Entity

Time filter

Source Type


Li J.,CAS Institute of Policy and Management | Wei L.,CAS Institute of Policy and Management | Wei L.,China National Institute of Standardization | Li G.,CAS Institute of Policy and Management | And 2 more authors.
Decision Support Systems | Year: 2011

Credit risk analysis has long attracted a great deal of attention from both academic researchers and practitioners. However, because of the recent financial crisis, this field continues to draw ever increasingly attention. A multiple kernels multi-criteria programming approach based on evolution strategy (ES-MK-MCP) is proposed for credit decision making in this study. We introduce a linear combination of kernel functions to enhance the interpretability of credit classification models, and propose an alternative to optimize the parameters based on the evolution strategy. For illustration purpose, two UCI credit card data sets are used to verify the effectiveness and feasibility of the proposed model. As the experimental results reveal, the proposed ES-MK-MCP model is an efficient tool for credit risk analysis, especially for decision makers to identify the most relevant features. © 2010 Elsevier B.V. All rights reserved. Source


Gao S.,University of Auckland | Gao S.,China National Institute of Standardization | Lewis G.D.,University of Auckland | Ashokkumar M.,University of Melbourne | And 2 more authors.
Ultrasonics Sonochemistry | Year: 2014

A simple theoretical model based on shear forces generated by the collapse of the ultrasound cavities near the surface of a microorganism is proposed. This model requires two parameters which take into account the number of acoustic cavitation bubbles, and the resistance of the cell wall of the microorganism to the shear forces generated by bubble collapse. To validate the model, high-power low frequency (20 kHz) ultrasound was used to inactivate two microorganisms with very different sizes, viz., a bacterium, Enterobacter aerogenes and a yeast, Aureobasidium pullulans. The inactivation ratio was experimentally measured as a function of sonication time for different ultrasound power and for different initial cell numbers. For both E. aerogenes and A. pullulans the Log of the inactivation ratio decreased linearly with sonication time, and the rate of inactivation increased (D-value decreased) with the increase in sonication power. The rate of inactivation was also found, for both microorganisms, to increase with a decrease in the initial cell number. The fits, obtained using the proposed model, are in very good agreement with the experimental data. © 2013 Elsevier B.V. All rights reserved. Source


Gao S.,University of Auckland | Gao S.,China National Institute of Standardization | Lewis G.D.,University of Auckland | Ashokkumar M.,University of Melbourne | And 2 more authors.
Ultrasonics Sonochemistry | Year: 2014

The aim of this study was to determine the effects of high-intensity low-frequency (20 kHz) ultrasound treatment on the viability of bacteria suspension. More specifically, we have investigated the relationship between the deactivation efficiency and the physical (size, hydrophobicity) and biological (gram-status, growth phase) properties of the microbes. Enterobacter aerogenes, Bacillus subtilis, Staphylococcus epidermidis, S. epidermidis SK and Staphylococcus pseudintermedius were chosen for this study owing to their varying physical and biological properties. The survival ratio of the bacteria suspension was measured as a function of the ultrasound power (up to 13 W) for a constant sonication time of 20 min. Transmission electron microscopy was used to evaluate the ultrasound-induced damages to the microbes. Ultrasound treatment resulted in lethal damage to E. aerogenes and B. subtilis (up to 4.5-log reduction), whereas Staphylococcus spp. were not affected noticeably. Further, E. aerogenes suspensions were more sensitive to ultrasonication in exponential growth phase than when they were in stationary phase. The results of this study demonstrate that the main reason for bacterial resistance to ultrasonic deactivation is due to the properties of the bacterial capsule. Microbes with a thicker and "soft" capsule are highly resistant to ultrasonic deactivation process. © 2013 Elsevier B.V. All rights reserved. Source


Gao S.,University of Auckland | Gao S.,China National Institute of Standardization | Hemar Y.,University of Auckland | Ashokkumar M.,University of Melbourne | And 2 more authors.
Water Research | Year: 2014

High-frequency (850kHz) ultrasound was used to inactivate bacteria and yeast at different growth phases under controlled temperature conditions. Three species of bacteria, Enterobacter aerogenes, Bacillus subtilis and Staphylococcus epidermidis as well as a yeast, Aureobasidium pullulans were considered. The study shows that high-frequency ultrasound is highly efficient in inactivating the bacteria in both their exponential and stationary growth phases, and inactivation rates of more than 99% were achieved. TEM observation suggests that the mechanism of bacteria inactivation is mainly due to acoustic cavitation generated free radicals and H2O2. The rod-shaped bacterium B. subtilis was also found to be sensitive to the mechanical effects of acoustic cavitation. The study showed that the inactivation process continued even after ultrasonic processing cessed due to the presence of H2O2, generated during acoustic cavitation. Compared to bacteria, the yeast A. pullulans was found to be more resistant to high-frequency ultrasound treatment. © 2014 Elsevier Ltd. Source


Wei L.,China National Institute of Standardization | Chen Z.,Northeastern University China | Li J.,CAS Institute of Policy and Management
Information Sciences | Year: 2011

Not only different databases but two classes of data within a database can also have different data structures. SVM and LS-SVM typically minimize the empirical φ-risk; regularized versions subject to fixed penalty (L 2 or L1 penalty) are non-adaptive since their penalty forms are pre-determined. They often perform well only for certain types of situations. For example, LS-SVM with L2 penalty is not preferred if the underlying model is sparse. This paper proposes an adaptive penalty learning procedure called evolution strategies (ES) based adaptive Lp least squares support vector machine (ES-based Lp LS-SVM) to address the above issue. By introducing multiple kernels, a Lp penalty based nonlinear objective function is derived. The iterative re-weighted minimal solver (IRMS) algorithm is used to solve the nonlinear function. Then evolution strategies (ES) is used to solve the multi-parameters optimization problem. Penalty parameterp, kernel and regularized parameters are adaptively selected by the proposed ES-based algorithm in the process of training the data, which makes it easier to achieve the optimal solution. Numerical experiments are conducted on two artificial data sets and six real world data sets. The experiment results show that the proposed procedure offer better generalization performance than the standard SVM, the LS-SVM and other improved algorithms. © 2011 Elsevier Inc. All rights reserved. Source

Discover hidden collaborations