Clustering, classification, and factor analysis are three popular data mining techniques. In this dissertation, we investigate these methods in high dimensional data analysis. Since there are much more features than the sample sizes and most of the features are non-informative in high dimensional data, dimension reduction is necessary before clustering or classification can be made. In the first part of this dissertation, we reinvestigate an existing clustering procedure, optimal discriminant clustering (ODC; Zhang and Dai, 2009), and propose to use cross-validation to select the tuning parameter. Then we develop a variation of ODC, sparse optimal discriminant clustering (SODC) for high dimensional data, by adding a group-lasso type of penalty to ODC. We also demonstrate that both ODC and SDOC can be used as a dimension reduction tool for data visualization in cluster analysis. In the second part, three existing sparse principal component analysis (SPCA) methods, Lasso-PCA (L-PCA), Alternative Lasso PCA (AL-PCA), and sparse principal component analysis by choice of norm (SPCABP) are applied to a real data set the International HapMap Project for AIM selection to genome-wide SNP data, the classification accuracy is compared for them and it is demonstrated that SPCABP outperforms the other two SPCA methods. Third, we propose a novel method called sparse factor analysis by projection (SFABP) based on SPCABP, and propose to use cross-validation method for the selection of the tuning parameter and the number of factors. Our simulation studies show that SFABP has better performance than the unpenalyzed factor analysis when they are applied to classification problems.

本文的主要目的，是找出一個簡單且有效的方法，預測台灣的股市報酬。比較許多不同的研究後，我發現無論面對多重共線性亦或變動要素結構等問題，主要要素模型(Principal Components Analysis)都可以表現地比其他模型優異。因此，在此篇文章中，我結合了資產訂價理論(Asset Pricing Theory)與主要要素模型的概念，來預測台灣八大產業股票指數的報酬。分析結果顯示，雖然主要要素模型在本文中的預測表現不如預期，但是整體仍優於隨機漫步(Random Walk)的預測。這意味著，主要要素模型對台灣股市的預測，可以在某種程度上推翻效率市場假說(Efficient Market Hypothesis)。 / The original purpose of this paper is to find a useful and simple way to forecast the return rates of Taiwan stock market. Comparing different empirical studies, I found that no matter with problems of multicollinearity or changing factor structure, the Principal Components Analysis (PCA) can usually outperform other models. Therefore, I combined the concepts of Asset Pricing Theory (APT) and PCA, to predict the movements of eight industrial indexes return rates of Taiwan stock market. The analysis indicates that, although PCA forecasting results couldn’t be very impressive in Taiwan stock market, it still can perform better than Random Walk Regression. That means the forecasting results of PCA to Taiwan stock market can overthrow the Efficient Market Hypothesis (EMH), which represents the trends of stock return rates are unpredictable, to some extents.

Thesis (Ph. D.)--Ohio State University, 2007. / Title from first page of PDF file. Includes bibliographical references (p. 204-214).

International bond market linkages are examined using monthly bond yield data and total return indices on government bonds with ten years to maturity. The bond yield data covers a nineteen-year period from January 1990 to July 2008, while the bond total return index data covers a nine-year period from August 2000 to July 2008. The international bond markets included in the study are Australia, Canada, Germany, Japan, the United Kingdom, and the United States. The examination of international bond market linkages across these markets has important implications for the formulation of effective portfolio diversification strategies. The empirical analysis is carried out in three phases: the preliminary analysis, the principal component analysis (PCA), and the cointegration analysis. For each analysis and for each set of data the full sample period is first analysed and subsequently a five-year rolling window approach is implemented. Accordingly, this makes it possible to capture the time-varying nature of international bond market linkages. The preliminary analysis examines the bond market trends over the sample period, provides descriptive statistics, and reports the correlation coefficients between the selected bond markets. The PCA investigates the interrelationships among the bond markets according to their common sources of movement and identifies which markets tend to move together. The cointegration analysis is carried out using the Johansen cointegration procedure and investigates whether there is long-run comovement between South Africa and the selected bond markets. Where cointegration is found, Vector Error-Correction Models (VECMs) are estimated in order to examine the long-run equilibrium relationships in addition to their short-run adjustments over time. The empirical analysis results were robust, and overall integration between SA and the selected major bond markets remained weak and sporadic. In addition, the results showed that even after accounting for exchange rate differentials, international bond market diversification remained beneficial for a South African investor; and since international bond market linkages remained weak with no observable trend, international bond market diversification will remain beneficial for some time to come for a South African investor.

