The classification of technical equipment states, using a neural nets approach

The Maser thesis is devoted to developing a model to technical state of gas turbine engine estimation. The approaches to preparation data, especially to handle unbalanced data were presented in the thesis. In order to efficient estimation of model performance, the special metric was chosen. Goal of the master thesis is analyzing of monitoring parameters data and developing a model of technical state of GTE estimation based on the data.

Handling the Problem of Unbalanced Data Sets in the Classification of Technical Equipment States

Questions of handling unbalanced data considered in this article. As models for classification, PNN and MLP are used. Problem of estimation of model performance in case of unbalanced training set is solved. Several methods (clustering approach and boosting approach) considered as useful to deal with the problem of input data.

Petrographical classification and mineral assemblage of basement clasts from the Quaternary and Miocene sections of sediment core CRP-1 (Table 1)

Petrographical and mineral chemistry data are described for the mist representative basement lithologies occurring as clasts (pebble grain-size class) from the CRP-1 drillhole. Most pebbles consits of either undeformed or foliated biotite with or without hornblende monzogranites. Other rock types include biotite with or without garnet syenogranitr, biotite-hornblende granodiorite, tonalite, monzogranitic porphyries, haplogranite, quartz-monzonite (restricted to the Quaternary section), Ca-silicate rocks and biotite amphibolite (restricted to the Miocene strata). The common and ubiquitous occurence of biotite with or without hornblende monzogranite pebbles, in both the Quaternary and Miocene sections, apparently mirrors the dominance of these rock types in the granitoid assemblages which are presently exposed in the upper Precambrian-lower Paleozoic basement of the south Victoria Land. The other CRP-1 pebble lithologies show petrographical features which consitently support a dominant supply from areas of the Transantarctic Mountains located to the west and south-west of the CRP-1 site, and they thus furthercorroborate a model of local provenance for the supply of basement clasts to the CRP-1 sedimentary strata.

A Knowledge-based System for Intelligent Support in Pharmacogenomics Evidence Assessment: Ontology-driven Evidence Representation, Retrieval, Classification and Interpretation

Thesis (Ph.D.)--University of Washington, 2016-06

Examining Gender Differences in Writing Skill with Latent Factor Modeling

Thesis (Master's)--University of Washington, 2016-06

Constructing and evaluating alternative prediction models for identifying clinically relevant cases in an STD clinic

Thesis (Master's)--University of Washington, 2016-06

Phobic anxiety in 11 nations: Part I: Dimensional constancy of the five-factor model

The Fear Survey Schedule-III (FSS-III) was administered to a total of 5491 students in Australia, East Germany, Great Britain, Greece, Guatemala, Hungary, Italy, Japan, Spain, Sweden, and Venezuela, and submitted to the multiple group method of confirmatory analysis (MGM) in order to determine the cross-national dimensional constancy of the five-factor model of self-assessed fears originally established in Dutch, British, and Canadian samples. The model comprises fears of bodily injury-illness-death, agoraphobic fears, social fears, fears of sexual and aggressive scenes, and harmless animals fears. Close correspondence between the factors was demonstrated across national samples. In each country, the corresponding scales were internally consistent, were intercorrelated at magnitudes comparable to those yielded in the original samples, and yielded (in 93% of the total number of 55 comparisons) sex differences in line with the usual finding (higher scores for females). In each country, the relatively largest sex differences were obtained on harmless animals fears. The organization of self-assessed fears is sufficiently similar across nations to warrant the use of the same weight matrix (scoring key) for the FSS-III in the different countries and to make cross-national comparisons feasible. This opens the way to further studies that attempt to predict (on an a priori basis) cross-national variations in fear levels with dimensions of national cultures. (C) 2002 Elsevier Science Ltd. All rights reserved.

Reliability measures for local nodes assessment in classification trees

Most of the modem developments with classification trees are aimed at improving their predictive capacity. This article considers a curiously neglected aspect of classification trees, namely the reliability of predictions that come from a given classification tree. In the sense that a node of a tree represents a point in the predictor space in the limit, the aim of this article is the development of localized assessment of the reliability of prediction rules. A classification tree may be used either to provide a probability forecast, where for each node the membership probabilities for each class constitutes the prediction, or a true classification where each new observation is predictively assigned to a unique class. Correspondingly, two types of reliability measure will be derived-namely, prediction reliability and classification reliability. We use bootstrapping methods as the main tool to construct these measures. We also provide a suite of graphical displays by which they may be easily appreciated. In addition to providing some estimate of the reliability of specific forecasts of each type, these measures can also be used to guide future data collection to improve the effectiveness of the tree model. The motivating example we give has a binary response, namely the presence or absence of a species of Eucalypt, Eucalyptus cloeziana, at a given sampling location in response to a suite of environmental covariates, (although the methods are not restricted to binary response data).

