<p>This dissertation investigates the influence of spatially explicit animal behavior active subsidy distribution patterns. Active subsidies are animal-transported consumption and resources transfers from donor to recipient ecosystems. Active subsidies influence ecosystem structure, function and services in recipient ecosystems. Even though active subsidies affect ecosystem dynamics, most ecosystem models consider the influence of spatially-explicit animal behavior on active subsidy distributions, limiting the ability to predict corresponding spatial impacts across ecosystems. Spatial subsidy research documents the need for systematic models and analyses frameworks to provide generally insights into the relationship between animal space use behavior and active subsidy patterns, and advance knowledge of corresponding ecosystem impacts for a variety of taxa and ecological scenarios.</p> <p> </p> <p>To advance spatial subsidy research, this dissertation employs a combined individual-based and movement ecology approach in abstract modeling frameworks to systematically investigate the influence of 1) animal movement behavior given mortality (chapter 2), 2) animal sociality (chapter 3) and 3) landscape heterogeneity (chapter 4) on active subsidy distribution. This dissertation shows that animal movement behavior, sociality and landscape heterogeneity influence the extent and intensity of active distribution and impacts in recipient ecosystems. Insights from this dissertation demonstrate that accounting for these factors in the development of ecosystem models will consequentially enhance their utility for predicting active subsidy spatial patterns and impacts. This dissertation advances spatial subsidy research by providing a road map for developing a comprehensive, unifying framework of the relationship between animal behavior and active subsidy distributions.</p>

<div><div>Designing trajectories in dynamically complex environments is challenging and can easily become intractable via solely manual design efforts. Thus, the problem is recast to blend traditional astrodynamics approaches with machine learning to develop a rapid and flexible trajectory design framework. This framework incorporates knowledge of the spacecraft performance specifications via the computation of Accessible Regions (ARs) that accommodate specific spacecraft acceleration levels for varied mission scenarios in a complex multi-body dynamical regime. Specifically, pathfinding agents, via Heuristically Accelerated Reinforcement Learning (HARL) and Dijkstra's algorithms, engage in a multi-dimensional combinatorial search to sequence advantageous natural states emerging from the ARs to construct initial guesses for end-to-end transfers. These alternative techniques incorporate various design considerations, for example, prioritizing computational time versus the pursuit of globally optimal solutions to meet multi-objective mission goals. The initial guesses constructed by pathfinding agents then leverage traditional numerical corrections processes to deliver continuous transport of a spacecraft from departure to destination. Solutions computed in the medium-fidelity Circular Restricted Three Body (CR3BP) model are then transitioned to a higher-fidelity ephemeris regime where the impact of time-dependent gravitational influences from multiple bodies is also explored.</div><div><br></div><div>A broad trade-space arises in this investigation in large part due to the rich and diverse dynamical flows available in the CR3BP. These dynamical pathways included in the search space via: (i) a pre-discretized database of known periodic orbit families; (ii) flow-models of these families of orbits/arcs `trained' via the supervised learning algorithms Artificial Neural Networks (ANNs) and Support Vector Machines (SVMs); and, finally (iii) a free-form search that permits selection of both chaotic and ordered motion. All three approaches deliver variety in the constructed transfer paths. The first two options offer increased control over the nature of the transfer geometry while the free-form approach eliminates the need for a priori knowledge about available flows in the dynamical environment. The design framework enables varied transfer scenarios including orbit-orbit transport, s/c recovery during contingency events, and rendezvous with a pre-positioned object at an arrival orbit. Realistic mission considerations such as altitude constraints with respect to a primary are also incorporated.</div></div>

At present, the combination of high drug development costs and external pressure to lower consumer prices is forcing the pharmaceutical industry to innovate in ways unlike ever before. One of the main drivers of increased productivity in research and development recently has been the application of computational methods to the drug discovery process. While this investment has generated promising insights in many cases, there is still much progress to be made.