Building strength models for high-performance concrete at different ages using genetic operation trees, nonlinear regression, and neural networks

Because the behavior of HPC at early age may be rather different at late age, this study proposed to establish the strength models of HPC at different ages, and to explore the difference between these models. A large number of experimental data were used to compare accuracy of the three model building techniques, nonlinear regression analysis (NLRA), back-propagation networks (BPN), and genetic operation trees (GOT). The results showed: (1) when NLRA was employed to establish the prediction model, the approach to establish HPC strength models based on the three separate data sets was more accurate than that used to establish HPC strength models for the total data set. (2) If an explicit formula is necessary, GOT is the best choice to build concrete strength models at medium and late ages (i.e., more than 14 days), while NLRA provides greater accuracy at early ages (i.e., less than 14 days); otherwise, BPN is the best choice.     

Cascade of genetic algorithm and decision tree for cancer classification on gene expression data

Abstract: Cancer classification, through gene expression data analysis, has produced remarkable results, and has indicated that gene expression assays could significantly aid in the development of efficient cancer diagnosis and classification platforms. However, cancer classification, based on DNA array data, remains a difficult problem. The main challenge is the overwhelming number of genes relative to the number of training samples, which implies that there are a large number of irrelevant genes to be dealt with. Another challenge is from the presence of noise inherent in the data set. It makes accurate classification of data more difficult when the sample size is small. We apply genetic algorithms (GAs) with an initial solution provided by t statistics, called t‐GA, for selecting a group of relevant genes from cancer microarray data. The decision‐tree‐based cancer classifier is built on the basis of these selected genes. The performance of this approach is evaluated by comparing it to other gene selection methods using publicly available gene expression data sets. Experimental results indicate that t‐GA has the best performance among the different gene selection methods. The Z‐score figure also shows that some genes are consistently preferentially chosen by t‐GA in each data set.     

An econometric analysis of SARS and Avian Flu on international tourist arrivals to Asia

This paper compares the impacts of SARS and human deaths arising from Avian Flu on international tourist arrivals to Asia. The effects of SARS and human deaths from Avian Flu are compared directly according to the number of human deaths. The nature of the short run and long run relationship is examined empirically by estimating a static line fixed effect model and a difference transformation dynamic model, respectively. Empirical results from the static fixed effect and difference transformation dynamic models are consistent, and indicate that both the short run and long run SARS effect have a more significant impact on international tourist arrivals than does Avian Flu. In addition, the effects of deaths arising from both SARS and Avian Flu suggest that SARS is more important to international tourist arrivals than is Avian Flu. Thus, while Avian Flu is here to stay, its effect is currently not as significant as that of SARS.     

Gene selection and sample classification on microarray data based on adaptive genetic algorithm/k-nearest neighbor method

Recently, microarray technology has widely used on the study of gene expression in cancer diagnosis. The main distinguishing feature of microarray technology is that can measure thousands of genes at the same time. In the past, researchers always used parametric statistical methods to find the significant genes. However, microarray data often cannot obey some of the assumptions of parametric statistical methods, or type I error may be over expanded. Therefore, our aim is to establish a gene selection method without assumption restriction to reduce the dimension of the data set. In our study, adaptive genetic algorithm/k-nearest neighbor (AGA/KNN) was used to evolve gene subsets. We find that AGA/KNN can reduce the dimension of the data set, and all test samples can be classified correctly. In addition, the accuracy of AGA/KNN is higher than that of GA/KNN, and it only takes half the CPU time of GA/KNN. After using the proposed method, biologists can identify the relevant genes efficiently from the sub-gene set and classify the test samples correctly.     

A multiple-kernel support vector regression approach for stock market price forecasting

Support vector regression has been applied to stock market forecasting problems. However, it is usually needed to tune manually the hyperparameters of the kernel functions. Multiple-kernel learning was developed to deal with this problem, by which the kernel matrix weights and Lagrange multipliers can be simultaneously derived through semidefinite programming. However, the amount of time and space required is very demanding. We develop a two-stage multiple-kernel learning algorithm by incorporating sequential minimal optimization and the gradient projection method. By this algorithm, advantages from different hyperparameter settings can be combined and overall system performance can be improved. Besides, the user need not specify the hyperparameter settings in advance, and trial-and-error for determining appropriate hyperparameter settings can then be avoided. Experimental results, obtained by running on datasets taken from Taiwan Capitalization Weighted Stock Index, show that our method performs better than other methods.     

