reduced by measuring VO₂ after three min of exercise. Measuring VO₂ at this time provided a new linear extrapolation method consisting of ten regression points spread evenly from BLT and ALT. This method produced an ETED with increased precision compared to using regression equations developed from intensities BLT with no forced y-intercept value; (95%CI (L), 0.70±0.26 versus 1.85±1.10, P<0.01; SEP(L/Watt), 0.07±0.02 versus 0.28±0.17; P<0.01). Including a forced y-intercept value with five regression points either BLT or ALT increased the precision of estimating the total energy demand to the same level as when using 10 regression points, (5 points BLT + y-intercept versus 5 points ALT + y-intercept versus 10 points; 95%CI(l), 0.61±0.32, 0.87±0.40, 0.70±0.26; SEP(L/Watt), 0.07±0.03, 0.08±0.04, 0.07±0.02; p>0.05). The VO₂-power regression can be designed using a reduced number of regression points... ABSTRACT FROM AUTHOR

Solving regression problems using competitive ensemble models

The use of ensemble models in many problem domains has increased significantly in the last fewyears. The ensemble modeling, in particularly boosting, has shown a great promise in improving predictive performance of a model. Combining the ensemble members is normally done in a co-operative fashion where each of the ensemble members performs the same task and their predictions are aggregated to obtain the improved performance. However, it is also possible to combine the ensemble members in a competitive fashion where the best prediction of a relevant ensemble member is selected for a particular input. This option has been previously somewhat overlooked. The aim of this article is to investigate and compare the competitive and co-operative approaches to combining the models in the ensemble. A comparison is made between a competitive ensemble model and that of MARS with bagging, mixture of experts, hierarchical mixture of experts and a neural network ensemble over several public domain regression problems that have a high degree of nonlinearity and noise. The empirical results showa substantial advantage of competitive learning versus the co-operative learning for all the regression problems investigated. The requirements for creating the efficient ensembles and the available guidelines are also discussed.

Improving an inverse model of sheet metal forming by neural network based regression

An inverse model for a sheet meta l forming process aims to determine the initial parameter levels required to form the final formed shape. This is a difficult problem that is usually approached by traditional methods such as finite element analysis. Formulating the problem as a classification problem makes it possible to use well established classification algorithms, such as decision trees. Classification is, however, generally based on a winner-takes-all approach when associating the output value with the corresponding class. On the other hand, when formulating the problem as a regression task, all the output values are combined to produce the corresponding class value. For a multi-class problem, this may result in very different associations compared with classification between the output of the model and the corresponding class. Such formulation makes it possible to use well known regression algorithms, such as neural networks. In this paper, we develop a neural network based inverse model of a sheet forming process, and compare its performance with that of a linear model. Both models are used in two modes, classification mode and a function estimation mode, to investigate the advantage of re-formulating the problem as a function estimation. This results in large improvements in the recognition rate of set-up parameters of a sheet metal forming process for both models, with a neural network model achieving much more accurate parameter recognition than a linear model.

There's SEM and 'SEM' : a critique of the use of PLS regression in information systems research

In disciplines other than IS, the use of covariance-based structural equation modelling (SEM) is the mainstream method for SEM analysis, and for confirmatory factor analysis (CFA). Yet a body of IS literature has developed arguing that PLS regression is a superior tool for these analyses, and for establishing reliability and validity. Despite these claims, the views underlying this PLS literature are not universally shared. In this paper the authors review the PLS and mainstream SEM literatures, and describe the key differences between the two classes of tools. The paper also canvasses why PLS regression is rarely used in management, marketing, organizational behaviour, and that branch of psychology concerned with good measurement – psychometrics. The paper offers some practical options to Australasian researchers seeking greater mastery of SEM, and also acts as a roadmap for readers who want to check for themselves what the mainstream SEM literature has to say.

Publication selection bias in minimum-wage research? A meta-regression analysis

Card and Krueger's meta-analysis of the employment effects of minimum wages challenged existing theory. Unfortunately, their meta-analysis confused publication selection with the absence of a genuine empirical effect. We apply recently developed meta-analysis methods to 64 US minimum-wage studies and corroborate that Card and Krueger's findings were nevertheless correct. The minimum-wage effects literature is contaminated by publication selection bias, which we estimate to be slightly larger than the average reported minimum-wage effect. Once this publication selection is corrected, little or no evidence of a negative association between minimum wages and employment remains.

Unions and profits : a meta-regression analysis

The effect of unions on profits continues to be an unresolved theoretical and empirical issue. In this paper, clustered data analysis and hierarchical linear meta-regression models are applied to the population of forty-five econometric studies that report 532 estimates of the direct effect of unions on profits. Unions have a significant negative effect on profits in the United States, and this effect is larger when market-based measures of profits are used. Separate meta-regression analyses are used to identify the effects of market power and long-lived assets on profits, as well as the sources of union-profit effects. The accumulated evidence rejects market power as a source of union-profit effects. While the case is not yet proven, there is some evidence in support of the appropriation of quasi-rent hypothesis. There is a clear need for further American and non-American primary research in this area.

