Abstract Existing classification analysis techniques (e.g., decision tree induction, backpropagation neural network, k-nearest neighbor classification, etc.) generally exhibit satisfactory classification effectiveness when dealing with data with non-skewed class distribution. However, real-world applications (e.g., churn prediction and fraud detection) often involve highly skewed data in decision outcomes (e.g., 2% churners and 98% non-churners). Such a highly skewed class distribution problem, if not properly addressed, would imperil the resulting learning effectiveness and might result in a ¡§null¡¨ prediction system that simply predicts all instances as having the majority decision class as the training instances (e.g., predicting all customers as non-churners). In this study, we extended the multi-classifier class-combiner approach and proposed a clustering-based multi-classifier class-combiner technique to address the highly skewed class distribution problem in classification analysis. In addition, we proposed four distance-based methods for selecting a subset of instances having the majority decision class for lowering the degree of skewness in a data set. Using two real-world datasets (including mortality prediction for burn patients and customer loyalty prediction), empirical results suggested that the proposed clustering-based multi-classifier class-combiner technique generally outperformed the traditional multi-classifier class-combiner approach and the four distance-based methods. Keywords: Data Mining, Classification Analysis, Skewed Class Distribution Problem, Decision Tree Induction, Multi-classifier Class-combiner Approach, Clustering-based Multi-classifier Class-combiner Approach

Characteristics of impinging-jet atomizers in the context of application in liquid propulsion systems are studied in this thesis. A review of past studies on impinging jets revealed the necessity of a correlation in terms of injector parameters for predicting Sauter Mean Diameter (SMD) of a spray. So, an experimental study of atomization in doublet and triplet impinging jet injectors is conducted using water as the stimulant? The major injector parameters considered are orifice diameter, impingement angle and jet velocity. Relative influences of these parameters are explained in terms of a single parameter, specific normal momentum. SMD of the spray reduces as specific normal momentum is increased. A universal expression between non-dimensional SMD and specific normal momentum is obtained, which satisfactorily predicts SMD in doublets as well as triplets. Noting that practical impinging injectors are likely to have skewness (partial impingement), the study is extended to understand the behavior of such jets. In perfectly impinging doublet, a high aspect ratio ellipse-like mass distribution pattern is obtained with major axis normal to the plane of two jets whereas in skewed jets the major axis turns from its normal position. A simple correlation is obtained, which shows that this angle of turn is a function of skewness fraction and impingement angle only and is independent of injection velocity. Experimental data from both mass distribution and photographic technique validate this prediction. SMD is found to decrease as skewness is increased. This may be the combined effect of shearing of liquid sheet at the point of impingement and more sheet elongation. Hence, skewness turns out to be an important parameter in controlling drop size.

銀行在評分模型建置完成後需進行驗證工作，以瞭解評分模型是否能有效評出客戶的風險層級，穩健地估計區別鑑別力指標為驗證工作中的重點。在先前的文獻中假設正常授信戶與違約戶分數分配為常態分配。但在實際資料中，分配未必定為常態。因此本文接著探討在正常授信戶與違約授信戶之分配為混合分配，即兩分數分配為偏斜常態分配下，何種方法可以對於估計AUC具有較高的穩定性。本文比較五種估計AUC的方法，分別為常態核，經驗分配，曼惠尼近似，最大摡似法和EM演算法。模擬結果呈現(1)投信戶組合分配為兩常態分配下，最大摡似法在大部分違約率下都可以得到較窄的信賴區間。(2)組合分配為一常態與一偏斜常態及兩偏斜常態分配下，EM演算法在大部分情況有較窄的信賴區間，其中在兩偏斜常態分配下，表現更佳。(3)曼惠尼近似建構的信賴區間寬度最大，代表曼惠尼近似是較保守的估計方法。 / Banks face discrimination after constructing the rating systems to figure out whether the systems can discriminate defaulting and non-defaulting borrowers. Literature assumed the two score distribuion are normal distributed. However, the real data may not be normal distribuions. We assum the two score distribuions are skewed normal distribuions to discuss which method has more robustness to estimate the AUC value.Under skewed distribution, we propose EM algorithm to estimate the population parametric. If used properly, information about the population properties may be used to get better accuracy of estimation the AUC value.Numerical results show the EM algorithm method , comparing with other methods, has robustness in detect the rating systems have discirmatory power.

