Spelling suggestions: "subject:"datadriven"" "subject:"datadrivet""
181 |
Evaluation of training for building based data managers within a scientifically based reading research programEvans, Michele Denise 29 September 2004 (has links)
No description available.
|
182 |
Building Models for Prediction and Forecasting of Service QualityHellberg, Johan, Johansson, Kasper January 2020 (has links)
In networked systems engineering, operational datagathered from sensors or logs can be used to build data-drivenfunctions for performance prediction, anomaly detection, andother operational tasks [1]. Future telecom services will share acommon communication and processing infrastructure in orderto achieve cost-efficient and robust operation. A critical issuewill be to ensure service quality, whereby different serviceshave very different requirements. Thanks to recent advances incomputing and networking technologies we are able to collect andprocess measurements from networking and computing devices,in order to predict and forecast certain service qualities, such asvideo streaming or data stores. In this paper we examine thesetechniques, which are based on statistical learning methods. Inparticular we will analyze traces from testbed measurements andbuild predictive models. A detailed description of the testbed,which is localized at KTH, is given in Section II, as well as in[2]. / Inom nätverk och systemteknik samlas operativ data från sensorer eller loggar som sedan kan användas för att bygga datadrivna funktioner för förutsägelser om prestanda och andra operationella uppgifter [1]. Framtidens teletjänster kommer att dela en gemensam kommunikation och bearbetnings infrastruktur i syfte att uppnå kostnadseffektiva och robusta nätverk. Ett kritiskt problem med detta är att kunna garantera en hög servicekvalitet. Detta problem uppstår till stor del som ett resultat av att olika tjänster har olika krav. Tack vare nyliga avanceringar inom beräkning och nätverksteknologi har vi kunnat samla in användningsmätningar från nätverk och olika datorenheter för att kunna förutspå servicekvalitet för exempelvis videostreaming och lagring av data. I detta arbete undersöker vi data med hjälp av statistiska inlärningsmetoder och bygger prediktiva modeller. En mer detaljerat beskrivning av vår testbed, som är lokaliserad på KTH, finns i [2]. / Kandidatexjobb i elektroteknik 2020, KTH, Stockholm
|
183 |
Developing the right features : the role and impact of customer and product data in software product developmentFabijan, Aleksander January 2016 (has links)
Software product development companies are increasingly striving to become data-driven. The access to customer feedback and product data has been, with products increasingly becoming connected to the Internet, demonetized. Systematically collecting the feedback and efficiently using it in product development, however, are challenges that large-scale software development companies face today when being faced by large amounts of available data. In this thesis, we explore the collection, use and impact of customer feedback on software product development. We base our work on a 2-year longitudinal multiple-case study research with case companies in the software-intensive domain, and complement it with a systematic review of the literature. In our work, we identify and confirm that large-software companies today collect vast amounts of feedback data, however, struggle to effectively use it. And due to this situation, there is a risk of prioritizing the development of features that may not deliver value to customers. Our contribution to this problem is threefold. First, we present a comprehensive and systematic review of activities and techniques used to collect customer feedback and product data in software product development. Next, we show that the impact of customer feedback evolves over time, but due to the lack of sharing of the collected data, companies do not fully benefit from this feedback. Finally, we provide an improvement framework for practitioners and researchers to use the collected feedback data in order to differentiate between different feature types and to model feature value during the lifecycle. With our contributions, we aim to bring software companies one step closer to data-driven decision making in software product development.
