Matching Items (12,531)
Filtering by

Clear all filters

161901-Thumbnail Image.png
Description
The need of effective forecasting models for multi-variate time series has been underlined by the integration of sensory technologies into essential applications such as building energy optimizations, flight monitoring, and health monitoring. To meet this requirement, time series prediction techniques have been expanded from uni-variate to multi-variate. However, due to

The need of effective forecasting models for multi-variate time series has been underlined by the integration of sensory technologies into essential applications such as building energy optimizations, flight monitoring, and health monitoring. To meet this requirement, time series prediction techniques have been expanded from uni-variate to multi-variate. However, due to the extended models’ poor ability to capture the intrinsic relationships among variates, naïve extensions of prediction approaches result in an unwanted rise in the cost of model learning and, more critically, a significant loss in model performance. While recurrent models like Long Short-Term Memory (LSTM) and Recurrent Neural Network Network (RNN) are designed to capture the temporal intricacies in data, their performance can soon deteriorate. First, I claim in this thesis that (a) by exploiting temporal alignments of variates to quantify the importance of the recorded variates in relation to a target variate, one can build a more accurate forecasting model. I also argue that (b) traditional time series similarity/distance functions, such as Dynamic Time Warping (DTW), which require that variates have similar absolute patterns are fundamentally ill-suited for this purpose, and that should instead quantify temporal correlation in terms of temporal alignments of key “events” impacting these series, rather than series similarity. Further, I propose that (c) while learning a temporal model with recurrence-based techniques (such as RNN and LSTM – even when leveraging attention strategies) is challenging and expensive, the better results can be obtained by coupling simpler CNNs with an adaptive variate selection strategy. Putting these together, I introduce a novel Selego framework for variate selection based on these arguments, and I experimentally evaluate the performance of the proposed approach on various forecasting models, such as LSTM, RNN, and CNN, for different top-X% percent variates and different forecasting time in the future (lead), on multiple real-world data sets. Experiments demonstrate that the proposed framework can reduce the number of recorded variates required to train predictive models by 90 - 98% while also increasing accuracy. Finally, I present a fault onset detection technique that leverages the precise baseline forecasting models trained using the Selego framework. The proposed, Selego-enabled Fault Detection Framework (FDF-Selego) has been experimentally evaluated within the context of detecting the onset of faults in the building Heating, Ventilation, and Air Conditioning (HVAC) system.
ContributorsTiwaskar, Manoj (Author) / Candan, K. Selcuk (Thesis advisor) / Sapino, Maria Luisa (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2021
161902-Thumbnail Image.png
Description
The GlueX experiment housed in Hall D of the Thomas Jefferson National Laboratory was created to map the light meson spectrum in order to contribute to the Standard Model of particle physics by strengthening our understanding of the strong interaction. GlueX is a medium-energy photoproduction experiment that utilizes a linearly

The GlueX experiment housed in Hall D of the Thomas Jefferson National Laboratory was created to map the light meson spectrum in order to contribute to the Standard Model of particle physics by strengthening our understanding of the strong interaction. GlueX is a medium-energy photoproduction experiment that utilizes a linearly polarized photon beam to create hadronic forms of matter. By mapping the light meson spectrum, the GlueX collaboration hopes to identify meson states forbidden by the Constituent Quark Model. As a main research objective, the GlueX collaboration is searching for hybrid $q\bar{q}g$ meson states that exhibit exotic quantum numbers. One hybrid meson candidate is the $\eta'_1$, which is predicted to decay to $K^\ast\bar{K}$ and have a mass near $2.3~\mathrm{GeV}$ (\citeauthor{qn_exotic_status}, \citeyear{qn_exotic_status}; \citeauthor{hybrid_mesons}, \citeyear{hybrid_mesons}). At this time, very few meson states have been identified in the $2.0~\mathrm{GeV}$ mass region. This dearth of evidence for existing states requires any tool developed to search for meson states above $2.0~\mathrm{GeV}$ must be verified by looking at known meson states. In order to search for the $\eta'_1$ hybrid meson candidate in $\gamma p \rightarrow pK^+K^-\gamma\gamma$ events, meson states decaying $K^\ast\bar{K}$ that contribute to the low mass region must be identified, defined in this document as particles having masses between $1400$ and $1600~\mathrm{MeV}$. Identifying what meson states exist in the low mass region is also critical to mapping the light meson spectrum and determining the quark-gluonic content of those meson states. The results of a partial wave analysis (PWA) of $\gamma p \rightarrow pX$ where $X\rightarrow K^\ast\bar{K}$ from $\gamma p \rightarrow pK^+K^-\gamma\gamma$ events in GlueX are presented. In the $J=0$ invariant mass distribution, the $\eta(1405)$ and $\eta(1475)$ are identified, adding to the debate as to whether two pseudoscalar mesons exist in the low mass region. For the $J=1$ distribution, the $f_1(1420)$ and $f_1(1510)$ axial vector mesons are seen, where the former helps further elaborate on the $E\iota$ puzzle of the twentieth century \citep{E_iota_puzzle}. With respect to the controversy of meson states in the low mass region, evidence for the existence of the $f_2(1430)$ meson is strengthened in the $J=2$ distribution, and the $f'_2(1525)$ state is seen. This work lays a foundation for the ASU Meson Physics Group to continue a wider search for hybrid mesons in the $\gamma p \rightarrow pK^+K^-\gamma\gamma$ reaction topology.
ContributorsCole, Sebastian Miles (Author) / Dugger, Michael (Thesis advisor) / Ritchie, Barry (Committee member) / Alarcon, Ricardo (Committee member) / Shovkovy, Igor (Committee member) / Arizona State University (Publisher)
Created2021
161903-Thumbnail Image.png
Description
This thesis argues that physical landscapes, from intentional sites of memory to average public spaces, play a foundational role in the formation and continuation of the official politics of memory that underpins Serbian cultural memory and collective identity. Thus, in order to understand the complexities of the Serbian collective identity,