<div><br></div><div>There currently exists a dichotomy in the types of algorithms employed which are roughly defined by the extent to which they compromise predictive accuracy for computational efficiency, and vice versa. Many computational drug discovery algorithms exist which yield commendable predictive power but are typically associated with overwhelming resource costs. High-throughput methods are also available, but often suffer from disappointing and inconsistent performance. <br></div><div><br></div><div>In the world of kinase inhibitor design, which often takes advantage of such computational tools, small molecules tend to have myriad side effects. These are usually caused by off-target binding, especially with other kinases (given the large size of the enzyme family and overall structural conservation), and so inhibitors with tunable selectivity are generally desirable. This issue is compounded when considering therapeutically relevant targets like Abelson Protein Tyrosine Kinase (Abl) and Lymphocyte Specific Protein Tyrosine Kinase (Lck) which have opposing effects in many cancers. <br></div><div><br></div><div>This work attempts to solve both of these problems by creating a methodology which incorporates high-throughput computational drug discovery methods, modern machine learning techniques, and novel protein-ligand binding descriptors based on backbone hydrogen bond (dehydron) wrapping, chosen because of their potential in differentiating between kinases. Using this approach, a procedure was developed to quickly screen focused chemical libraries (in order to narrow the domain of applicability and keep medicinal chemistry at the forefront of development) for detection of selective kinase inhibitors. In particular, five pharmacologically relevant kinases were investigated to provide a proof of concept, including those listed above.</div><div><br></div><div>Ultimately, this work shows that dehydron wrapping indeed has predictive value, though it's likely hindered by common and current issues derived from noisy training data, imperfect feature selection algorithms, and simplifying assumptions made by high-throughput algorithms used for structural determination. It also shows that the procedure's predictive value varies depending on the target, leading to the conclusion that the utility of dehydron wrapping for drug design is not necessarily universal, as originally thought. However, for those targets which are amenable to the concept, there are two major benefits: relatively few features are required to produce modest results; and those structural features chosen are easily interpretable and can thereby improve the overall design process by pointing out regions to optimize within any given lead. Of the five kinases explored, Src and Lck are shown in this work to fit particularly well with the general hypothesis; given their importance in treating cancer and evading off-target related side effects, the developed methodology now has the potential to play a major role in the development of drug candidates which specifically inhibit and avoid these kinases.<br></div>

<div>Magnetic resonance imaging (MRI) has become an invaluable tool in health care. Despite its popularity, there is still an ever-increasing need for faster scans and better image quality. Multi-coil MRI, which uses multiple transmit and/or receive coils, holds the potential to address many of these MRI challenges. Multi-coil MRI systems can utilize parallel transmission (pTx) technology using multi-dimensional radio-frequency (RF) pulses for parallel excitation. The pTx platform is shown to be superior in high-field MRI. Therefore, this dissertation is focused on the RF pulse design and optimization on an MRI system with multiple transceiver coils.</div><div> </div><div>This dissertation addresses three major research topics. First, we investigate the optimization of pTx RF pulses when considering both transmitters and receivers of the MRI system. We term this framework multiple-input multiple-output (MIMO) MRI. The RF pulse design method is modeled by minimizing the excitation error while simultaneously maximizing the signal-to-noise ratio (SNR) of the reconstructed MR image. It further allows a key trade-off between the SNR and the excitation accuracy. Additionally, multiple acceleration factors, different numbers of used receive coils, maximum excitation error tolerance, and different excitation patterns are simulated and analyzed within this model. For a given excitation pattern, our method is shown to improve the SNR by 18-130% under certain acceleration schemes, as compared to conventional parallel transmission methods, while simultaneously controlling the excitation error in a desired scope.</div><div> </div><div>Second, we propose a pTx RF pulse design method that controls the peak local specific absorption rates (SARs) using a compressed set of SAR matrices. RF power, peak local SARs, excitation accuracy, and SNR are simultaneously controlled in the designed pTx RF pulses. An alternative compression method using k-means clustering algorithm is proposed for an upper-bounded estimation of peak local SARs. The performance of the pTx design method is simulated using a human head model and an eight-channel transceiver coil array. The proposed method reduces the 10-g peak local SAR by 44.6-54.2%, as compared to the unconstrained pTx approach, when it has a pre-defined lower bound of SNR and an upper bound of excitation error tolerance. The k-means clustering-based SAR compression model shows its efficiency as it generates a narrower and more accurate overestimation bound than the conventional SAR compression model.</div><div> </div><div>Finally, we propose two machine learning based pTx RF pulse design methods and test them for the ultra-fast pTx RF pulse prediction. The two methods proposed are the kernelized ridge regression (KRR) based pTx RF pulse design and the feedforward neural network (FNN) based pTx RF pulse design. These two methods learn the training pTx RF pulses from the extracted key features of their corresponding B1+ fields. These methods are compared with other supervised learning methods (nearest-neighbor methods, etc.). All learned pTx RF pulses should be reasonably SAR-efficient because training pTx RF pulses are SAR-efficient. Longer computation time and pre-scan time are the drawbacks of the current pTx approach, and we address this issue by instantaneously predicting pTx RF pulses using well-trained machine learning models.</div>