|
184 |
Investigating Documentation Requirements & Practices in Swedish Long-Term Elder Care / Kartläggning av dokumentationen av patienthälsodata inom äldreomsorgenHussein, Yacquub Omar January 2024 (has links)
Background The aging population, particularly in Sweden, has heightened the need for innovative solutions in long-term elder care. This demographic shift results in a wealth of health data, from vital signs to social interactions. Accurate documentation of this data is crucial for providing effective care. However, there is a lack of systematic understanding of the types of health data collected in Swedish elder care. This study aims to bridge the gap between idealized and actual documentation practices. Methods The method combines literature reviews with qualitative techniques such as interviews, aiming to thoroughly understand the documentation of health data in Swedish long-term care facilities for older adults. By coding and categorizing the content into manageable data, researchers can identify patterns, trends, and relationships within the media being studied. Results Current documentation practices in Swedish long-term elder care were examined, revealing a variety of health data types collected, including personal information, assessments, care plans, and incident reports. Legal guidance from the Social Authority is deemed insufficient by managers, particularly concerning the Act on Coherent Documentation (SVOD).Documentation practices vary among municipalities due to operational differences and municipal autonomy. These findings highlight complexities and gaps, necessitating further research and improvement efforts. Conclusion Within the context of long-term elder care in Sweden, effective documentation practices play a crucial role. However, there is vagueness surrounding the specific health data that should be and is documented. This is what has been revealed in my study, drawing from interviews with multiple informants. While digital documentation offers promise, addressing complexity requires coordinated efforts. Streamlining regulations, enhancing interoperability, and improving usability are essential steps toward improving elder care documentation practices. / Bakgrund Den åldrande befolkningen i världen, och särskilt i Sverige, har ökat behovet av innovativa lösningar inom äldreomsorgen då denna demografiska förändring har resulterat i en ackumuleringen av stora mängder hälsodata. För att kunna tillhandahålla effektiv vård så är det viktigt att den samlade hälsodatan dokumenteras korrekt. Det finns dock en brist i den systematiska förståelsen för vilka typer av hälsodata som samlas in inom svensk äldreomsorg. Syftet med denna studie är därför att undersöka klyftan mellan hur hälsodata bör dokumenteras enligt lagar och regler, och hur det faktiskt dokumenteras i praktiken. Metoder Studien kombinerar dokumentanalyser med intervjuer, och har som mål att djupgående förstå dokumentationen av hälsoinformation i svenska vård-och omsorgsboenden. Genom att koda och kategorisera resultatet från dokumentanalyserna och intervjudata kunde mönster, trender och relationer inom det studerade materialet identifieras. Resultat Nuvarande dokumentationspraxis inom svensk äldreomsorg undersöktes. Resultatet visade att en mängd olika typer av hälsoinformation samlas in, däribland personlig information, bedömningar, vårdplaner och incidentrapporter. Juridisk vägledning från Socialstyrelsen bedöms vara otillräcklig av chefer, särskilt när det gäller lagen om sammanhållen dokumentation (SVOD). Dokumentationspraxis varierar mellan kommuner på grund av operativa skillnader samt det kommunala självstyret. Dessa resultat belyser problemet med hur hälsoinformation dokumenteras och de brister som kräver ytterligare forskning och förbättringsåtgärder. Slutsats Inom ramen för äldreomsorgen i Sverige spelar effektiva dokumentationsmetoder en avgörande roll. Det råder dock oklarhet kring vad för hälsoinformation som bör och faktiskt dokumenteras. Detta är vad som avslöjats i denna studie, baserat på intervjuer med flera informanter. Även om digital dokumentation är fördelaktigt så kräver hanteringen av de påvisade problemensamordnade insatser. Förtydligandet av regelverk, förbättrad interoperabilitet och ökadanvändbarheten är viktiga steg mot att förbättra dokumentationspraxis inom äldreomsorgen.
|
185 |
A qualitative assessment and optimization of URANS modelling for unsteady cavitating flowsApte, Dhruv Girish 07 June 2024 (has links)
Cavitation is characterized by the formation of vapor bubbles when the pressure in a working fluid drops sharply below the vapor pressure. These bubbles, upon exiting the low-pressure region burst emanating tremendous amounts of energy. Unsteady cavitating flows have been influential in several aspects from being responsible for erosion damage and vibrations in hydraulic engineering devices to being used for non-invasive medical surgeries and drilling for geothermal energy. While the phenomenon has been investigated using both experimental and numerical methods, it continues to pose a challenge for numerical modelling techniques due to its flow unsteadiness and the cavitation-turbulence interaction. One of the principal aspects to modelling cavitation requires the coupling of a cavitation and a turbulence model. While, scale-resolving turbulence modelling techniques like Direct Numerical Simulations (DNS) and Large Eddy Simulations (LES) upto a certain extent may seem an intuitive solution, the physical complexities involved with cavitation result in extremely high computational costs. Thus, Unsteady Reynolds-Averaged Navier-Stokes (URANS) models have been widely utilized as a workhorse for cavitating simulations. However, URANS models are unable to reproduce the periodic