A study of maintenance policies for second-hand products

This paper proposes two maintenance schemes for reused products. In order to reduce the purchasing cost or conform to the concept of environmental protection, reuse is considered as one of the most efficient strategies. However, the initial quality of reused product is often inferior to the new one and then product failures will occur frequently during usage period. Therefore, two periodical preventive maintenance (PM) policies in this paper are developed to decrease the high failure rate of the second-hand products. When a second-hand product with Weibull life time distribution of known age is intended to be used for a pre-specified period of time, the optimal number of PM actions and the corresponding maintenance degree are derived such that the expected total maintenance cost is minimized. The impact of providing preventive maintenance is illustrated through numerical examples.     

Modeling analysis of a triaxial microaccelerometer with piezoelectric thin-film sensing using energy method

This study applies energy method to derive the system modeling of a triaxial microaccelerometer that consists of a quadri-beam suspension, a seismic mass, and displacement transducers using piezoelectric thin films. Two suspension beams support both ends of the seismic mass, which is fabricated by anisotropic etching of silicon. An out-of-plane acceleration will result in a symmetric bend, and in-plane accelerations will produce asymmetric bend and torsion of the suspension beams. Two piezoelectric thin-film transducers are arranged at both ends of each suspension beam. Eight transducers in total are interconnected such that triaxial accelerations can be measured selectively. The structure stiffness of the suspension beams considers both the silicon beams and piezoelectric films by the use of the laminated beam theory. Therefore, the analytical model is applicable to the accelerometers with thick piezoelectric films. The model is based on the anisotropic material properties of Silicon and PZT and Euler’s beam equation with the assumptions that smaller strains and stresses are negligible. The analytical results of the resonant frequencies and sensor sensitivities to triaxial accelerations are presented and confirmed by finite element analysis.     

The adaptive fuzzy time series model with an application to Taiwan’s tourism demand

In this study, an adaptive fuzzy time series model for forecasting Taiwan’s tourism demand is proposed to further enhance the predicted accuracy. We first transfer fuzzy time series data to the fuzzy logic group, assign weights to each period, and then use the proposed adaptive fuzzy time series model for forecasting in which an enrollment forecasting values is applied to obtain the smallest forecasting error. Finally, an illustrated example for forecasting Taiwan’s tourism demand is used to verify the effectiveness of proposed model and confirmed the potential benefits of the proposed approach with a very small forecasting error MAPE and RMSE.     

Genetic algorithm-based heuristics for an open shop scheduling problem with setup,processing, and removal times separated

This report proposes a solution to the open shop scheduling problem with the objective of minimizing total job tardiness in the system. Some practical processing restrictions, such as independent setup and dependent removal times, are taken into account as well. The addressed problem is first described as a 0–1 integer programming model, and is then solved optimally. Subsequently, some hybrid genetic-based heuristics are proposed to solve the problem in an acceptable computation time. To demonstrate the adaptability of these heuristics, some performance comparisons are made with solutions provided by running either a mathematical programming model or certain classic meta-heuristics such as genetic algorithm, simulated annealing, and tabu search in various manufacturing scenarios. The experimental results show that the hybrid genetic-based heuristics perform well, especially the DGA. However, these heuristics require some more additional computations but are still acceptable.     

An expert system to classify microarray gene expression data using gene selection by decision tree

Gene selection can help the analysis of microarray gene expression data. However, it is very difficult to obtain a satisfactory classification result by machine learning techniques because of both the curse-of-dimensionality problem and the over-fitting problem. That is, the dimensions of the features are too large but the samples are too few. In this study, we designed an approach that attempts to avoid these two problems and then used it to select a small set of significant biomarker genes for diagnosis. Finally, we attempted to use these markers for the classification of cancer. This approach was tested the approach on a number of microarray datasets in order to demonstrate that it performs well and is both useful and reliable.