No description available.

This thesis is concerned with characterizing the quality of Hidden Markov modeling when learning from limited data. It introduces a new perspective on different sources of errors to describe the impact of undermodeling. Our view is that modeling errors can be decomposed into two primary sources of errors: the approximation error and the estimation error. This thesis takes a first step towards exploring the approximation error of low order HMMs that best approximate the true system of a HMM. We introduce the notion minimality and show that best approximations of the true system with complexity greater or equal to the order of a minimal system are actually equivalent realizations. Understanding this further allows us to explore integer lumping and to present a new way named weighted lumping to find realizations. We also show that best approximations of order strictly less than that of a minimal realization are truly approximations; they are incapable of mimicking the true system exactly. Our work then proves that the resulting approximation error is non-decreasing as the model order decreases, verifying the intuitive idea that increasingly simplified models are less and less descriptive of the true system.

No description available.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第24569号 / 工博第5075号 / 新制||工||1972(附属図書館) / 京都大学大学院工学研究科都市社会工学専攻 / (主査)教授 須崎 純一, 教授 宇野 伸宏, 准教授 松島 格也 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

There are two themes in this dissertation, that is, the effect of universal demand law change on corporate risk-taking in the U.S. property & casualty insurance industry, and the application of hidden Markov model to auto telematics data. The first chapter presents my study in the first theme and the rest two chapters present the other theme. In Chapter 1, "Does Shareholder Litigation Affect Corporate Risk-Taking? Evidence from the Property-Casualty Insurance Industry", I explore whether shareholder litigation affects corporate risk-taking differently depending on distinct organizational structures. I use a law change, called Universal Demand (UD) Law, as an exogenous shock and develop three risk-taking measures that are unique in the U.S. property-casualty insurance industry: leverage risk, asset risk, and underwriting risk. The insurance industry provides an interesting opportunity for the study as shareholders in mutual insurers are an ambiguous concept in the legal world, as opposed to the common argument in the insurance literature. The results show that along with UD law adoption, insurers increase their risk-taking. After taking organizational structures into account, the impact of the law change differentiates. Stock insurers increase all three risk-taking measures while mutual insurers decrease their Leverage Risk and increase Asset Risk measures. For different time windows, stock insurers respond faster with respect to their Asset Risk compared to mutual insurers. In addition, I proceed to examine the main economic channel for the impact and find that the free cash flow argument is not the main channel. Chapters 2 and 3 present the study in auto telematics data using a proprietary data source. Both studies are based on the application of hidden Markov model (HMM). Specifically, Chapter 2, "Auto Insurance Pricing Using Telematics Data: Application of a Hidden Markov Model", develops an HMM-based clustering framework to predict auto insurance losses using driving characteristics extracted from telematics data. Through a simulation experiment based on a proprietary telematics data set, I show that HMM can effectively classify driving trips using model-implied hidden states, and HMM-based pricing methods provide better predictive power measured by both deviance statistics and mean squared error. Importantly, the proposed framework not only enables us to price usage-based insurances at a granular level, but it is also viable for estimating long-term insurance losses utilizing the limiting properties of HMM. Chapter 3, "Theoretical Framework of a 3-Layer Hidden Markov Model for Auto Insurance Pricing", is a theoretical extension of the second chapter to improve the framework at a more granular level. I develop a 3-layer HMM for risk classification, which links driving behavior characteristics with risk classes and loss estimation. The proposed model presents a direct structure among all variables and utilizes time series data without aggregation. Furthermore, this study provides a theoretical framework to estimate the 3-layer HMM using the Expectation-Maximization (EM) algorithm. The parameters of Bernoulli distributed loss count (per unit of time) and Gamma distributed loss severity can be solved at least numerically, and the negative definite Hessian matrix indicates that the solution of the first-order condition of the log-likelihood function achieves its local maximum. / Business Administration/Risk Management and Insurance