vapor shedding observed in experiments and thus, are often corrected by empirical correction. Recently, some models termed as hybrid RANS-LES models that behave as RANS or LES depending on location of flow have been introduced and employed to model cavitating flows. In addition, there has also been a rise in defining some frameworks that use data from high-fidelity simulations or experiments to drive numerical algorithms and aid standard turbulence modelling procedures for accurately simulating turbulent flows. This dissertation is aimed at (1) evaluating the abilities of these corrections, traditional URANS and hybrid RANS-LES models to model cavitation and (2) optimizing the URANS modelling strategy by designing a methodology driven by experimental data to augment the turbulence modelling to simulate cavitating flow in a converging-diverging nozzle. / Doctor of Philosophy / The famous painting Arion on the Dolphin by the French artist François Boucher shows a dolphin rescuing the poet Arion from the choppy seas after being thrown overboard. Today, seeing silhouettes of dolphins swimming near the shore as the Sun sets is a calming sight. However, as these creatures splash their fins in the water, these fins create a drastic pressure difference resulting in the formation of ribbons of vapor bubbles. As the bubbles exit the low-pressure zones, they collapse and release tremendous amounts of energy. This energy manifests in the form of shockwaves rendering this pleasant sight to the human eye, extremely painful for dolphins. These shocks also impact the metal blades in hydraulic machinery like pumps and ship propellers. This dissertation aims to investigate the physics driving this phenomenon using accurate numerical simulations. We first conduct two-dimensional simulations and observe that standard numerical techniques to model the turbulence are unable to simulate cavitation accurately. The investigation is then extended to three-dimensional simulations using hybrid RANS-LES models that aim to strike a delicate balance between accuracy and efficiency. It is observed that these models are able to reproduce the flow dynamics as observed in experiments but are extremely expensive in terms of computational costs due to the three-dimensional nature of the calculations. The investigation then switches to a data-driven approach where a machine learning algorithm driven by experimental data informs the standard turbulence models and is able to simulate cavitating flows accurately and efficiently.
|
186 |
Data-Focused Decision Making: One School's JourneyKretzer, Sandra A. 13 April 2012 (has links)
The use and analysis of data has become a keystone in national policy for educational improvement and a foundational condition in the award of federal grant monies (U.S. Department of Education, 2008, 2009a, 2009b, 2010). Principals are expected to lead their schools in the use of data and are accountable for adequate yearly progress (AYP) for the No Child Left Behind Act (NCLB). Effective use of data can move educators toward student centric learning plans and interventions which improve achievement. While current literature emphasizes the importance of assessment data used to guide sound instructional decisions, gathering scores and generating reports by grade and level does little at individual schools unless there is strong site-based leadership to guide faculty and staff in targeting areas of improvement, implementing a plan, monitoring progress, and adjusting actions.
This qualitative case study describes how the principal's leadership guided a journey of data-focused decision making at one middle school. This dissertation describes use of data in decision-making processes to promote student learning from the perspective of a school which has been implementing data-focused decision making for several years and was selected for its established use of student assessment data. This research focused on the processes individuals and groups use to better understand and use data within a school context and the role of school leaders in supporting these actions.
The intent of this case study is to describe and understand how school leaders make the use of data an integral part of the operation within a middle school in a large suburban mid-Atlantic school district. By looking at how principals embed data analysis and interpretation in the decision-making processes of the school and engage teachers in the use of data to promote student learning, findings could be useful as a guide to other educational leaders as they implement site based actions and related professional development for school-based leaders and teachers. / Ed. D.
|
187 |
Balanced asphalt mix design and pavement distress predictive models based on machine learningLiu, Jian 22 September 2022 (has links)
Traditional asphalt mix design procedures are empirical and need random and lengthy trials in a laboratory, which can cost much labor, material resources, and finance. The initiative (Material Genome initiative) was launched by President Obama to revitalize American manufacturing. To achieve the objective of the MGI, three major tools which are computational techniques, laboratory experiments, and data analytics methods are supposed to have interacted. Designing asphalt mixture with laboratory and computation simulation methods has developed in recent decades. With the development of data science, establishing a new design platform for asphalt mixture based on data-driven methods is urgent. A balanced mix design, defined as an asphalt mix design simultaneously considering the ability of asphalt mixture to resist pavement distress, such as rutting, cracking, IRI (international roughness index), etc., is still the trend of future asphalt mix design.