<div> <div> <div> <p>Real-time pricing (RTP) of electricity for consumers has long been argued to be crucial for realizing the many envisioned benefits of demand flexibility in a smart grid. However, many details of how to actually implement a RTP scheme are still under debate. Since most of the organized wholesale electricity markets in the US implement a two-settlement mechanism, with day-ahead electricity price forecasts guiding financial and physical transactions in the next day and real-time ex post prices settling any real-time imbalances, it is a natural idea to let consumers respond to the day-ahead prices in real-time. However, if such an idea is not controlled properly, the inherent closed-loop operation may lead consumers to all respond in the same fashion, causing large swings of real-time demand and prices, which may jeopardize system stability and increase consumers’ financial risks. </p><p><br></p> <p>To overcome the potential uncertainties and undesired demand peak caused by “selfish” behaviors by individual consumers under RTP, in this research, we develop a fully decentralized price-driven demand response (DR) approach under game- theoretical frameworks. In game theory, agents usually make decisions based on their belief about competitors’ states, which needs to maintain a large amount of knowledge and thus can be intractable and implausible for a large population. Instead, we propose using regret-based learning in games by focusing on each agent’s own history and utility received. We study two learning mechanisms: bandit learning with incomplete information feedback, and low regret learning with full information feedback. With the learning in games, we establish performance guarantees for each individual agent (i.e., regret minimization) and the overall system (i.e., bounds on price of anarchy).</p><p><br></p></div></div></div><div><div><div> <p>In addition to the game-theoretical framework for price-driven demand response, we also apply such a framework for peer-to-peer energy trading auctions. The market- based approach can better incentivize the development of distributed energy resources (DERs) on demand side. However, the complexity of double-sided auctions in an energy market and agents’ bounded rationality may invalidate many well-established theories in auction design, and consequently, hinder market development. To address these issues, we propose an automated bidding framework based on multi-armed bandit learning through repeated auctions, and is aimed to minimize each bidder’s cumulative regret. We also use such a framework to compare market outcomes of three different auction designs. </p> </div> </div> </div>

<p>Developing a reliable in-situ non-destructive testing method to determine the strength of in-place concrete is critical because a fast-paced construction schedule exposes concrete pavement and/or structures undergoing substantial loading conditions, even at their early ages. Conventional destructive testing methods, such as compressive and flexural tests, are very time-consuming, which may cause construction delays or cost overruns. Moreover, the curing conditions of the tested cylindrical samples and the in-place concrete pavement/structures are quite different, which may result in different strength values. An NDT method that could directly correlate the mechanical properties of cementitious materials with the sensing results, regardless of the curing conditions, mix design, and size effect is needed for the in-situ application.</p><p>The piezoelectric sensor-based electromechanical impedance (EMI) technique has shown promise in addressing this challenge as it has been used to both monitor properties and detect damages on the concrete structure. Due to the direct and inverse effects of piezoelectric, this material can act as a sensor, actuator, and transducer. This research serves as a comprehensive study to investigate the feasibility and efficiency of using piezoelectric sensor-based EMI to evaluate the strength of newly poured concrete. To understand the fundamentals of this method and enhance the durability of the sensor for in-situ monitoring, this work started with sensor fabrication. It has studied two types of polymer coating on the effect of the durability of the sensor to make it practical to be used in the field.