<p>The decision making processes in today's manufacturing systems represent very complex and challenging tasks. The desired flexibility in terms of the functionality of a machine adds more components to the machine. The real time monitoring and reporting generates large streams of data. However the intelligent and real time processing of this large collection of system data is at the core of the manufacturing decision support tools. </p> <p>This thesis outlines the use of Frequent Episodes in Event Sequences and Hidden Markov Modeling of throughput, quality and maintenance data to model the deterioration of performance in the components that make up the manufacturing system. The thesis also introduces the concept of decision points and outlines how to integrate the total cost function in a business model. </p> This thesis deals with the following three topics: <p>First, the component-based data structure of the manufacturing system is outlined especially throughput, quality and maintenance data. In this approach, the manufacturing system is considered as a group of components that interact with each other and with raw materials to produce the manufactured product. This interaction creates a considerable amount of data which can be associated with the relevant components of the system. The relations between the manufacturing components are established on a physical and logical basis. The components properties are clearly defined in database tables specifically created for this application. The thesis also discusses the web services in manufacturing systems and the portable technologies used in plant decision support tools. </p> <p>Second, the thesis presents a novel application of Frequent Episodes in Event Sequences to identify patterns in the deterioration of performance in a component using frequent episodes of operational failures, quality failures and maintenance activities. A Hidden Markov Model (HMM) is used to model each deterioration episode to estimate the states of performance and the transition rates between the states. The thesis compares the results generated by this model to other existing models of component performance deterioration while emphasizing the benefits ofthe proposed model through the use of the plant data.</p> <p>Finally the thesis presents a methodology usmg HMM probability distributions and Bayesian Decision theory framework to provide a set of decisions and recommendations under the condition of data uncertainty. The results of this analysis are then integrated in the plant maintenance business model.</p> <p>It is worthwhile mentioning that to develop the techniques and validate the results in this research; a Manufacturing Execution System (MES) was developed to operate in an automotive engine plant. All the data and results in this research are based on the plant data. The MES which was developed in this research provided significant benefits in the plant and was adapted by many other GM plants around the world.</p> / Thesis / Doctor of Philosophy (PhD)

Rapidly accumulating genomic data as a result of high-throughput sequencing has necessitated development of efficient computational methods to decode the biological information underlying these data. DNA composition varies across structurally or functionally different regions of a genome as well as those of distinct evolutionary origins. We adapted an integrative framework that combines a top-down, recursive segmentation algorithm with a bottom-up, agglomerative clustering algorithm to decipher compositionally distinct regions in genomes. The recursive segmentation procedure entails fragmenting a genome into compositionally distinct segments within a statistical hypothesis testing framework. This is followed by an agglomerative clustering procedure to group compositionally similar segments within the same framework. One of our main objectives was to decipher distinctive evolutionary patterns in sex chromosomes via unraveling the underlying compositional heterogeneity. Application of this approach to the human X-chromosome provided novel insights into the stratification of the X chromosome as a consequence of punctuated recombination suppressions between the X and Y from the distal long arm to the distal short arm. Novel "evolutionary strata" were identified particularly in the X conserved region (XCR) that is not amenable to the X-Y comparative analysis due to massive loss of the Y gametologs following recombination cessation. Our compositional based approach could circumvent the limitations of the current methods that depend on X-Y (or Z-W for ZW sex determination system) comparisons by deciphering the stratification even if only the sequence of sex chromosome in the homogametic sex (i.e. X or Z chromosome) is available. These studies were extended to the plant sex chromosomes which are known to have a number of evolutionary strata that formed at the initial stage of their evolution, presenting an opportunity to examine the onset of stratum formation on the sex chromosomes. Further applications included detection of horizontally acquired DNAs in extremophilic eukaryote, Galdieria sulphuraria, which encode variety of potentially adaptive functions, and in the taxonomic profiling of metagenomic sequences. Finally, we discussed how the Markovian segmentation and clustering method can be made more sensitive and robust for further applications in biological and biomedical sciences in future.

No description available.

No description available.

No description available.