The service life of asphalt pavement mainly depends on the properties of the asphalt mixture. Whether asphalt mixture has good properties also depends on advanced asphalt mix design methods. Scientific mix design methods can improve engineering properties of asphalt mixture, further extending pavement life and preventing early distress of flexible pavement. Additionally, in traditional asphalt mix design procedures, the capability to resist pavement distress (rutting, IRI, and fatigue cracking) of a mixture is always evaluated based on laboratory performance tests (Hamburg wheel tracking device, Asphalt Pavement Analyzer, repeated flexural bending, etc.). However, there is an inevitable difference between laboratory tests and the real circumstance where asphalt mixture experiences because the pavement condition (traffic, climate, pavement structure) is varying and complex. The successful application examples of machine learning (ML) in all kinds of fields make it possible to establish the predictive models of pavement distress, with the inputs which contain asphalt concrete materials properties involved in the mix design process.
Therefore, this study utilized historical data acquired from laboratory records, the LTPP dataset, and the NCHRP 1-37A report, data analytics and processing methods, as well as ML models to establish pavement distress predictive models, and then developed an automated and balanced mix design procedure, further lying a foundation to achieve an MGI mix design in the future. Specifically, the main research content can be divided into three parts:1. Established ML models to capture the relationship between properties of the binder, aggregates properties, gradation, asphalt content (effective and absorbed asphalt content), gyration numbers, and mixture volumetric properties for developing cost-saving Superpave and Marshall mix design methods; 2. Developed pavement distress (rutting, IRI, and fatigue cracking) predictive models, based on the inputs of asphalt concrete properties, other pavement materials information, pavement structure, climate, and traffic; 3. Proposed and verified an intelligent and balanced asphalt mix design procedure by combining the mixture properties prediction module, pavement distress predictive models and criteria, and non-dominated Sorting genetic algorithm-Ⅱ (NSGA-Ⅱ). It was discovered determining total asphalt content through predicting effective and absorbed asphalt content indirectly with ML models was more accurate than predicting total asphalt content directly with ML models; Pavement distress predictive models can achieve better predictive results than the calibrated prediction models of Mechanistic-Empirical Pavement Design Guide (MEPDG); The design results for an actual project of surface asphalt course suggested that compared to the traditional ones, the asphalt contents of the 12.5 mm and 19 mm Nominal Maximum Aggregate Size (NMAS) mixtures designed by the automated mix design procedure drop by 7.6% and 13.2%, respectively; the percent passing 2.36 mm sieve of the two types of mixtures designed by the proposed mix design procedure fall by 17.8% and 10.3%, respectively. / Doctor of Philosophy / About 96% of roads are paved with asphalt mixture. Asphalt mixture consists of asphalt, aggregates, and additives. Asphalt mix design refers to the process to determine the proper proportion of aggregates, asphalt, and additives. Traditional asphalt mix design procedures in laboratories are empirical and cost much labor, material resources, and finance. Pavement distresses, for example, cracks are important indicators to assess pavement condition. With the development of data science, machine learning (ML) has been applied to various fields by predicting desired targets. The multi-objective optimization refers to determining the optimal solution of a multiple objectives problem. The study applied ML methods to predict asphalt mixture components' proportions and pavement distress with historical experimental data and pavement condition records from literature and an open-source database. Specifically, the main research content can be divided into three parts:1. Established ML models to predict the proportion of asphalt when aggregates are given; 2. Built ML models to predict pavement distress from pavement materials information, pavement structure, climate, and traffic; 3. Develop a digital asphalt mix design procedure by combining the pavement distress prediction models and a multi-objective optimization algorithm.
|
188 |
Data-driven X-ray Tomographic Imaging and Applications to 4D Material CharacterizationWu, Ziling 05 January 2021 (has links)
X-ray tomography is an imaging technique to inspect objects' internal structures with externally measured data by X-ray radiation non-destructively. However, there are concerns about X-ray radiation damage and tomographic acquisition speed in real-life applications. Strategies with insufficient measurements, such as measurements with insufficient dosage (low-dose) and measurements with insufficient projection angles (sparse-view), have been proposed to relieve these problems but are generally compromising imaging quality. Such a dilemma inspires the development of advanced tomographic imaging techniques, in particular, deep learning algorithms to improve reconstruction results with insufficient measurements. The overall aim of this thesis is to design efficient and robust data-driven algorithms with the help of prior knowledge from physics insights and measurement models.