This thesis argues that physical landscapes, from intentional sites of memory to average public spaces, play a foundational role in the formation and continuation of the official politics of memory that underpins Serbian cultural memory and collective identity. Thus, in order to understand the complexities of the Serbian collective identity, the landscapes that underpin such an identity must first be understood. Building off prior findings, the three landscapes to be considered relate to three pivotal moments in Serbian nation-building and identity formation: the end of the Ottoman presence, World War II and Yugoslavia, and the wars of the 1990s. This thesis put surveys of Serbian landscapes, which map both sites of remembrance and sites left to be forgotten in Belgrade, as well as oral histories with local young-adult Serbians in conversation in order to elucidate the extent to which individual conceptions of the past and of the Serbian identity correlate to the official politics of memory in Serbia. Young-adult Serbians have been selected, as their only personal experience with each moment of history under consideration is generational memory and state narratives of the past. Ultimately, this study seeks to expand and verify the themes of remembrance found in Serbia as well as understand how the reconstruction of the past, starting from the end of the Ottoman presence to the 1990s war, has figured into the various nation-building projects in Serbia. Building on Halbwachs and Nora, this study understands culture memory as dependent on objectivized culture, like buildings, which naturally challenges the traditional separation of memory and history. Though it does not represent the full Serbian public, this study demonstrates the limited role the physical landscape has in shaping the understanding of the past held by the Serbians interviewees.
ContributorsStull, Madeline (Author) / Manchester, Laurie (Thesis advisor) / Cichopek-Gajraj, Anna (Committee member) / Thompson, Victoria (Committee member) / Arizona State University (Publisher)
Created2021
161904-Thumbnail Image.png
Description
Stressors to marine environments are predicted to increase and affect the well-being of marine ecosystems and coastal communities. Marine protected areas (MPAs) are one most widely implemented interventions for marine stressors. Despite the implementation of thousands of protected areas worldwide, people are still striving to understand their dynamics as they