On a resampling approach for tests on the number of clusters with mixture model-based clustering of tissue samples

We consider the problem of assessing the number of clusters in a limited number of tissue samples containing gene expressions for possibly several thousands of genes. It is proposed to use a normal mixture model-based approach to the clustering of the tissue samples. One advantage of this approach is that the question on the number of clusters in the data can be formulated in terms of a test on the smallest number of components in the mixture model compatible with the data. This test can be carried out on the basis of the likelihood ratio test statistic, using resampling to assess its null distribution. The effectiveness of this approach is demonstrated on simulated data and on some microarray datasets, as considered previously in the bioinformatics literature. (C) 2004 Elsevier Inc. All rights reserved.

Using the EM Algorithm to Train Neural Networks: Misconceptions and a New Algorithm for Multiclass Classification

The expectation-maximization (EM) algorithm has been of considerable interest in recent years as the basis for various algorithms in application areas of neural networks such as pattern recognition. However, there exists some misconceptions concerning its application to neural networks. In this paper, we clarify these misconceptions and consider how the EM algorithm can be adopted to train multilayer perceptron (MLP) and mixture of experts (ME) networks in applications to multiclass classification. We identify some situations where the application of the EM algorithm to train MLP networks may be of limited value and discuss some ways of handling the difficulties. For ME networks, it is reported in the literature that networks trained by the EM algorithm using iteratively reweighted least squares (IRLS) algorithm in the inner loop of the M-step, often performed poorly in multiclass classification. However, we found that the convergence of the IRLS algorithm is stable and that the log likelihood is monotonic increasing when a learning rate smaller than one is adopted. Also, we propose the use of an expectation-conditional maximization (ECM) algorithm to train ME networks. Its performance is demonstrated to be superior to the IRLS algorithm on some simulated and real data sets.

Sexual abuse, antisocial behaviour and substance use: gender differences in young community adolescents

Objective: To investigate gender-specific relationships between self-reported sexual abuse, antisocial behaviour and substance use in a large community sample of adolescents. Method: A cross-sectional study of students aged, on average, 13 (n = 2596), 14 (n = 2475) and 15 years (n = 2290), from 27 schools in South Australia with a questionnaire including sexual abuse, frequency and severity of substance use, depressive symptomatology (CES-D), family functioning (McMaster Family Assessment Device), and antisocial behaviour (an adapted 22-item Self-Report Delinquency Scale). Logistic regression analyses using HLM V5.05 with a population-average model were conducted. Results: In the model considered, reported sexual abuse is significantly independently associated with antisocial behaviour, controlling for confounding factors of depressive symptomatology and family dysfunction, with increased risks of three- to eightfold for sexually abused boys, and two- to threefold for sexually abused girls, compared to nonabused. Increased risks of extreme substance use in sexually abused girls (age 13) and boys (ages 13-15) are more than fourfold, compared to nonabused. Age differences were not statistically significant. Conclusion: Childhood sexual abuse is a risk factor for the development of antisocial behaviour and substance use in young adolescents. Clinicians should be aware of gender differences.

Classification and analysis of integrating frameworks in multiscale modelling

Chemical engineers are turning to multiscale modelling to extend traditional modelling approaches into new application areas and to achieve higher levels of detail and accuracy. There is, however, little advice available on the best strategy to use in constructing a multiscale model. This paper presents a starting point for the systematic analysis of multiscale models by defining several integrating frameworks for linking models at different scales. It briefly explores how the nature of the information flow between the models at the different scales is influenced by the choice of framework, and presents some restrictions on model-framework compatibility. The concepts are illustrated with reference to the modelling of a catalytic packed bed reactor. (C) 2004 Elsevier Ltd. All rights reserved.

Subordinate-manager gender combination and perceived leadership style influence on emotions, self-esteem and organizational commitment

A theoretical model was developed to investigate the relationships among subordinate-manager gender combinations, perceived leadership style, experienced frustration and optimism, organization-based self-esteem and organizational commitment. The model was tested within the context of a probabilistic structural model, a discrete Bayesian network, using cross-sectional data from a global pharmaceutical company. The Bayesian network allowed forward inference to assess the relative influence of gender combination and leadership style on the emotions, self-esteem and commitment consequence variables. Further, diagnostics from backward inference were used to assess the relative influence of variables antecedent to organizational commitment. The results showed that gender combination was independent of leadership style and had a direct impact on subordinates' levels of frustration and optimism. Female manager-female subordinate had the largest probability of optimism, while male manager teamed with a male subordinate had the largest probability of frustration. Furthermore, having a female manager teamed up with a male subordinate resulted in the lowest possibility of frustration. However, the findings show that the gender issue is not simply female managers versus male managers, but is concerned with the interaction of the subordinate-manager gender combination and leadership style in a nonlinear manner. (C) 2003 Elsevier Inc. All rights reserved.

An economic analysis of life expectancy by gender with application to the United States

This paper presents an economic model to explain the behavior of life expectancy of both sexes. It explicitly examines the relationship between the gender gap in life expectancy and the gender gap in pay. It shows that as the latter narrows over the course of economic development, the former may initially expand but will eventually shrink. Simulation results from our model accord with the behavior of life expectancy for both sexes since the 1940s in the United States. (C) 2003 Elsevier B.V. All rights reserved.