We first introduce a hierarchical synthesis CNN (HSCNN), which is a knowledge-incorporated data-driven tomographic reconstruction method for sparse-view and low-dose tomography with a split-and-synthesis approach. This proposed learning-based method informs the forward model biases based on data-driven learning but with reduced training data. The learning scheme is robust against sampling bias and aberrations introduced in the forward modeling. High-fidelity X-ray tomographic imaging reconstruction results are obtained with a very sparse number of projection angles for both numerical simulated and physics experiments. Comparison with both conventional non-learning-based algorithms and advanced learning-based approaches shows improved accuracy and reduced training data size. As a result of the split-and-synthesis strategy, the trained network could be transferable to new cases.
We then present a deep learning-based enhancement method, HDrec (hybrid-dose reconstruction algorithm), for low-dose tomography reconstruction via a hybrid-dose acquisition strategy composed of textit{extremely sparse-view normal-dose measurements} and textit{full-view low-dose measurements}. The training is applied for each individual sample without the need of transferring the trained models for other samples. Evaluation of two experimental datasets under different hybrid-dose acquisition conditions shows significantly improved structural details and reduced noise levels compared to results with traditional analytical and regularization-based iterative reconstruction methods from uniform acquisitions under the same amount of total dosage. Our proposed approach is also more efficient in terms of single projection denoising and single image reconstruction. In addition, we provide a strategy to distribute dosage smartly with improved reconstruction quality. When the total dosage is limited, the strategy of combining a very few numbers of normal-dose projections and with not-too-low full-view low-dose measurements greatly outperforms the uniform distribution of the dosage throughout all projections.
We finally apply the proposed data-driven X-ray tomographic imaging reconstruction techniques, HSCNN and HDrec, to the dynamic damage/defect characterization applications for the cellular materials and binder jetting additive manufacturing. These proposed algorithms improve data acquisition speeds to record internal dynamic structure changes.
A quantitative comprehensive framework is proposed to study the dynamic internal behaviors of cellular structure, which contains four modules: (i) In-situ fast synchrotron X-ray tomography, which enables collection of 3D microstructure in a macroscopic volume; (ii) Automated 3D damage features detection to recognize damage behaviors in different scales; (iii) Quantitative 3D structural analysis of the cellular microstructure, by which key morphological descriptors of the structure are extracted and quantified; (iv) Automated multi-scale damage structure analysis, which provides a quantitative understanding of damage behaviors.
In terms of binder jetting materials, we show a pathway toward the efficient acquisition of holistic defect information and robust morphological representation through the integration of (i) fast tomography algorithms, (ii) 3D morphological analysis, and (iii) machine learning-based big data analysis.
The applications to two different 4D material characterization demonstrate the advantages of these proposed tomographic imaging techniques and provide quantitative insights into the global evolution of damage/defect beyond qualitative human observation. / Doctor of Philosophy / X-ray tomography is a nondestructive imaging technique to visualize interior structures of non-transparent objects, which has been widely applied to resolve implicit 3D structures, such as human organs and tissues for clinical diagnosis, contents of baggage for security check, internal defect evolution during additive manufacturing, observing fracturing accompanying mechanical tests, and etc. Multiple planar measurements with sufficient X-ray exposure time among different angles are desirable to reconstruct the unique high-quality 3D internal distribution. However, there are practical concerns about X-ray radiation damage to biology samples or long-time acquisition for dynamic experiments in real-life applications. Insufficient measurements by reducing the number of total measurements or the time for each measurement, are proposed to solve this problem but doing so usually leads to the sacrifice of the reconstruction quality. Computational algorithms are developed for tomographic imaging under these insufficient measurement conditions to obtain reconstructions with improved quality.
Deep learning has been successfully applied to numerous areas, such as in recognizing speech, translating languages, detecting objects, and etc. It has also been applied to X-ray tomographic imaging to improve the reconstruction results by learning the features through thousands to millions of corrupted and ideal reconstruction pairs. The aim of this thesis to design efficient deep learning-based algorithms with the help of physical and measurement priors to reduce the number of training datasets.
We propose two different deep learning-based tomographic imaging techniques to improve reconstruction results with reduced training data under different insufficient measurement conditions. One way is to incorporate prior knowledge of the physics models to reduce the required amount of ground truth data, from thousands to hundreds. The training data requirement is further simplified with another hybrid measurement strategy, which could be implemented on each individual sample with only several high-quality measurements. In the end, we apply these two proposed algorithms to different dynamic damage/defect behavior characterization applications.