A textura e a cor têm influência na aquisição, consumo, aceitação e preferência de biscoitos. Alguns ingredientes e etapas de processo podem influenciar diretamente nestes parâmetros. O açúcar é um dos principais ingredientes utilizados nas formulações de biscoitos, sendo o tamanho, ou diâmetro dos cristais um fator importante para o comportamento da massa e consequentemente para a textura e cor do produto. Tendo em vista que a movimentação do açúcar na indústria de alimentos é realizada geralmente por transporte pneumático,e os cristais são quebrados, alterando a granulometria e consequentemente interferindo nas características do produto. Neste contexto, o objetivo deste trabalho foi avaliar a influência do transporte pneumático na granulometria e cor de açúcar cristal e consequentes modificações no comportamento da massa e nos parâmetros textura e cor de biscoitos rosca sabor leite. Para tal, foram realizadas análises de granulometria e cor de açúcar cristal, bem como análises da massa e dos biscoitos elaborados. O açúcar foi obtido de três diferentes fornecedores, codificados como A, B e C, e foi aplicado na produção das massas e dos biscoitos roscas sabor leite em duas condições: antes do transporte pneumático (ATP) e depois de submetido ao transporte pneumático (DTP), totalizando seis produções diferentes de massa e de biscoitos. Todos os demais ingredientes e condições de processo foram mantidas sem alteração em todas as produções. Através da análise de granulometria foi determinado o valor de diâmetro médio dos cristais de açúcar e a avaliação colorimétrica possibilitou determinar os valores dos intervalos de cores L, a* e b* (CIELab) do açúcar. O comportamento reológico da massa de biscoitos foi avaliado através de análise de consistência, estabilidade e dureza. Os biscoitos rosca sabor leite foram submetidos a análises instrumentais e sensoriais (teste descritivo e afetivo), onde foram avaliadas a textura (dureza e fraturabilidade) e cor (L, a* e b*). A granulometria nas amostras de açúcar ATP foi significativamente maior (p<0,05) que nas amostras de DTP. Por outro lado, a análise de cor nas amostras de açúcar ATP apresentou luminosidade menor do que nas DTP (p<0,05). A dureza da massa foi maior (p<0,05) nas amostras onde foi aplicado o açúcar DTP. A textura do biscoito (dureza e a fraturabilidade) com açúcar ATP foram significativamente menores que os produzidos com açúcar DTP. A cor do produto com açúcar DTP foi maior (p<0,05) que no produto com açúcar ATP. A ACP mostrou correlação entre os dados instrumentais e sensoriais. De acordo com os resultados obtidos observou-se que o transporte pneumático influencia diretamente na granulometria e cor do açúcar, bem como na textura e cor do produto final. Sendo assim, conclui-se que a utilização de açúcar cristal submetido ao transporte pneumático (DTP) na produção de biscoitos rosca sabor leite resulta em produtos mais escuros, e com maior dureza. / The texture and color influence the acquisition, consumption, acceptance and preference biscuits. Some ingredients and stages of process can directly influence these parameters. Sugar is one of the main ingredients used in the formulations biscuits, and the crystal size an important factor in the mass performance and consequently to the texture and color of the product. Having in view that the movement of the sugar in the food industry is carried out usually by pneumatic transport, and the crystals are broken by changing the particle size and consequently interfering with the characteristics of the product. In this context, the objective of this study was to evaluate the influence of pneumatic conveying in particle size and color of crystal sugar and consequent changes in mass behavior and parameters texture and color of thread biscuits flavored milk. For such, particle size and color of crystal sugar analyzes were conducted, as well as the dough analysis and biscuits prepared. The sugar was obtained from three different suppliers, coded as A, B and C, and has been applied in production of dough and biscuits flavored milk thread, in two conditions: before the pneumatic transport (BPT) and after being subjected to pneumatic transport (APT), totaling six different productions of dough and biscuits. All the other ingredients and process conditions were maintained without change in all productions. Through the analysis of particle size was determined the value of mean aperture of sugar crystals and the colorimetric evaluation allowed us to determine the values of color ranges L, a * and b * (Cielab) of sugar. The rheological behavior of cookie dough was evaluated using analysis of consistency stability and hardness. The biscuits thread flavor milk have been subjected to instrumental and sensory analysis (descriptive and affective tests), which were evaluated texture (hardness and fracturability) and color (L, a * and b *). The particle size distribution in the samples of sugar BPT was significantly higher (p< 0.05) than in samples of APT. On the other hand, the analysis of color in the samples of sugar BPT showed brightness lower than in APT (p< 0.05).The hardness of the dough was higher (p <0.05) in samples where the Sugar APT was applied. The biscuit texture (hardness and fracturability) sugar BPT were significantly lower than those produced with APT sugar. The product color APT sugar was higher (p <0.05) in the product with sugar ATP. The PCA showed correlation between instrumental and sensory analysis. According to the results obtained, it was observed that the pneumatic transport directly influences the size and color of sugar, as well as the texture and color of the final product. Thus, it is concluded that the use of crystal sugar submitted to pneumatic transport (APT) in the production of biscuits milk flavor results in products darker, and with greater hardness.