Submitted by RENATO MORI (rmori3@hotmail.com) on 2017-09-20T05:58:01Z No. of bitstreams: 1 dissertacao_VaRArmaGarchSkewt.pdf: 3267680 bytes, checksum: 6a8a935c128bb04a8a4f91fb592de3a8 (MD5) / Approved for entry into archive by Thais Oliveira (thais.oliveira@fgv.br) on 2017-09-20T17:58:58Z (GMT) No. of bitstreams: 1 dissertacao_VaRArmaGarchSkewt.pdf: 3267680 bytes, checksum: 6a8a935c128bb04a8a4f91fb592de3a8 (MD5) / Made available in DSpace on 2017-09-21T13:36:32Z (GMT). No. of bitstreams: 1 dissertacao_VaRArmaGarchSkewt.pdf: 3267680 bytes, checksum: 6a8a935c128bb04a8a4f91fb592de3a8 (MD5) Previous issue date: 2017-08-22 / A proposta do estudo é aplicar ao Ibovespa, modelo paramétrico de VaR de 1 dia, com distribuição dos retornos dinâmica, que procura apreciar características empíricas comumente apresentadas por séries financeiras, como clusters de volatilidade e leptocurtose. O processo de retornos é modelado como um ARMA com erros GARCH que seguem distribuição t assimétrica. A metodologia foi comparada com o RiskMetrics e com modelos ARMA-GARCH com distribuição dos erros normal e t. Os modelos foram estimados diariamente usando uma janela móvel de 1008 dias. Foi verificado pelos backtests de Christoffersen e de Diebold, Gunther e Tay que dentre os modelos testados, o ARMA(2,2)- GARCH(2,1) com distribuição t assimétrica apresentou os melhores resultados. / The proposal of the study is to apply to Ibovespa a 1 day VaR parametric model, with dynamic distribution of returns, that aims to address empirical features usually seen in financial series, such as volatility clustering and leptocurtosis. The returns process is modeled as an ARMA with GARCH residuals that follow a skewed t distribution. The methodology was compared to RiskMetrics and to ARMA-GARCH with normal and t distributed residuals. The models were estimated every daily period using a window of 1008 days. By the backtests of Christoffersen and Diebold, Gunther and Tay, among the tested models, the ARMA(2,2)-GARCH(2,1) with skewed t distribution has given the best results.

With the continuous expansion of data availability in many large-scale, complex, and networked systems, such as surveillance, security, the Internet, and finance, it becomes critical to advance the fundamental understanding of knowledge discovery and analysis from raw data to support decision-making processes. Existing knowledge discovery and data analyzing techniques have shown great success in many real-world applications such as applying Automatic Target Recognition (ATR) methods to detect targets of interest in imagery, drug activity prediction, computer vision recognition, and so on. Among these techniques, Multiple-Instance (MI) learning is different from standard classification since it uses a set of bags containing many instances as input. The instances in each bag are not labeled | instead the bags themselves are labeled. In this area many researchers have accomplished a lot of work and made a lot of progress. However, there still exist some areas which are not covered. In this thesis, we focus on two topics of MI learning: (1) Investigating the relationship between MI learning and other multiple pattern learning methods, which include multi-view learning, data fusion method and multi-kernel SVM. (2) Dealing with the class imbalance problem of MI learning. In the first topic, three different learning frameworks will be presented for general MI learning. The first uses multiple view approaches to deal with MI problem, the second is a data fusion framework, and the third framework, which is an extension of the first framework, uses multiple-kernel SVM. Experimental results show that the approaches presented work well on solving MI problem. The second topic is concerned with the imbalanced MI problem. Here we investigate the performance of learning algorithms in the presence of underrepresented data and severe class distribution skews. For this problem, we propose three solution frameworks: a data re-sampling framework, a cost-sensitive boosting framework and an adaptive instance-weighted boosting SVM (with the name IB_SVM) for MI learning. Experimental results - on both benchmark datasets and application datasets - show that the proposed frameworks are proved to be effective solutions for the imbalanced problem of MI learning.

This master thesis deals with calculations and simulations of multi-speed induction motors. In the first part, basic principle of these machines is described. Next, there are given three examples of pole-changing stator windings and their properties are analyzed (winding factors, magnetomotive force). Main part includes simulations and measuring of the specific motor, which is produced by company Siemens Mohelnice. This is the motor with 2/1 pole-changing in Y/YY connection. First, influence of skewed rotor slots on current and torque is analyzed. Next, parametres of motor from simulations are given and these results are compared to measured values.