</p><p>The mortar and concrete samples with various mix designs were prepared to ascertain whether the results of the proposed sensing technique were affected by the different mixtures. The EMI measurement and compressive strength testing methods (ASTM C39, ASTM C109) were conducted in the laboratory. The experimental results of mortar samples with different water-to-cement ratios (W/C) and two types of cement (I and III) showed that the correlation coefficient (R²) is higher than 0.93 for all mixes. In the concrete experiments, the correlation coefficient between the EMI sensing index and compressive strength of all mixes is higher than 0.90. The empirical estimation function was established through a concrete slab experiment. Moreover, several trial implementations on highway construction projects (I-70, I-74, and I-465) were conducted to monitor the real-time strength development of concrete. The data processing method and the reliable index of EMI sensing were developed to establish the regression model to correlate the sensing results with the compressive strength of concrete. It has been found that the EMI sensing method and its related statistical index can effectively reflect the compressive strength gain of in-place concrete at different ages.</p><p>To further investigate the in-situ compressive strength of concrete for large-scale structures, we conducted a series of large concrete slabs with the dimension of 8 feet × 12 feet × 8 inches in depth was conducted at outdoor experiments field to simulate real-world conditions. Different types of compressive strength samples, including cast-in-place (CIP) cylinder (4” × 6”) – (ASTM C873), field molded cylinder (4” × 8”) – (ASTM C39), and core drilled sample (4” × 8”) – (ASTM C42) were prepared to compare the compressive strength of concrete. The environmental conditions, such as ambient temperatures and relative humidity, were also recorded. The in-situ EMI monitoring of concrete strength was also conducted. The testing ages in this study were started from 6 hours after the concrete cast was put in place to investigate the early age results and continued up to 365 days (one year) later for long-term monitoring. The results indicate that the strength of the CIP sample is higher than the 4” x 8” molded cylinder , and that core drilled concrete is weaker than the two aforementioned. The EMI results obtained from the slab are close to those obtained from CIP due to similar curing conditions. The EMI results collected from 4 × 8-inch cylinder samples are lower than slab and CIP, which aligns with the mechanical testing results and indicates that EMI could capture the strength gain of concrete over time.</p><p>The consequent database collected from the large slab tests was used to build a prediction model for concrete strength. The Artificial Neuron Network (ANN) was investigated and experimented with to optimize the prediction of performances. Then, a sensitivity analysis was conducted to discuss and understand the critical parameters to predict the mechanical properties of concrete using the ML model. A framework using Generative Adversarial Network (GAN) based on algorithms was then proposed to overcome real data usage restrictions. Two types of GAN algorithms were selected for the data synthesis in the research: Tabular Generative Adversarial Networks (TGAN) and Conditional Tabular Generative Adversarial Networks (CTGAN). The testing results suggested that the CTGAN-NN model shows improved testing performances and higher computational efficiency than the TGAN model. In conclusion, the AI-guided concrete strength sensing and prediction approaches developed in this dissertation will be a steppingstone towards accomplishing the reliable and intelligent assessment of in-situ concrete structures.</p><br>

Dynamic spectrum access enabled by cognitive radio networks are envisioned to drive the next generation wireless networks that can increase spectrum utility by opportunistically accessing unused spectrum. Due to the policy constraint that there could be no interference to the primary (licensed) users, secondary cognitive radios have to continuously sense for primary transmissions. Typically, sensing reports from multiple cognitive radios are fused as stand-alone observations are prone to errors due to wireless channel characteristics. Such dependence on cooperative spectrum sensing is vulnerable to attacks such as Secondary Spectrum Data Falsification (SSDF) attacks when multiple malicious or selfish radios falsify the spectrum reports. Hence, there is a need to quantify the trustworthiness of radios that share spectrum sensing reports and devise malicious node identification and robust fusion schemes that would lead to correct inference about spectrum usage. In this work, we propose an anomaly monitoring technique that can effectively capture anomalies in the spectrum sensing reports shared by individual cognitive radios during cooperative spectrum sensing in a multi-channel distributed network. Such anomalies are used as evidence to compute the trustworthiness of a radio by its neighbours. The proposed anomaly monitoring technique works for any density of malicious nodes and for any physical environment. We propose an optimistic trust heuristic for a system with a normal risk attitude and show that it can be approximated as a beta distribution. For a more conservative system, we propose a multinomial Dirichlet distribution based conservative trust framework, where Josang*s Belief model is used to resolve any uncertainty in information that might arise during anomaly monitoring. Using a machine learning approach, we identify malicious nodes with a high degree of certainty regardless of their aggressiveness and