Stressors to marine environments are predicted to increase and affect the well-being of marine ecosystems and coastal communities. Marine protected areas (MPAs) are one most widely implemented interventions for marine stressors. Despite the implementation of thousands of protected areas worldwide, people are still striving to understand their dynamics as they vary in their efficacy and many MPAs have not met their objectives. Additionally, those that have often fail to protect the ecosystem services and cultural values necessary for human community health. Thus, research has expanded to include analyses of the human and social dimensions that may limit their effectiveness. This dissertation explores the role of community engagement in marine protected areas and perceptions of environmental changes in coastal communities. Currently, existing research on the roles of community engagement in marine conservation interventions is limited, particularly in the island-states of the Caribbean region. This dissertation contains a review of the literature to understand the nuances of community engagement in relation to MPAs. Through the review, it was determined that primary forms of engagement are interviews and surveys, and respondents primarily included businesses, community members, fishers, and resource users. To better understand the perceptions and practices on-the-ground, key informants were interviewed across the Caribbean. There are strong desires to conduct community engagement for innumerable benefits, but there are barriers that some participants have overcome. Sharing information between MPA sites offers an opportunity to effectively engage community members. For the local case study, Charlotteville, Trinidad and Tobago, a small, coastal fishing town in the northeast region of Tobago was selected to understand the role of perceptions of environmental changes. There were strong ties of environmental and social changes, with an emphasis on the impacts of environmental stressors to human health. The heterogeneity and diversity of responses in this chapter highlight the need to consider who is engaged in community engagement activities.
ContributorsBernard, Miranda Lynn (Author) / Gerber, Leah (Thesis advisor) / Buzinde, Christine (Committee member) / Schoon, Michael (Committee member) / Kittinger, Jack (Committee member) / Cheng, Samantha (Committee member) / Arizona State University (Publisher)
Created2021
161905-Thumbnail Image.png
Description
The current study explores the extent to which different processing strategies affect comprehension accuracy and integration of information across multiple texts. Reading comprehension of single texts is a difficult task, in which the challenges are compounded by the need to integrate information across texts. Processing strategies, such as self-explanation and

The current study explores the extent to which different processing strategies affect comprehension accuracy and integration of information across multiple texts. Reading comprehension of single texts is a difficult task, in which the challenges are compounded by the need to integrate information across texts. Processing strategies, such as self-explanation and source-evaluation, help reduce the challenges that readers face when attempting to comprehend texts. Self-explanation has been a successful strategy for coherence-building processes in single text comprehension, but the benefits for supporting inter-textual comprehension have not yet been explored. Source-evaluation supports identification of different sources, which helps resolve inconsistencies between texts; yet it remains unclear whether sourcing alone supports comprehension within as well as between texts. Think-aloud is a strategy intended to encourage further processing of the text without providing any explicit comprehension strategy. The differences between these two strategies prompts questions regarding the adequacy of either strategy for supporting inferencing and integration within and across texts. In this study, participants (n=80) were randomly assigned to one of three strategy conditions: self-explanation, source-evaluation, or think-aloud. Students read four texts after which they completed three types of open-ended comprehension questions (i.e., textbase, intra-textual inference, and inter-textual inference), a source memory task, and individual difference measures. Prior knowledge and reading skill were strongly correlated (r = .65) and showed moderate correlations (r = .31 to .60) with participants’ comprehension accuracy, total number of integrations within their responses, and their memory for sources. Participants were more likely to respond accurately and demonstrate integrations across texts for the text-based questions in comparison to the more challenging inference questions. There was a marginal effect of condition on comprehension question accuracy, wherein participants who self-explained responded more accurately than those who engaged in the think-aloud task. In addition, those in the self-explanation or source-evaluation conditions recalled more sources than those in the think-aloud condition. There were no significant differences in performance between the self-explanation and the source-evaluation conditions. Overall, the results of this study indicate that encouraging students to self-explain and/or evaluate sources while they read multiple documents enhances comprehension and memory for sources.
ContributorsPerret, Cecile Aline (Author) / McNamara, Danielle S (Thesis advisor) / Brewer, Gene (Committee member) / Glenberg, Arthur (Committee member) / Arizona State University (Publisher)
Created2021
161906-Thumbnail Image.png
Description
In many real-world machine learning classification applications, well labeled training data can be difficult, expensive, or even impossible to obtain. In such situations, it is sometimes possible to label a small subset of data as belonging to the class of interest though it is impractical to manually label all data

In many real-world machine learning classification applications, well labeled training data can be difficult, expensive, or even impossible to obtain. In such situations, it is sometimes possible to label a small subset of data as belonging to the class of interest though it is impractical to manually label all data not of interest. The result is a small set of positive labeled data and a large set of unknown and unlabeled data. This is known as the Positive and Unlabeled learning (PU learning) problem, a type of semi-supervised learning. In this dissertation, the PU learning problem is rigorously defined, several common assumptions described, and a literature review of the field provided. A new family of effective PU learning algorithms, the MLR (Modified Logistic Regression) family of algorithms, is described. Theoretical and experimental justification for these algorithms is provided demonstrating their success and flexibility. Extensive experimentation and empirical evidence are provided comparing several new and existing PU learning evaluation estimation metrics in a wide variety of scenarios. The surprisingly clear advantage of a simple recall estimate as the best estimate for overall PU classifier performance is described. Finally, an application of PU learning to the field of solar fault detection, an area not previously explored in the field, demonstrates the advantage and potential of PU learning in new application domains.
ContributorsJaskie, Kristen P (Author) / Spanias, Andreas (Thesis advisor) / Blain-Christen, Jennifer (Committee member) / Tepedelenlioğlu, Cihan (Committee member) / Thiagarajan, Jayaraman (Committee member) / Arizona State University (Publisher)
Created2021
161907-Thumbnail Image.png
Description
Women’s contributions to agriculture are an essential factor in achieving food security in developing countries. In rice production, women’s involvement is usually limited to their labor participation. Differences in gender roles within the household hinder women from accessing productive resources and services compared to their male counterparts, leading to a