This thesis has used the Fama and French five-factor model (FF5M) and proposed an alternative model. The proposed model is named the Fama and French five-factor heteroscedastic student's model (FF5HSM). The model utilises an ARMA model for the returns with the FF5M factors incorporated and a GARCH(1,1) model for the volatility. The FF5HSM uses returns data from the FF5M's portfolio construction for the Japanese stock market and the five risk factors. The portfolio's capture different levels of market capitalisation, and the factors capture market risk. The ARMA modelling is used to address the autocorrelation present in the data. To deal with the heteroscedasticity in daily returns of stocks, a GARCH(1,1) model has been used. The order of the GARCH-model has been concluded to be reasonable in academic literature for this type of data. Another finding in earlier research is that asset returns do not follow the assumption of normality that a regular regression model assumes. Therefore, the skewed student's t-distribution has been assumed for the error terms. The result of the data indicates that the FF5HSM has a better in-sample fit than the FF5M. The FF5HSM addresses heteroscedasticity and autocorrelation in the data and minimises them depending on the portfolio. Regardingforecasting, both the FF5HSM and the FF5M are accurate models depending on what portfolio the model is applied on.

This master thesis presents a comparative analysis of the starting performance and losses at rated operation for a 15 kW, 4-pole industrial induction motor, mounted with standard skewed, unskewed and unskewed asymmetrical die-cast aluminium rotors through measurements and simulations. It is a well-known fact that rotor skewing suppresses the synchronous torques at low speeds and also reduces the audible noise of the machine. However, the casting process results in a low resistive path between the rotor bars and the iron laminations, for skewed rotors, this promotes the flow of inter-bar currents. These currents, flowing between the rotorbars, increase the harmonic torques during a start and create additional losses at rated operation. For standard unskewed rotors, these losses are ideally zero, but these rotors may produce high audible noise. Studies have shown that rotors with asymmetrical rotor slot pitch can reduce the audible noise level in unskewed machines. By removing the skew, the inter-bar current losses are suppressed to a negligible level; ultimately increased machine efficiency is obtained. In this work the electrical performance is verified through measurements on the built prototypes. Direct-on-line starts and rated performance for motors with different rotor slot arrangements is simulated using 2D FEM tool FCSmek. The three prototypes are tested in the laboratory according to IEC 60034-2-1 standard and the simulation results are in good agreement with the measured results. An additional test for the measurement of high frequency delta connected stator winding currents for each prototype machine is also performed, in order to study the losses induced in the stator winding. Results have shown that by introducing the proposed asymmetry in the rotor slots, the synchronous torques at low speeds are suppressed effectively, thus, improving the starting performance of the asymmetrical rotor compared to the standard unskewed rotor. Additionally, a higher pull-out torque is obtained for the unskewed rotor motor compared to the standard skewed rotor motor. However, the losses were more or less re-distributed in the unskewed rotor motor, resulting in similar efficiency as the standard skewed rotor motor. One important observation is that; to capture the inter-bar current losses which are estimated to be 5.5% of the total losses, requires more accurate methods of measurements than the existing. And sufficient repeatability must be achieved; alternatively one should rely on statistical data obtained from measurements on several number of motors.

Population control programs such as family planning and the introduction of sex identification technologies has helped to create skewed sex ratios in northern India and particularly in the state of Haryana. Due to a surplus of men and the numbers of missing females, an organized business of bride trafficking has emerged where poor women from eastern and northeastern states of India are bought and brought to Haryana for the purpose of marriage. This thesis explores how skewed sex ratios have contributed to the phenomenon of bride trafficking in Haryana guided by the theoretical framework of violences of development which argues that there is a hidden paradox within development. This was done by conducting a conventional content analysis in order to create a deeper understanding of the phenomenon. There is scarce research on bride trafficking, therefore this study contributes with extended knowledge in order to shed a light on the increasing trade with females.

This thesis investigates the strength behavior of a representative highly skewed and horizontally curved bridge as well as analysis and design procedures for these types of structures. The bridge responses at and above a number of limits in the AASHTO (2007) Specifications are considered. The study includes the evaluation of various attributes of the elastic analysis of the subject bridge. These attributes include: (1) the accuracy of 3-D grid versus 3-D FEA models, (2) first-order versus second-order effects during the construction, (3) the ability to predict layover at bearing lines using simplified equations and (4) the benefit of combining the maximum and concurrent major-axis and flange lateral bending values due to live load compared to combining the maximums due to different live loads when checking the section resistances. The study also addresses the ability of different AASHTO 2007 resistance equations to capture the ultimate strength behavior. This is accomplished by comparing the results from full nonlinear 3-D FEA studies to the elastic design and analysis results. Specifically the use of the 2007 AASHTO moment based one-third rule equations is evaluated for composite sections in positive bending.