variations introduced by the pathloss environment. We also propose extensions to the anomaly monitoring technique that facilitate learning about strategies employed by malicious nodes and also utilize the misleading information they provide. We also devise strategies to defend against a collaborative SSDF attack that is launched by a coalition of selfish nodes. Since, defense against such collaborative attacks is difficult with popularly used voting based inference models or node centric isolation techniques, we propose a channel centric Bayesian inference approach that indicates how much the collective decision on a channels occupancy inference can be trusted. Based on the measured observations over time, we estimate the parameters of the hypothesis of anomalous and non-anomalous events using a multinomial Bayesian based inference. We quantitatively define the trustworthiness of a channel inference as the difference between the posterior beliefs associated with anomalous and non-anomalous events. The posterior beliefs are updated based on a weighted average of the prior information on the belief itself and the recently observed data. Subsequently, we propose robust fusion models which utilize the trusts of the nodes to improve the accuracy of the cooperative spectrum sensing decisions. In particular, we propose three fusion models: (i) optimistic trust based fusion, (ii) conservative trust based fusion, and (iii) inversion based fusion. The former two approaches exclude untrustworthy sensing reports for fusion, while the last approach utilizes misleading information. All schemes are analyzed under various attack strategies. We propose an asymmetric weighted moving average based trust management scheme that quickly identifies on-off SSDF attacks and prevents quick trust redemption when such nodes revert back to temporal honest behavior. We also provide insights on what attack strategies are more effective from the adversaries* perspective. Through extensive simulation experiments we show that the trust models are effective in identifying malicious nodes with a high degree of certainty under variety of network and radio conditions. We show high true negative detection rates even when multiple malicious nodes launch collaborative attacks which is an improvement over existing voting based exclusion and entropy divergence techniques. We also show that we are able to improve the accuracy of fusion decisions compared to other popular fusion techniques. Trust based fusion schemes show worst case decision error rates of 5% while inversion based fusion show 4% as opposed majority voting schemes that have 18% error rate. We also show that the proposed channel centric Bayesian inference based trust model is able to distinguish between attacked and non-attacked channels for both static and dynamic collaborative attacks. We are also able to show that attacked channels have significantly lower trust values than channels that are not– a metric that can be used by nodes to rank the quality of inference on channels.

<div>Hot surface ignition is a safety design concern for serval industries including mining, aviation, automotive, boilers, and maritime applications. Bleed air ducts, exhaust pipes, combustion liners, and machine tools that are operated at elevated temperatures may be a source of ignition that needs to be accounted for during design. An apparatus for the measurements of minimum hot surface ignition temperature (MHSIT) of 3 aviation fluids (Jet-A, Hydraulic Oil (MIL-PRF-5606) and Lubrication Oil (MIL-PRF-23699)) has been developed. This study expands a widely utilized database of values of MHSIT. The study will expand the current range of design parameters including air temperature, crossflow velocity, fluid temperature, global equivalence ratio, injection method, and the effects of pressure. The expanded data are utilized to continue the development of a physics-anchored data dependent system and machine learning model for the estimation of MHSIT.</div><div><br></div><div>The aviation industry, including Rolls Royce, currently use a database of MHSIT values resulting from experiments conducted in 1988 at the Air Force Research Laboratory (AFRL) within the Wright Patterson Air Force Base in Dayton, OH. Over the three decades since these experiments, the range of operating conditions have significantly broadened in most applications including high performance aircraft engines. For example, the cross-stream air velocities (V) have increased by a factor of two (from ~3.4 m/s to ~6.7 m/s). Expanding the known database to document MHSIT for a range of fuel temperatures (TF), air temperatures (TA), pressure (P) and air velocities (V) is of great interest to the aviation industry. MHSIT data for current aviation fluids such as Jet-A and MIL-PRF-23699 (lubrication oil) and their relation to the design parameters have recently been under investigation in a generic experimental apparatus. </div><div><br></div><div>The current work involves utilization of this generic experimental apparatus to further the understanding of MHSIT through the investigation of intermediate air velocities, global equivalence ratios, injection method, and the effects of pressure. This study investigates the effects of air velocity in a greater degree of granularity by utilizing 0.6 m/s increments. This is done to capture the uncertainty seen in MHSIT