Women’s contributions to agriculture are an essential factor in achieving food security in developing countries. In rice production, women’s involvement is usually limited to their labor participation. Differences in gender roles within the household hinder women from accessing productive resources and services compared to their male counterparts, leading to a gender gap in rice productivity. With the steady growth of rice productivity experienced in eastern India, it is essential to reduce the gender gap by providing women equal access to resources. However, there is little information on how the gender gap can be addressed between married couples in a patriarchal family structure like India. This dissertation analyzes the potential impact on rice productivity and input use when the spouse (wife) in the household has given access to resources (e.g., rice variety and credit). The first chapter analyzes the impact of a married couple’s decision-making strategy in choosing rice varieties on rice productivity and input use using an endogenous switching regression. The second chapter estimates the effect of access to financial services on technical efficiency using a stochastic production frontier framework. The last chapter evaluates how joint decision-making strategy influences the inverse relationship between farm size and rice productivity following a yield approach and quantile regression. The findings show that joint decision-making strategy choice leads to a higher rice yield and fertilizer usage while lower labor requirements. Regarding spouse access to financial resources, results show a significant difference in technological and managerial gaps. However, that households with access have a lower predicted rice yield than households without access. The last chapter shows that joint decision-making in the family still left the inverse relationship unchanged in examining the inverse relationship. The dissertation provides two significant implications. First, results provide evidence of gender-differentiated preferences for rice variety within the household that can affect rice productivity and input use. Second, the spouse’s access to credit does not necessarily lead to an increase in rice productivity. Thus, determining the primary purpose of why households avail financial services would be essential in analyzing its impact on productivity to avoid misleading results.
ContributorsMalabayabas, Maria Luz Lazaro (Author) / Mishra, Ashok K (Thesis advisor) / Englin, Jeffrey (Committee member) / Manfredo, Mark (Committee member) / Aggarwal, Rimjhim (Committee member) / Arizona State University (Publisher)
Created2021
161853-Thumbnail Image.png
Description
The current study used a Solomon four-group, experimental design to investigate the influence racist hate speech has on college students' anxiety, affective state, and attentional functioning. This study also examined if racist hate speech has differential impacts between students of color (i.e. targets of racist hate speech) and White students

The current study used a Solomon four-group, experimental design to investigate the influence racist hate speech has on college students' anxiety, affective state, and attentional functioning. This study also examined if racist hate speech has differential impacts between students of color (i.e. targets of racist hate speech) and White students (i.e., non-targets of racist hate speech). Participants included 591 undergraduate students predominantly from Arizona (n = 553, 93.57%). Participants were randomly assigned to one of four groups (i.e., pretest experimental, pretest control, posttest only experimental, and posttest only control). Participants assigned to the experimental condition read a vignette containing a classroom incident of racist hate speech, while participants assigned to the control condition read a vignette containing a classroom incident of speech remarking on a university’s football team. Repeated measures, within-between interaction analyses of variance as well as Spearman's bivariate correlations were conducted. Findings revealed that exposure to racist hate speech in a classroom setting can raise state anxiety for students of color and White students. Unexpectedly, exposure to racist hate speech reduced positive affect among White students, and previous experiences witnessing racist hate speech was associated with greater anxiety and attentional difficulty for White students; however, students of color did not experience changes in affective outcomes following exposure to racist hate speech, and previous experiences with racist hate speech were not associated with affective or attentional outcomes for students of color. The present study and future research on this topic can help to inform university policies and campus initiatives to support students impacted by racist discourse and create a more inclusive learning environment for all students.
ContributorsCruz, Samantha Nicole (Author) / Spanierman, Lisa B (Thesis advisor) / Capielo Rosario, Cristalís (Committee member) / Bebout, Lee (Committee member) / Arizona State University (Publisher)
Created2021
Description
Current Li-ion battery technologies are limited by the low capacities of theelectrode materials and require developments to meet stringent performance demands for future energy storage devices. Electrode materials that alloy with Li, such as Si, are one of the most promising alternatives for Li-ion battery anodes due to their high capacities. Tetrel (Si,

