The problem of choosing a good parameter setting for a better generalization performance in a learning task is the so-called model selection. A nested uniform design (UD) methodology is proposed for efficient, robust and automatic model selection for support vector machines (SVMs). The proposed method is applied to select the candidate set of parameter combinations and carry out a k-fold cross-validation to evaluate the generalization performance of each parameter combination. In contrast to conventional exhaustive grid search, this method can be treated as a deterministic analogue of random search. It can dramatically cut down the number of parameter trials and also provide the flexibility to adjust the candidate set size under computational time constraint. The key theoretic advantage of the UD model selection over the grid search is that the UD points are "far more uniform" and "far more space filling" than lattice grid points. The better uniformity and space-filling phenomena make the UD selection scheme more efficient by avoiding wasteful function evaluations of close-by patterns. The proposed method is evaluated on different learning tasks, different datasets as well as different SVM algorithms.

The initial emittable concentration (Cm,0) is a key parameter characterizing the emission behaviors of formaldehyde from building materials, which is highly dependent on temperature but has seldom been studied. Our previous study found that Cm,0 is much less than the total concentration (C0,total, used for labeling material in many standards) of formaldehyde. Because Cm,0 and not C0,total directly determines the actual emission behaviors, we need to determine the relationship between Cm,0 and C0,total so as to use Cm,0 as a more appropriate labeling index. By applying statistical physics theory, this paper derives a novel correlation between the emittable ratio (Cm,0/C0,total) and temperature. This correlation shows that the logarithm of the emittable ratio multiplied by power of 0.5 of temperature is linearly related to the reciprocal of temperature. Emissions tests for formaldehyde from a type of medium density fiberboard over the temperature range of 25.0-80.0 °C were performed to validate the correlation. Experimental results indicated that Cm,0 (or emittable ratio) increased significantly with increasing temperature, this increase being 14-fold from 25.0 to 80.0 °C. The correlation prediction agreed well with experiments, demonstrating its effectiveness in characterizing physical emissions. This study will be helpful for predicting/controlling the emission characteristics of pollutants at various temperatures.

The emission rate is considered to be a good indicator of the emission characteristics of formaldehyde and volatile organic compounds (VOCs) from building materials. In contrast to the traditional approach that focused on an experimental study, this paper uses a theoretical approach to derive a new correlation to characterize the relationship between the emission rate and temperature for formaldehyde emission. This correlation shows that the logarithm of the emission rate by a power of 0.25 of the temperature is linearly related to the reciprocal of the temperature. Experimental data from the literature were used to validate the derived correlation. The good agreement between the correlation and experimental results demonstrates its reliability and effectiveness. Using the derived correlation, the emission rate at temperatures other than the test condition can be obtained, greatly facilitating engineering applications. Further analysis indicates that the temperature-related emission rate of other scenarios, i.e., the standard emission reference and semi-volatile organic compounds (SVOCs), also conforms to the same correlation as that of formaldehyde. The molecular dynamics theory is introduced to preliminarily understand this phenomenon. Our new correlation should prove useful for estimating the emission characteristics of chemicals from materials that are subject to changes in temperature.

No abstract

Background:The prevalence of asthma and allergic diseases has increased rapidly in urban China since 2000. There has been limited study of associations between home environmental and lifestyle factors with asthma and symptoms of allergic disease in China. Methods:In a cross-sectional analysis of 2,214 children in Beijing, we applied a two-step hybrid Least Absolute Shrinkage and Selection Operator (LASSO) algorithm to identify environmental and lifestyle-related factors associated with asthma, rhinitis and wheeze from a wide range of candidates. We used group LASSO to select variables, using cross-validation as the criterion.