values above 3.0 m/s. Furthermore, this study also expands the understanding of the effects of injection method on the MHSIT value with the inclusion of spray injected lubrication oil (MIL-PRF-23699) and stream injected Jet-A. The effects of global equivalence ratio are examined for spray injected Jet-A by modulating the aviation fluid injection rate and the crossflow air velocity in tandem. </div><div><br></div><div>During previous experimental campaigns, it was found that MHSIT did not monotonically increase with crossflow air velocity as previously believed. This new finding inspired a set of experiments that found MHSIT in crossflow to have four proposed ignition regimes: conduction, convective cooling, turbulent mixing, and advection. The current study replicates the results from the initial set of experiments at new conditions and to determine the effects of surface temperature on the regimes. </div><div><br></div><div>The MHSIT of flammable liquids depends on several factors including leak type (spray or stream), liquid temperature, air temperature, velocity, and pressure. ASTM standardized methods for ignition are limited to stagnant and falling drops downward (autoignition) at atmospheric pressure (ASTM E659, ASTM D8211, and ASTM E1491) and at pressures from 218 to 203 kPa (ASTM G72). Past studies have shown that MHSIT decreases with increasing pressure, but the available databases lack results of extensive experimental investigation. Therefore, such data for pressures between 101 to 203 kPa are missing or inadequate. As such the generic experimental apparatus was modified to produce the 101 to 203 kPa air duct pressure levels representative of a typical turbofan engine. </div><div><br></div><div>Machine learning (ML) and deep learning (DL) have become widely available in recent years. Open-source software packages and languages have made it possible to implement complex ML based data analysis and modeling techniques on a wide range of applications. The application of these techniques can expedite existing models or reduce the amount of physical lab investigation time required. Three data sets were utilized to examine the effectiveness of multiple ML techniques to estimate experimental outcomes and to serve as a substitute for additional lab work. To achieve this complex multi-variant regressions and neural networks were utilized to create estimating models. The first data sets of interest consist of a pool fire experiment that measured the flame spread rate as a function of initial fuel temperature for 8 different fuels, including Jet-A, JP-5, JP-8, HEFA-50, and FT-PK. The second data set consists of hot surface ignition data for 9 fuels including 4 alternative piston engine fuels for which properties were not available. The third data set is the MHSIT data generated by the generic experimental apparatus during the investigations conducted to expand the understanding of minimum hot surface ignition temperatures. When properties were not available multiple imputation by chained equations (MICE) was utilized to estimate fluid properties. Training and testing data sets were split up to 70% and 30% of the respective data set being modeled. ML techniques were implemented to analyze the data and R-squared values as high as 92% were achieved. The limitation of machine learning models is also discussed along with the advantages of physics-based approaches. The current study has furthered the application of ML in combustion through use of the MHSIT database.</div>

A variety of complex data has broadened in many research fields such as epidemiology, genomics, and analytical chemistry with the development of science, technologies, and design scheme over the past few decades. For example, in epidemiology, the matched case-crossover study design is used to investigate the association between the clustered binary outcomes of disease and a measurement error in covariate within a certain period by stratifying subjects' conditions. In genomics, high-correlated and high-dimensional(HCHD) data are required to identify important genes and their interaction effect over diseases. In analytical chemistry, multiple time series data are generated to recognize the complex patterns among multiple classes. Due to the great diversity, we encounter three problems in analyzing those complex data in this dissertation. We have then provided several contributions to semiparametric and nonparametric methods for dealing with the following problems: the first is to propose a method for testing the significance of a functional association under the matched study; the second is to develop a method to simultaneously identify important variables and build a network in HDHC data; the third is to propose a multi-class dynamic model for recognizing a pattern in the time-trend analysis. For the first topic, we propose a semiparametric omnibus test for testing the significance of a functional association between the clustered binary outcomes and covariates with measurement error by taking into account the effect modification of matching covariates. We develop a flexible omnibus test for testing purposes without a specific alternative form of a hypothesis. The advantages of our omnibus test are demonstrated through simulation studies and 1-4 bidirectional matched data analyses from an epidemiology study. For the second topic, we propose a joint