In marketing, behaviour of users is analysed in order to discover which channels (for instance TV, Social media etc.) are important for increasing the user’s intention to buy a product. The search for better channel attribution models than the common last-click model is of major concern for the industry of marketing. In this thesis, a probabilistic model for channel attribution has been developed, and this model is demonstrated to be more data-driven than the conventional last- click model. The modelling includes an attempt to include the time aspect in the modelling which have not been done in previous research. Our model is based on studying different sequence length and computing conditional probabilities of conversion by using logistic regression models. A clickstream dataset from an online store was analysed using the proposed model. This thesis has revealed proof of that the last-click model is not optimal for conducting these kinds of analyses.

Heat exchanger networks (HENs) are the backbone of heat integration due to their ability in energy and environmental managements. This thesis deals with two issues on HENs. The first concerns with designing of economically optimal Heat exchanger network (HEN) whereas the second focus on optimal operation of HEN in the presence of uncertainties and disturbances within the network. In the first issue, a pinch technology based optimal HEN design is firstly implemented on a 3–streams heat recovery case study to design a simple HEN and then, a more complex HEN is designed for a coal-fired power plant retrofitted with CO2 capture unit to achieve the objectives of minimising energy penalty on the power plant due to its integration with the CO2 capture plant. The benchmark in this case study is a stream data from (Khalilpour and Abbas, 2011). Improvement to their work includes: (1) the use of economic data to evaluate achievable trade-offs between energy, capital and utility cost for determination of minimum temperature difference; (2) redesigning of the HEN based on the new minimum temperature difference and (3) its comparison with the base case design. The results shows that the energy burden imposed on the power plant with CO2 capture is significantly reduced through HEN leading to utility cost saving maximisation. The cost of addition of HEN is recoverable within a short payback period of about 2.8 years. In the second issue, optimal HEN operation considering range of uncertainties and disturbances in flowrates and inlet stream temperatures while minimizing utility consumption at constant target temperatures based on self-optimizing control (SOC) strategy. The new SOC method developed in this thesis is a data-driven SOC method which uses process data collected overtime during plant operation to select control variables (CVs). This is in contrast to the existing SOC strategies in which the CV selection requires process model to be linearized for nonlinear processes which leads to unaccounted losses due to linearization errors. The new approach selects CVs in which the necessary condition of optimality (NCO) is directly approximated by the CV through a single regression step. This work was inspired by Ye et al., (2013) regression based globally optimal CV selection with no model linearization and Ye et al., (2012) two steps regression based data-driven CV selection but with poor optimal results due to regression errors in the two steps procedures. The advantage of this work is that it doesn’t require evaluation of derivatives hence CVs can be evaluated even with commercial simulators such as HYSYS and UNISIM from among others. The effectiveness of the proposed method is again applied to the 3-streams HEN case study and also the HEN for coal-fired power plant with CO2 capture unit. The case studies show that the proposed methodology provides better optimal operation under uncertainties when compared to the existing model-based SOC techniques.

The data-driven adaptive algorithms are explored as a means of increasing the accuracy of Reynolds-averaged turbulence models. This dissertation presents two new data-driven adaptive computational models for simulating turbulent flow, where partial-but-incomplete measurement data is available. These models automatically adjust (i.e., adapts) the closure coefficients of the Reynolds-averaged Navier-Stokes (RANS) k-ω turbulence equations to improve agreement between the simulated flow and a set of prescribed measurement data. The first approach is the data-driven adaptive RANS k-ω (D-DARK) model. It is validated with three canonical flow geometries: pipe flow, the backward-facing step, and flow around an airfoil. For all 3 test cases, the D-DARK model improves agreement with experimental data in comparison to the results from a non-adaptive RANS k-ω model that uses standard values of the closure coefficients. The second approach is the Retrospective Cost Adaptation (RCA) k-ω model. The key enabling technology is that of retrospective cost adaptation, which was developed for real-time adaptive control technology, but is used in this work for data-driven model adaptation. The algorithm conducts an optimization, which seeks to minimize the surrogate performance, and by extension the real flow-field error. The advantage of the RCA approach over the D-DARK approach is that it is capable of adapting to unsteady measurements. The RCA-RANS k-ω model is verified with a statistically steady test case (pipe flow) as well as two unsteady test cases: vortex shedding from a surface-mounted cube and flow around a square cylinder. The RCA-RANS k-ω model effectively adapts to both averaged steady and unsteady measurement data.

