Abstract. Classifier fusion strategies have shown great potential to enhance the performance of pattern recognition systems. There is an agreement among researchers in classifier combination that the major factor for producing better accuracy is the diversity in the classifier team. Re-sampling based approaches like bagging, boosting and random subspace generate multiple models by training a single learning algorithm on multiple random replicates or sub-samples, in either feature space or the sample domain. In the present study we proposed a hybrid random subspace fusion scheme that simultaneously utilizes both the feature space and the sample domain to improve the diversity of the classifier ensemble. Experimental results using two protein mass spectra datasets of ovarian cancer demonstrate the usefulness of this approach for six learning algorithms (LDA, 1-NN, Decision Tree, Logistic Regression, Linear SVMs and MLP). The results also show that the proposed strategy outperforms three conventional re-sampling based ensemble algorithms on these datasets.
Background: Certain murine leukemia viruses (MLVs) are capable of inducing progressive spongiform motor neuron disease in susceptible mice upon infection of the central nervous system (CNS). The major CNS parenchymal target of these neurovirulent retroviruses (NVs) are the microglia, whose infection is largely coincident with neuropathological changes. Despite this close association, the role of microglial infection in disease induction is still unknown. In this paper, we investigate the interaction of the highly virulent MLV, FrCasE, with microglia ex vivo to evaluate whether infection induces specific changes that could account for neurodegeneration. Specifically, we compared microglia infected with FrCasE, a related non-neurovirulent virus (NN) F43/Fr57E, or mock-infected, both at a basic virological level, and at the level of cellular gene expression using quantitative real time RT-PCR (qRT-PCR) and Afffymetrix 430A mouse gene chips.
Methods for protein secondary structure prediction have improved significantly in recent years. This has lead to enhanced protein homology modeling efforts. Protein homology modeling involves the sub-task of identifying a set of homologous proteins from a protein database when given as input the amino acid sequence of a qnery protein, with the ultimate goal of using the resulting set of homologous proteins as a starting point for predicting the 3D structure of the qnery protein. Previous work has indicated that improvements can be made when combining secondary structure sequence alignment using a 3-state structure symbol alphabet together with primary amino acid sequence alignment methods. These approaches typically use a local alignment algorithm. We compare the performance of several dynamic programming alignment algorithms on the task of aligning secondary structure sequences using an %state secondary structure alphabet. Our results indicate that the typical use of a local alignment algorithm may not be best when aligning protein secondary structure information.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.