semiparametric kernel machine network approach to provide a connection between variable selection and network estimation. Our approach is a unified and integrated method that can simultaneously identify important variables and build a network among them. We develop our approach under a semiparametric kernel machine regression framework, which can allow for the possibility that each variable might be nonlinear and is likely to interact with each other in a complicated way. We demonstrate our approach using simulation studies and real application on genetic pathway analysis. Lastly, for the third project, we propose a Bayesian focal-area detection method for a multi-class dynamic model under a Bayesian hierarchical framework. Two-step Bayesian sequential procedures are developed to estimate patterns and detect focal intervals, which can be used for gas chromatography. We demonstrate the performance of our proposed method using a simulation study and real application on gas chromatography on Fast Odor Chromatographic Sniffer (FOX) system. / Doctor of Philosophy / A variety of complex data has broadened in many research fields such as epidemiology, genomics, and analytical chemistry with the development of science, technologies, and design scheme over the past few decades. For example, in epidemiology, the matched case-crossover study design is used to investigate the association between the clustered binary outcomes of disease and a measurement error in covariate within a certain period by stratifying subjects' conditions. In genomics, high-correlated and high-dimensional(HCHD) data are required to identify important genes and their interaction effect over diseases. In analytical chemistry, multiple time series data are generated to recognize the complex patterns among multiple classes. Due to the great diversity, we encounter three problems in analyzing the following three types of data: (1) matched case-crossover data, (2) HCHD data, and (3) Time-series data. We contribute to the development of statistical methods to deal with such complex data. First, under the matched study, we discuss an idea about hypothesis testing to effectively determine the association between observed factors and risk of interested disease. Because, in practice, we do not know the specific form of the association, it might be challenging to set a specific alternative hypothesis. By reflecting the reality, we consider the possibility that some observations are measured with errors. By considering these measurement errors, we develop a testing procedure under the matched case-crossover framework. This testing procedure has the flexibility to make inferences on various hypothesis settings. Second, we consider the data where the number of variables is very large compared to the sample size, and the variables are correlated to each other. In this case, our goal is to identify important variables for outcome among a large amount of the variables and build their network. For example, identifying few genes among whole genomics associated with diabetes can be used to develop biomarkers. By our proposed approach in the second project, we can identify differentially expressed and important genes and their network structure with consideration for the outcome. Lastly, we consider the scenario of changing patterns of interest over time with application to gas chromatography. We propose an efficient detection method to effectively distinguish the patterns of multi-level subjects in time-trend analysis. We suggest that our proposed method can give precious information on efficient search for the distinguishable patterns so as to reduce the burden of examining all observations in the data.

Structural health monitoring and building assessment are crucial to acquire structures’ states and maintain their conditions. Besides human-labor surveys that are subjective, time-consuming, and expensive, autonomous image and video analysis is a faster, more efficient, and non-destructive way. This thesis focuses on crack detection from videos, crack segmentation from images, and building assessment from street view images. For crack detection from videos, three approaches are proposed based on local binary pattern (LBP) and support vector machine (SVM), deep convolution neural network (DCNN), and fully-connected network (FCN). A parametric Naïve Bayes data fusion scheme is introduced that registers video frames in a spatiotemporal coordinate system and fuses information based on Bayesian probability to increase detection precision. For crack segmentation from images, the rotation-invariant property of crack is utilized to enhance the segmentation accuracy. The architectures of several approximately rotation-invariant DCNNs are discussed and compared using several crack datasets. For building assessment from street view images, a framework of multiple DCNNs is proposed to detect buildings and predict their attributes that are crucial for flood risk estimation, including founding heights, foundation types (pier, slab, mobile home, or others), building types (commercial, residential, or mobile home), and building stories. A feature fusion scheme is proposed that combines image feature with meta information to improve the predictions, and a task relation encoding network (TREncNet) is introduced that encodes task relations as network connections to enhance multi-task learning.