By means of data-driven marketing as well as big data technology, this paper presents the investigation of a case study from travel industry implemented by a combination of propensity model and a business model “2W1H”. The business model “2W1H” represents the purchasing behavior “What to buy”, “When to buy”, and “How to buy”. This paper presents the process of building propensity models for the application in behavioral targeting in travel industry.     Combined the propensity scores from predictive analysis and logistic regression with proper marketing and CRM strategies when communicating with travelers, the business model “2W1H” can perform personalized targeting for evaluating of marketing strategy and performance. By analyzing the business model “2W1H” and the propensity model on each business model, both the validation of the model based on training model and test data set, and the validation of actual marketing activities, it has been proven that predictive analytics plays a vital role in the implementation of travelers’ purchasing behavioral targeting in marketing.

A software component model is the cornerstone of any Component-based Software Development (CBSD) methodology. Such a model defines the modelling elements for constructing software systems. In software system modelling, it is necessary to capture the three elements of a system's behaviour: (i) control (ii) computation and (iii) data. Within a system, computations are performed according to the flow of control or the flow of data, depending on whether computations are control-driven or data-driven. Computations are function evaluations, assignments, etc., which transform data when invoked by control or data flow. Therefore a component model should be able to model control flow, data flow as well as computations. Current component models all model computations, but beside computations tend to model either control flow only or data flow only, but not both. In this thesis, we present a new component model which can model both control flow and data flow. It contains modelling elements that capture control flow and data flow explicitly. Furthermore, the modelling of control flow is separate from that of data flow; this enables the modelling of both control-driven and data-driven computations. The feasibility of the model is shown by means of an implementation of the model, in the form of a prototype tool. The usefulness of the model is then demonstrated for a specific domain, the embedded systems domain, as well as a generic domain. For the embedded systems domain, unlike current models, our model can be used to construct systems that are both control-driven and data-driven. In a generic domain, our model can be used to construct domain models, by constructing control flows and data flows which together define a domain model.

This master's degree thesis focuses on the issues of visualization formerly preprocessed business and technical metadata in a business environment. Within the process of elabora-tion and usage of the collected data in the company, it is necessary to present the data to the users in a comfortable, comprehensible and clear way. The first goal of this thesis is to describe and to specify the term of metadata in the field of theory and on the level of busi-ness, their main structure, their occurrence in a non-visual manner and related places where we can find metadata in the heterogeneous business environment. This part also includes a short introduction to the usage of metadata that is related and originates from business in-telligence and a description of Company encyclopedia that can syndicate these resources for further utilization. When defined, the sources, destinations and purpose of technical and business metadata can be used in the second part of the thesis -- this part is aimed at model-ing the use cases for the visual component that can be applied to business and technical metadata. Use cases will be focused on the roles of the users that will use this component and to discover the primary demands and requirements of these users and the functionality that will be indispensable. After the use cases are defined we can process to the next stage of visual component development -- the data must be visualized itself and we have to find proper means to achieve this with user experience being the main focus. Then we have to encapsulate the visualization with a graphical user interface that will meet the requirements and demands of the users' roles specified by the use cases section by prototyping. Lastly, the results of the previous chapters are used to prototype the visual component suitable for a web environment which is based on principles of reusability, data-driven approach, and uses modern web technologies such as framework and library D3.js, HTML5, CSS3, and SVG.

This dissertation establishes a novel data-driven method to identify language network activation patterns in pediatric epilepsy through the use of the Principal Component Analysis (PCA) on functional magnetic resonance imaging (fMRI). A total of 122 subjects’ data sets from five different hospitals were included in the study through a web-based repository site designed here at FIU. Research was conducted to evaluate different classification and clustering techniques in identifying hidden activation patterns and their associations with meaningful clinical variables. The results were assessed through agreement analysis with the conventional methods of lateralization index (LI) and visual rating. What is unique in this approach is the new mechanism designed for projecting language network patterns in the PCA-based decisional space. Synthetic activation maps were randomly generated from real data sets to uniquely establish nonlinear decision functions (NDF) which are then used to classify any new fMRI activation map into typical or atypical. The best nonlinear classifier was obtained on a 4D space with a complexity (nonlinearity) degree of 7. Based on the significant association of language dominance and intensities with the top eigenvectors of the PCA decisional space, a new algorithm was deployed to delineate primary cluster members without intensity normalization. In this case, three distinct activations patterns (groups) were identified (averaged kappa with rating 0.65, with LI 0.76) and were characterized by the regions of: 1) the left inferior frontal Gyrus (IFG) and left superior temporal gyrus (STG), considered typical for the language task; 2) the IFG, left mesial frontal lobe, right cerebellum regions, representing a variant left dominant pattern by higher activation; and 3) the right homologues of the first pattern in Broca's and Wernicke's language areas. Interestingly, group 2 was found to reflect a different language compensation mechanism than reorganization. Its high intensity activation suggests a possible remote effect on the right hemisphere focus on traditionally left-lateralized functions. In retrospect, this data-driven method provides new insights into mechanisms for brain compensation/reorganization and neural plasticity in pediatric epilepsy.