Estimating the total energy demand for supra-maximal exercise using the VO2-power regression from an incremental exercise test

The VO_2-power regression and estimated total energy demand for a 6-minute supramaximal exercise test was predicted from a continuous incremental exercise test. Sub-maximal VO_2- power co-ordinates were established from the last 40 seconds(s) of 150-second exercise stages. The precision of the estimated total energy demand was determined using the 95% confidence interval (95% CI) of the estimated total energy demand. The linearity of the individual VO_2-power regression equations was determined using Pearson's correlation coefficient. The mean 95% CI of the estimated total energy demand was 5.9±2.5 mL O₂ Eq•^-1kg•min^-1, and the mean correlation coefficient was 0.9942±0.0042. The current study contends that the sub-maximal VO_2-power co-ordinates from a continuous incremental exercise test can be used to estimate supra-maximal energy demand without compromising the precision of the accumulated oxygen deficit (AOD) method.

Meta-regression analysis as the socio-economics of economics research

Meta-regression analysis (MRA) provides an empirical framework through which to integrate disparate economics research results, filter out likely publication selection bias, and explain their wide variation using socio-economic and econometric explanatory variables. In dozens of applications, MRA has found excess variation among reported research findings, some of which is explained by socio-economic variables (e.g., researchers’ gender). MRA can empirically model and test socio-economic theories about economics research. Here, we make two strong claims: socio-economic MRAs, broadly conceived, explain much of the excess variation routinely found in empirical economics research; whereas, any other type of literature review (or summary) is biased.

Managerial ownership concentration and agency conflict using logistic regression approach : evidence from Bursa Malaysia

This paper adopted logistic regression model to examine the relationship between level of managerial ownership concentration and agency conflict which are proxied by level of risk, firms leverage and firms dividend policy. The study covers a period of 5 years from 1997 through 2001. The study is based on the 100 blue-chip stocks, majority of which are derived from CI components. The findings suggest a positive and significant association between level of level of risk at lower level and managerial ownership while a negative and significant association is also evidenced between risk at higher level and managerial ownership concentration. While debt policy which serves as positive monitoring substitute for agency conflict is found to be positive and significant explaining the level of ownership concentration. Furthermore, dividend policies, which also serve as monitoring, substitute to reduce agency conflict between manager and external shareholders do not appear to have any significant impact on managerial ownership. On the other hand, the level of institutional ownership, which serves as external monitoring force, is found to have inverse impact on level of managerial ownership concentration. This is marginally significant at 10 level (p=.12). The findings, in part explain the argument that the managerial ownership help reduce agency conflict between outside equity holders and managers.

Regression of solid tumors by engineered overexpression of von Hippel-Lindau tumor suppressor protein and antisense hypoxia-inducible factor-1alpha

The von Hippel-Lindau tumor suppressor protein (pVHL) suppresses tumor formation by binding the alpha subunits of hypoxia-inducible factors (HIFs) responsible for stimulating tumor angiogenesis and glycolysis, targeting them for ubiquitination and proteasomal destruction. Loss of pVHL leads to the development of sporadic renal cell carcinomas (RCCs). In the present study, we sought to determine whether engineered overexpression of pVHL in tumors other than RCC can inhibit tumor growth, either as a monotherapy, or in combination with antisense HIF-1alpha therapy. Intratumoral injection of subcutaneous EL-4 thymic lymphomas with an expression plasmid encoding pVHL resulted in the downregulation of HIF-1alpha and vascular endothelial growth factor (VEGF). There was a concomitant reduction in tumor angiogenesis and increased tumor cell apoptosis due in part to downregulation of Bcl-2 expression. VHL therapy resulted in the complete regression of small (0.1 cm diameter) tumors whereas, in contrast, large (0.4 cm diameter) EL-4 tumors were only slowed in their growth. Nevertheless, large tumors completely regressed in response to intratumoral injection of a combination of antisense HIF-1alpha and VHL plasmids. Combination therapy resulted in increased losses of HIF-1alpha, VEGF, and tumor blood vessels, and increased tumor cell apoptosis. These novel results suggest that synergistic therapies that simultaneously block the expression or function of HIF-1alpha, and enhance the expression or function of VHL may be beneficial in the treatment of cancer.

OWA operators in linear regression and detection of outliers

We consider the use of Ordered Weighted Averaging (OWA) in linear regression. Our goal is to replace the traditional least squares, least absolute deviation, and maximum likelihood criteria with an OWA function of the residuals. We obtain several high breakdown robust regression methods as special cases (least median, least trimmed squares, trimmed likelihood methods). We also present new formulations of regression problem. OWA-based regression is particularly useful in the presence of outliers.

OWA operators in regression problems

We consider an application of fuzzy logic connectives to statistical regression. We replace the standard least squares, least absolute deviation, and maximum likelihood criteria with an ordered weighted averaging (OWA) function of the residuals. Depending on the choice of the weights, we obtain the standard regression problems, high-breakdown robust methods (least median, least trimmed squares, and trimmed likelihood methods), as well as new formulations. We present various approaches to numerical solution of such regression problems. OWA-based regression is particularly useful in the presence of outliers, and we illustrate the performance of the new methods on several instances of linear regression problems with multiple outliers.