Orientador: Robson Pederiva / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-26T01:39:31Z (GMT). No. of bitstreams: 1 Varanis_MarcusViniciusMonteiro_D.pdf: 5116959 bytes, checksum: b16ac36565b93c6bf49eb1863f7e9823 (MD5) Previous issue date: 2014 / Resumo: Motores elétricos são componentes de grande importância na maioria dos equipamentos de plantas industriais. As diversas falhas que ocorrem nas máquinas de indução podem gerar consequências severas no processo industrial. Os principais problemas estão relacionados à elevação dos custos de produção, piora nas condições do processo e de segurança e, sobretudo piora na qualidade do produto final. Muitas destas falhas mostram-se progressivas. Neste trabalho, apresenta-se uma contribuição ao estudo de Técnicas de Processamento de Sinais Baseadas na Transformada Wavelet para extração de parâmetros de Energia e Entropia a partir de sinais de vibração para detecção de falhas no regime não-estacionário (parada e partida do motor). Em conjunto com a transformada Wavelet utilizam-se métodos de redução de dimensionalidade como, a análise em componentes principais (PCA e a análise Linear Discriminante (LDA). O uso de uma bancada experimental mostra que os resultados da classificação têm alta precisão / Abstract: Electric motors are very important components in most industrial plants equipment. The several faults occurring in induction machines can generate severe consequences in the industrial process. The main problems are related to high production costs, worsening the conditions of process and security, and especially poor quality of the final product. Many of these failures are shown progressive. This work presents a contribution to the study of Signal Processing Techniques Based on Wavelet Packet Transform for extracting parameters of Energy and Entropy, together makes the use of dimensionality reduction methods like the Principal components Analysis (PCA) and Linear Dscriminant Analysis (LDA). This analysis is done from the acquisition of vibration signals in Non-Stationary state (stop and start the engine). The results show that the performance of classification has high accuracy based on experimental work / Doutorado / Mecanica dos Sólidos e Projeto Mecanico / Doutor em Engenharia Mecânica

The Carex nardina complex is a group of one to three species (C. nardina, C. hepburnii, C. stantonensis) and six taxa of unispicate sedges (Cyperaceae), the taxonomy of which has been controversial since the 1800s. As initial DNA phylogenies suggested that the complex was nested within Carex section Filifoliae and sister to C. elynoides, a species often confused with C. nardina and sympatric with it in the western North American Cordillera, analyses were conducted to determine whether C. hepburnii, C. stantonensis and other infraspecific taxa could be the result of hybridization. Morphometric and molecular analyses found no substantial evidence for hybridization and supported the recognition of no taxon beyond C. nardina. Consequently, this study concludes that the complex comprises a single variable species, Carex nardina, distributed throughout arctic North America south through the western Cordillera to New Mexico with a minor portion of its range in northeastern Russia, northwestern Scandinavia and Iceland.