Our methods achieve improved reconstruction results with greatly enhanced experimental speeds, which become suitable for dynamic 3D recording. Final results demonstrate the advantages of the proposed tomographic imaging techniques and provide quantitative insights into the global dynamic evolution inside the material. This quantitative analysis also provides a much more comprehensive understanding than qualitative human observation.
|
189 |
Domain-based Frameworks and Embeddings for Dynamics over NetworksAdhikari, Bijaya 01 June 2020 (has links)
Broadly this thesis looks into network and time-series mining problems pertaining to dynamics over networks in various domains. Which locations and staff should we monitor in order to detect C. Difficile outbreaks in hospitals? How do we predict the peak intensity of the influenza incidence in an interpretable fashion? How do we infer the states of all nodes in a critical infrastructure network where failures have occurred? Leveraging domain-based information should make it is possible to answer these questions. However, several new challenges arise, such as (a) presence of more complex dynamics. The dynamics over networks that we consider are complex. For example, C. Difficile spreads via both people-to-people and surface-to-people interactions and correlations between failures in critical infrastructures go beyond the network structure and depend on the geography as well. Traditional approaches either rely on models like Susceptible Infectious (SI) and Independent Cascade (IC) which are too restrictive because they focus only on single pathways or do not incorporate the model at all, resulting in sub-optimality. (b) data sparsity. Additionally, the data sparsity still persists in this space. Specifically, it is difficult to collect the exact state of each node in the network as it is high-dimensional and difficult to directly sample from. (c) mismatch between data and process. In many situations, the underlying dynamical process is unknown or depends on a mixture of several models. In such cases, there is a mismatch between the data collected and the model representing the dynamics. For example, the weighted influenza like illness (wILI) count released by the CDC, which is meant to represent the raw fraction of total population infected by influenza, actually depends on multiple factors like the number of health-care providers reporting the number and public tendency to seek medical advice. In such cases, methods which generalize well to unobserved (or unknown) models are required. Current approaches often fail in tackling these challenges as they either rely on restrictive models, require large volume of data, and/or work only for predefined models.
In this thesis, we propose to leverage domain-based frameworks, which include novel models and analysis techniques, and domain-based low dimensional representation learning to tackle the challenges mentioned above for networks and time-series mining tasks. By developing novel frameworks, we can capture the complex dynamics accurately and analyze them more efficiently. For example, to detect C. Difficile outbreaks in a hospital setting, we use a two-mode disease model to capture multiple pathways of outbreaks and discrete lattice-based optimization framework. Similarly, we propose an information theoretic framework which includes geographically correlated failures in critical infrastructure networks to infer the status of the network components. Moreover, as we use more realistic frameworks to accurately capture and analyze the mechanistic processes themselves, our approaches are effective even with sparse data. At the same time, learning low-dimensional domain-aware embeddings capture domain specific properties (like incidence-based similarity between historical influenza seasons) more efficiently from sparse data, which is useful for subsequent tasks. Similarly, since the domain-aware embeddings capture the model information directly from the data without any modeling assumptions, they generalize better to new models.
Our domain-aware frameworks and embeddings enable many applications in critical domains. For example, our domain-aware frameworks for C. Difficile allows different monitoring rates for people and locations, thus detecting more than 95% of outbreaks. Similarly, our framework for product recommendation in e-commerce for queries with sparse engagement data resulted in a 34% improvement over the current Walmart.com search engine. Similarly, our novel framework leads to a near optimal algorithms, with additive approximation guarantee, for inferring network states given a partial observation of the failures in networks. Additionally, by exploiting domain-aware embeddings, we outperform non-trivial competitors by up to 40% for influenza forecasting. Similarly, domain-aware representations of subgraphs helped us outperform non-trivial baselines by up to 68% in the graph classification task. We believe our techniques will be useful for variety of other applications in many areas like social networks, urban computing, and so on. / Doctor of Philosophy / Which locations and staff should we monitor to detect pathogen outbreaks in hospitals? How do we predict the peak intensity of the influenza incidence? How do we infer the failures in water distribution networks? These are some of the questions on dynamics over networks discussed in this thesis. Here, we leverage the domain knowledge to answer these questions. Specifically, we propose (a) novel optimization frameworks where we exploit domain knowledge for tractable formulations and near-optimal algorithms, and (b) low dimensional representation learning where we design novel neural architectures inspired by domain knowledge. Our frameworks capture the complex dynamics accurately and help analyze them more efficiently. At the same time, our low-dimensional embeddings capture domain specific properties more efficiently from sparse data, which is useful for subsequent tasks. Similarly, our domain-aware embeddings are inferred directly from the data without any modeling assumptions, hence they generalize better. The frameworks and embeddings we develop enable many applications in several domains. For example, our domain-aware framework for outbreak detection in hospitals has more than 95% accuracy. Similarly, our framework for product recommendation in e-commerce for queries with sparse data resulted in a 34% improvement over state-of-the-art e-commerce search engine. Additionally, our approach outperforms non-trivial competitors by up to 40% in influenza forecasting.