Current Li-ion battery technologies are limited by the low capacities of theelectrode materials and require developments to meet stringent performance demands for future energy storage devices. Electrode materials that alloy with Li, such as Si, are one of the most promising alternatives for Li-ion battery anodes due to their high capacities. Tetrel (Si, Ge, Sn) clathrates are a class of host-guest crystalline structures in which Tetrel elements form a cage framework and encapsulate metal guest atoms. These structures can form with defects such as framework/guest atom substitutions and vacancies which result in a wide design space for tuning materials properties. The goal of this work is to establish structure property relationships within the context of Li-ion battery anode applications. The type I Ba 8 Al y Ge 46-y clathrates are investigated for their electrochemical reactions with Li and show high capacities indicative of alloying reactions. DFT calculations show that Li insertion into the framework vacancies is favorable, but the migration barriers are too high for room temperature diffusion. Then, guest free type I clathrates are investigated for their Li and Na migration barriers. The results show that Li migration in the clathrate frameworks have low energy barriers (0.1- 0.4 eV) which suggest the possibility for room temperature diffusion. Then, the guest free, type II Si clathrate (Na 1 Si 136 ) is synthesized and reversible Li insertion into the type II Si clathrate structure is demonstrated. Based on the reasonable capacity (230 mAh/g), low reaction voltage (0.30 V) and low volume expansion (0.21 %), the Si clathrate could be a promising insertion anode for Li-ion batteries. Next, synchrotron X-ray measurements and pair distribution function (PDF) analysis are used to investigate the lithiation pathways of Ba 8 Ge 43 , Ba 8 Al 16 Ge 30 , Ba 8 Ga 15 Sn 31 and Na 0.3 Si 136 . The results show that the Ba-clathrates undergo amorphous phase transformations which is distinct from their elemental analogues (Ge, Sn) which feature crystalline lithiation pathways. Based on the high capacities and solid-solution reaction mechanism, guest-filled clathrates could be promising precursors to form alloying anodes with novel electrochemical properties. Finally, several high temperature (300-550 °C) electrochemical synthesis methods for Na-Si and Na-Ge clathrates are demonstrated in a cell using a Na β’’-alumina solid electrolyte.
ContributorsDopilka, Andrew (Author) / Chan, Candace K (Thesis advisor) / Zhuang, Houlong (Committee member) / Peng, Xihong (Committee member) / Sieradzki, Karl (Committee member) / Arizona State University (Publisher)
Created2021
161855-Thumbnail Image.png
Description
Image-based deep learning (DL) models are employed to enable the detection of critical heat flux (CHF) based on pool boiling experimental images. Most machine learning approaches for pool boiling to date focus on a single dataset under a certain heater surface, working fluid, and operating conditions. For new datasets collected

Image-based deep learning (DL) models are employed to enable the detection of critical heat flux (CHF) based on pool boiling experimental images. Most machine learning approaches for pool boiling to date focus on a single dataset under a certain heater surface, working fluid, and operating conditions. For new datasets collected under different conditions, a significant effort in re-training the model or developing a new model is required under the assumption that the new dataset has a sufficient amount of labeled data. This research is to explore supervised, semi-supervised, and unsupervised machine learning strategies that are formulated to adapt to two scenarios. The first is when the new dataset has limited labeled data available. This scenario was addressed in chapter 2 of this thesis, where Convolutional Neural Networks (CNNs) and Transfer learning (TL) were used in tackling such situations. The second scenario is when the new dataset has no labeled data available at all. In such cases, this research presents a methodology in Chapter 3, where one of the state-of-the-art Generative Adversarial Networks (GANs) called Fixed-Point GAN is deployed in collaboration with a regular CNN model to tackle the problem. To the best of my knowledge, the approaches presented in chapters 2 and 3 are the first of their kind to utilize TL and GANs to solve the boiling heat transfer problem within the heat transfer community and are a step forward towards obtaining a one-for-all general model.
ContributorsAl-Hindawi, Firas Al (Author) / Wu, Teresa TW (Thesis advisor) / Yoon, Hyunsoo HY (Thesis advisor) / Hu, Han HH (Committee member) / Iquebal, Ashif AI (Committee member) / Arizona State University (Publisher)
Created2021