Heavy metal contamination in the environment is a global issue that is likely to increase in the future. This report investigates a construction area in which increased concentrations of the heavy metals cadmium, cobalt, copper, nickel and zinc and a decreased pH-value has been observed in the surface water recipient. The focus is on assessing contamination characteristics and identifying suitable remediation methods to avoid a river protected by environmental quality standards further downstream from getting contaminated. The bedrock in the area is sulphide containing and releases acidic leachate when oxidising, which is assumed to have occurred due to plane blasting and filling of residual rock. The contamination characteristics were assessed with the statistical methods modified double mass analysis and principal components analysis. A water balance was established to obtain the flowrates, discharge volumes and to determine the masses of the released metals in the surface water. Identification of suitable remediation methods was performed through a literature study of available remediation methods and using the findings of the assessments as basis. The results showed that there was a significant increase in metal concentrations and decrease in pH-value roughly around the same time as blasting and filling of residual rocks in the area was begun and that there were elevated levels of sulphide and sulphur, but they could not be specifically linked to any media. The yearly masses of metals released from the area into the surface water were between 77-98 % higher than allowed by the established guidelines. By separating the water assumed to carry the majority of the contaminants from the remaining natural water in the watershed, the volume that needs to be treated can be halved. As the contamination is so extensive, a mixture of remediation methods was proposed, including installing green roofs to decrease the runoff from the area, confining the crushed rock with bentonite and installing a filter for fast, efficient reduction. For long-term remediation, it is suggested to optimise the existing sedimentation basins and wetlands. The conclusions were that it will be very expensive to remediate the contamination, due to the extent and magnitude, and that handling sulphide containing bedrock for construction purposes should be legally regulated in order to avoid negative environmental and economic impacts. / Förorening av tungmetaller i naturen är ett globalt problem som troligtvis kommer öka i framtiden. Den här rapporten undersöker en byggarbetsplats där ökade koncentrationer av metallerna kadmium, kobolt, koppar, nickel och zink samt ett minskat pH-värde har observerats i ytvattenrecipienten. Fokus ligger på att analysera föroreningskaraktärer och identifiera lämpliga åtgärdsmetoder för att undvika att en å nedströms som är skyddad av miljökvalitetsnormer ska förorenas. Berggrunden i området är sulfidförande och släpper ut surt lakvatten när den oxiderar, vilket är antaget har hänt till följd av plansprängning och utfyllnad av överblivet bergmaterial. Föroreningskaraktärerna analyserades med de statistiska metoderna modified double mass analysis och principalkomponentsanalys. En vattenbalans etablerades för att ta fram flöden, volymer och för att bestämma massorna av de frigjorda metallerna i ytvattnet. Identifiering av lämpliga åtgärdsmetoder gjordes med en litteraturstudie av tillgängliga metoder som grund. Resultaten visade att det är en signifikant ökning av metallkoncentrationer och minskning i pH-värde runt samma tid som sprängning och utfyllning av bergmaterial påbörjades samt att det är förhöjda halter av sulfid och svavel, men de kunde inte bli associerade med ett specifikt media. De årliga massorna av frigjorda metaller som släpps ut från området i ytvattnet är mellan 77-98 % högre än tillåtet av de etablerade riktlinjerna. Genom att separera vattnet som kan antas innehålla majoriteten av föroreningarna från det naturliga vattenflödet i avrinningsområdet kan volymen som behöver renas halveras. Eftersom föroreningen är så omfattande föreslås en kombination av åtgärdsmetoder; installation av gröna tak för att minska avrinningen från området, inneslutning av utfyllnadsmaterialet med bentonit och installation av ett filter för snabb, effektiv reduktion. För mer långsiktig rening föreslås det att optimera de existerande sedimentationsdammarna och våtmarken. Slutsatsen är att det kommer bli väldigt dyrt att åtgärda föroreningen på grund av dess omfattning, och hantering av sulfidförande berg för exploateringssyfte borde vara lagstadgat för att undvika miljömässiga och ekonomiska kostnader.

Laser Induced Breakdown Spectroscopy (LIBS) is showing great potential as an atomic analytical technique. With its ability to rapidly analyze all forms of matter, with little-to-no sample preparation, LIBS has many advantages over conventional atomic emission spectroscopy techniques. With the maturation of the technologies that make LIBS possible, there has been a growing movement to implement LIBS in portable analyzers for field applications. In particular, LIBS has long been considered the front-runner in the drive for stand-off detection of trace deposits of explosives. Thus there is a need for a better understanding of the relevant processes that are responsible for the LIBS signature and their relationships to the different system parameters that are helping to improve LIBS as a sensing technology. This study explores the use of LIBS as a method to detect random trace amounts of specific organic materials deposited on organic or non-metallic surfaces. This requirement forces the limitation of single-shot signal analysis. This study is both experimental and theoretical, with a sizeable component addressing data analysis using principal components analysis to reduce the dimensionality of the data, and quadratic discriminant analysis to classify the data. In addition, the alternative approach of ‘target factor analysis’ was employed to improve detection of organic residues on organic substrates. Finally, a new method of characterizing the laser-induced plasma of organics, which should lead to improved data collection and analysis, is introduced. The comparison between modeled and experimental measurements of plasma temperatures and electronic density is discussed in order to improve the present models of low-temperature laser induced plasmas.

Current methods of production forecasting such as decline curve analysis (DCA) or numerical simulation require years of historical production data, and their accuracy is limited by the choice of model parameters. Unconventional resources have proven challenging to apply traditional methods of production forecasting because they lack long production histories and have extremely variable model parameters. This research proposes a data-driven alternative to reservoir simulation and production forecasting techniques. We create a proxy-well model for predicting cumulative oil production by selecting statistically significant well completion parameters and reservoir information as independent predictor variables in regression-based models. Then, principal component analysis (PCA) is applied to extract key features of a well’s time-rate production profile and is used to estimate cumulative oil production. The efficacy of models is examined on field data of over 400 wells in the Eagle Ford Shale in South Texas, supplied from an industry database. The results of this study can be used to help oil and gas companies determine the estimated ultimate recovery (EUR) of a well and in turn inform financial and operational decisions based on available production and well completion data.