|
190 |
Model Reduction of Power NetworksSafaee, Bita 08 June 2022 (has links)
A power grid network is an interconnected network of coupled devices that generate, transmit and distribute power to consumers. These complex and usually large-scale systems have high dimensional models that are computationally expensive to simulate especially in real time applications, stability analysis, and control design. Model order reduction (MOR) tackles this issue by approximating these high dimensional models with reduced high-fidelity representations. When the internal description of the models is not available, the reduced representations are constructed by data. In this dissertation, we investigate four problems regarding the MOR and data-driven modeling of the power networks model, particularly the swing equations.
We first develop a parametric MOR approach for linearized parametric swing equations that preserves the physically-meaningful second-order structure of the swing equations dynamics. Parameters in the model correspond to variations in operating conditions. We employ a global basis approach to develop the parametric reduced model.
We obtain these local bases by $mathcal{H}_2$-based interpolatory model reduction and then concatenate them to form a global basis. We develop a framework to enrich this global basis based on a residue analysis to ensure bounded $mathcal{H}_2$ and $mathcal{H}_infty$ errors over the entire parameter domain.
Then, we focus on nonlinear power grid networks and develop a structure-preserving system-theoretic model reduction framework. First, to perform an intermediate model reduction step, we convert the original nonlinear system to an equivalent quadratic nonlinear model via a lifting transformation. Then, we employ the $mathcal{H}_2$-based model reduction approach, Quadratic Iterative Rational Krylov Algorithm (Q-IRKA). Using a special subspace structure of the model reduction bases resulting from Q-IRKA and the structure of the underlying power network model, we form our final reduction basis that yields a reduced model of the same second-order structure as the original model.
Next, we focus on a data-driven modeling framework for power network dynamics by applying the Lift and Learn approach. Once again, with the help of the lifting transformation, we lift the snapshot data resulting from the simulation of the original nonlinear swing equations such that the resulting lifted-data corresponds to a quadratic nonlinearity.
We then, project the lifted data onto a lower dimensional basis via a singular value decomposition. By employing a least-squares measure, we fit the reduced quadratic matrices to this reduced lifted data. Moreover, we investigate various regularization approaches.
Finally, inspired by the second-order sparse identification of nonlinear dynamics (SINDY) method, we propose a structure-preserving data-driven system identification method for the nonlinear swing equations. Using the special structure on the right-hand-side of power systems dynamics, we choose functions in the SINDY library of terms, and enforce sparsity in the SINDY output of coefficients.
Throughout the dissertation, we use various power network models to illustrate the effectiveness of our approaches. / Doctor of Philosophy / Power grid networks are interconnected networks of devices responsible for delivering electricity to consumers, e.g., houses and industries for their daily needs. There exist mathematical models representing power networks dynamics that are generally nonlinear but can also be simplified by linear dynamics.
Usually, these models are complex and large-scale and therefore take a long time to simulate. Hence, obtaining models of much smaller dimension that can capture the behavior of the original systems with an acceptable accuracy is a necessity. In this dissertation, we focus on approximation of power networks model through the swing equations. First, we study the linear parametric power network model whose operating conditions depend on parameters. We develop an algorithm to replace the original model with a model of smaller dimension and the ability to perform in different operating conditions.
Second, given an explicit representation of the nonlinear power network model, we approximate the original model with a model of the same structure but smaller dimension. In the cases where the mathematical models are not available but only time-domain data resulting from simulation of the model is at hand, we apply an already developed framework to infer a model of a small dimension and a specific nonlinear structure: quadratic dynamics. In addition, we develop a framework to identify the nonlinear dynamics while maintaining their original physically-meaningful structure.
|
Page generated in 0.0575 seconds