Training classifiers on skewed data can be technically challenging tasks, especially if the is high-dimensional simultaneously, tasks become more difficult. In biomedicine field, type often appears. this study, we try to deal with problem by combining asymmetric bagging ensemble classifier (asBagging) that has been presented in previous work and an improved random subspace (RS) generation strategy called feature (FSS). Specifically, FSS a novel method promote balance level between accuracy...

DNA microarray technology can measure the activities of tens thousands genes simultaneously, which provides an efficient way to diagnose cancer at molecular level. Although this strategy has attracted significant research attention, most studies neglect important problem, namely, that datasets are skewed, causes traditional learning algorithms produce inaccurate results. Some have considered yet they merely focus on binary-class problem. In paper, we dealt with multiclass imbalanced...

DNA Microarray technology is an innovative methodology in experimental molecular biology, which has produced huge amounts of valuable data the profile gene expression. Many clustering algorithms have been proposed to analyze expression data, but little guidance available help choose among them. The evaluation feasible and applicable becoming important issue today's bioinformatics research. In this paper we first experimentally study three major algorithms: Hierarchical Clustering (HC),...

The echo-state-network approach for training recurrent neural networks can yield good results. However, the results depend on experience of network design. It usually requires multiple tests and random chances. Through our study effects spectral radius internal weight matrix results, we propose to develop a method that improve echo-state by introducing dynamic radius. Our experiments verify new algorithm is significantly better than original it stable.

The purpose of this paper is to demonstrate the design and implementation GISolve - a grid-based problem solving environment for computationally intensive geographic information analysis based on geo-middleware. geo-middleware resides between existing grid middleware applications manage heterogeneous dynamic resources behalf applications. At same time, provides adaptive domain decomposition solutions parallel Based these solutions, also schedules distributed tasks manages data transfers. In...

The iterative reconstruction (IR) algorithms are able to generate CT images with higher quality compared the filtered backprojection method when projection data is noisy or radiation dose low. IR can also be used incomplete. major disadvantage of its high demand on computation and slow reconstruction. To improve time performance we parallelized four representative algorithms: EM, SART their ordered subset (OS) versions a Linux PC cluster. A micro-forward-back-projector was implemented for...

Based on the concept depictions and systems analysis of cloud computing e-commerce features, we proposed a service-oriented Turban model with computing, which called "Cloud-based enterprise framework". The framework can be used to analyze feasibilities cloud-computing-based impacts SME in aspects technology, service, business model. Finally propose roadmap "adoption-implementation-evaluation-innovation" as development strategies for based computing.

After we briefly review representative analytic and iterative reconstruction algorithms for X-ray computed tomography (CT), address the need faster by parallel computing techniques. For a decent, conebeam usually takes hours on regular PC, since most of take more than 60 iterations even longer. In order to speedup performance, people introduce various acceleration methodologies including algorithm improvements, chip utilization, technique. This paper focuses computation using computing. The...

A protein structural class (PSC) belongs to the most basic but important classification in structures. The prediction technique of has been developing for decades. Two popular indices are amino-acid-frequency (AAF) based, and amino-acid-arrangement (AAA) with long-term correlation (LTC) - based indices. They were proposed many works. Both have its pros cons. For example, AAF index focuses on a statistical analysis, while AAA-LTC emphasizes long-term, biological significance. Unfortunately,...

In many real world applications, class imbalance problems occur frequently, causing great underestimation for the classification performance of minority classes. recent years, much effective solutions have been proposed to address this problem. Ho

Speckle is a multiplicative noise that degrades the visual evaluation in ultrasound imaging. In medical image processing, speckle suppression has become very essential exercise for diagnose. The recent advancements devices necessitate need of more robust despeckling techniques enhancing imaging routine clinical practice. Many denoising have been proposed effective noise. This paper compiles performance various B-mode images.

Resource discovery technologies for Grids and Peer-to-Peer (P2P) systems share some characteristics. However, P2P such as Skip Graph cannot simply be applied to Grid resource because complex grid resources need searched by using multi-attribute queries. This paper proposes a new approach (SkipMard) that extends structure support SkipMard provides prefix matching routing algorithm resolve queries, introduces the concepts of "layer" "crossing layer nearest neighbor" into data structure. To...

The Second Symposium on Computations in Bioinformatics and Bioscience (SCBB07) was held Iowa City, Iowa, USA, August 13–15, 2007. This annual event attracted dozens of bioinformatics professionals students, who are interested solving emerging computational problems bioscience, from China, Japan, Taiwan the United States. Scientific Committee symposium selected 18 peer-reviewed papers for publication this supplemental issue BMC Bioinformatics. These cover a broad spectrum topics biology...

In this paper, the melting of ice 1h is studied using molecular dynamics (MD). Common potential functions employed in MD simulations include SPC/E, TIP4P, TIP5P, TIP4/ice and TIP5P/E. We first conducted bulks then speed ice/water interface during melting. It found that various potentials result similar ice-melting phenomena. The compared with analytical solution for phase change problem. also size effects temperature gradient on

Accurately-predicting the readability of text documentation is important for educators, writers and learners. In perspective linguistics, many researchers study by analyzing semantics, vocabulary, syntax, expression, stylish, cultural. The considerations these facts are combined together to generate a common predictor. this paper, we first review status field with conventional methods being used assess evaluate readability. Our emphasis on feature selection, since features commonly effects...

Genetic association studies have been used to map disease-causing genes. A newly introduced statistical method, called exhaustive haplotype study, analyzes genetic information consisting of different numbers and combinations DNA sequence variations along a chromosome. Such involve large number calculations subsequently high computing power. It is possible develop parallel algorithms codes perform the on performance (HPC) system. However, most existing commonly-used statistic packages for are...

Due to the limited availability of energy within network nodes, energy-efficient routing mechanism is one most critical issues in wireless sensor networks. In this paper, we propose a multi-path algorithm with energy-aware based on ant colony system. The introduces three improved rules: Ant Marginalization Rule, State Transition Rule and Global Pheromone Update solve problems local convergence, optimization for transferring data respectively. It can avoid using up nodes optimal path while...

There exists a significant difference between traditional data processing and today's big technology, especially structure information extraction. Such brings higher education systems great challenge which promotes their changes in the concept methodology. This paper analyses architecture of data-enhanced with necessary evaluation components institutes.

Abstract A volume of fluid finite difference method based computer code was developed to simulate the heat transfer and squeezing flow in hot‐plate welding process. The simulation results include melt displacement, temperature distribution, stress build‐up relaxation. By implementing crystallization kinetic model Part I, formation stress‐induced crystal structure affected zone can be revealed.