abstract: Drawn from a trio of manuscripts, this dissertation evaluates the sustainability contributions and implications of deploying underutilized spaces for alternative uses at multiple scales: urban, regional and continental. The first paper considers the use of underutilized spaces at the urban scale for urban agriculture (UA) to meet local sustainability goals in Phoenix, Arizona. Through a data-driven analysis, it demonstrates UA can meet 90% of annual demand for fresh produce, supply local produce in all food deserts, reduce areas underserved by public parks by 60%, and displace >50,000 tons of carbon-dioxide emissions from buildings. The second paper considers marginal agricultural land use for bioenergy crop cultivation to meet future liquid fuels demand from cellulosic biofuels sustainably and profitably. At a wholesale fuel price of $4 gallons-of-gasoline-equivalent, 30 to 90.7 billion gallons of cellulosic biofuels can be supplied by converting 22 to 79.3 million hectares of marginal lands in the Eastern United States (U.S.). Displacing marginal croplands (9.4-13.7 million hectares) reduces stress on water resources by preserving soil moisture. This displacement is comparable to existing land use for first-generation biofuels, limiting food supply impacts. Coupled modeling reveals positive hydroclimate feedback on bioenergy crop yields that moderates the land footprint. The third paper examines the sustainability implications of expanding use of marginal lands for corn cultivation in the Western Corn Belt, a commercially important and environmentally sensitive U.S. region. Corn cultivation on lower quality lands, which tend to overlap with marginal agricultural lands, is shown to be nearly three times more sensitive to changes in crop prices. Therefore, corn cultivation disproportionately expanded into these lands following price spikes. Underutilized spaces can contribute towards sustainability at small and large scales in a complementary fashion. While supplying fresh produce locally and delivering other benefits in terms of energy use and public health, UA can also reduce pressures on croplands and complement non-urban food production. This complementarity can help diversify agricultural land use for meeting other goals, like supplying biofuels. However, understanding the role of market forces and economic linkages is critical to anticipate any unintended consequences due to such re-organization of land use. / Dissertation/Thesis / Doctoral Dissertation Geography 2020

The energy consumed by heating, ventilating, and air conditioning (HVAC) systems has increased in the past two decades. Thus, improving efficiency of HVAC systems has gained more and more attentions. This concern has posed challenges for modeling and optimizing HVAC systems. The traditional methods, such as analytical and statistical methods, are usually computationally complex and involve assumptions that may not hold in practice since HVAC system is a complex, nonlinear, and dynamic system. Data-mining approach is a novel science aiming at extracting system characteristics, identifying models and recognizing patterns from large-size data set. It has proved its power in modeling complex and nonlinear systems through various effective and successful applications in industrial, business, and medical areas. Classical data-mining techniques, such as neural networks and boosting tree have been largely applied into modeling HVAC systems in literature. Evolutionary computation, including swarm intelligence, have rapidly developed in the past decades and then applied to improving the performance of HVAC systems. This research focuses on modeling, optimizing, and controlling an HVAC system. Data-mining algorithms are first utilized to extract predictive models from experimental data set at Energy Resource Station in Ankeney. Evolutionary algorithms are then employed to solve the optimization models converted from the above data-driven models. In the optimization process, two set points of the HVAC system, supply air duct static pressure set point and supply air temperature set point, are controlled aiming at improving the energy efficiency and maintaining thermal comfort. The methodology presented in this research is applicable to various industrial processes other than HVAC systems.

Magister Commercii (Industrial Psychology) - MCom(IPS) / People analytics refer to people-related, data-driven, processes (e.g. trend analyses and data management) aimed at describing and evaluating the effectiveness and efficiency of people management practices and processes in support of business outcomes in order to inform and improve people management initiatives and performance as well as business decision making.