No description available.

Functional data analysis (FDA) studies data in the form of measurements over a domain as whole entities. Our first focus is on the post-hoc analysis with pairwise and contrast comparisons of the popular functional ANOVA model comparing groups of functional data. Existing contrast tests assume independent functional observations within group. In reality, this assumption may not be satisfactory since functional data are often collected continually overtime on a subject. In this work, we introduce a new linear contrast test that accounts for time dependency among functional group members. For a significant contrast test, it can be beneficial to identify the region of significant difference. In the second part, we propose a non-parametric regression procedure to obtain a locally sparse estimate of functional contrast. Our work is motivated by a biomedical study using Raman spectroscopy to monitor hemodialysis treatment near real-time. With contrast test and sparse estimation, practitioners can monitor the progress of the hemodialysis within session and identify important chemicals for dialysis adequacy monitoring. In the third part, we propose a functional data model for degradation analysis of functional data. Motivated by degradation analysis application of rechargeable Li-ion batteries, we combine state-of-the-art functional linear models to produce fully functional prediction for curves on heterogenous domains. Simulation studies and data analysis demonstrate the advantage of the proposed method in predicting degradation measure than existing method using aggregation method. / Doctor of Philosophy / Functional data analysis (FDA) studies complex data structure in the form of curves and shapes. Our work is motivated by two applications concerning data from Raman spectroscopy and battery degradation study. Raman spectra of a liquid sample are curves with measurements over a domain of wavelengths that can identify chemical composition and whose values signify the constituent concentrations in the sample. We first propose a statistical procedure to test the significance of a functional contrast formed by spectra collected at beginning and at later time points during a dialysis session. Then a follow-up procedure is developed to produce a sparse representation of the contrast functional contrast with clearly identified zero and nonzero regions. The use of this method on contrast formed by Raman spectra of used dialysate collected at different time points during hemodialysis sessions can be adapted for evaluating the treatment efficacy in real time. In a third project, we apply state-of-the-art methodologies from FDA to a degradation study of rechargeable Li-ion batteries. Our proposed methods produce fully functional prediction of voltage discharge curves allowing flexibility in monitoring battery health.

In South Africa the main electricity supplier “Eskom” is struggling to meet the increasing demand. To lower the problematic electricity demand, demand side management projects are implemented by large electricity consumers. Measuring equipment is installed as part of these projects to monitor and manage the electricity consumption. Measured data is stored and can be analysed to produce information used for energy management. This, however, is a difficult and time-consuming task, because there are large volumes of data to filter through. It is repetitive work which can be automated. Various data analysis methods are available. These include plotting charts and tables using software packages or data management products. Manually analysing the data using different methods and software packages can be a long and painstaking process especially with large volumes of historical data. Information needs to be customised for different users. For example, managers need to view the end power usage and the amount of electrical energy that can be saved or was saved. Technical personnel need more detail about the electricity consumption by individual components in their system. To interpret the data in different ways a powerful tool is needed. There are many existing tools and software packages available, but most of them focus on buildings or factories. The software packages also have fixed reporting methods that are usually not customisable. In this study a customisable data analysis interface was developed to provide a solution for all the different needs. This interface is user friendly without limiting its customisable functionality. Data is received via emails, processed and then stored in a database hosted by a web server. Users access a website and configure custom charts and tables using the available data. The charts and tables are then displayed on the client’s own home page when the client logs onto the website. This interface was implemented on a website. The results of the interface were tested by automating existing reports using the customisable data analysis interface. Also when compared with the previous data analysis methods it was easily customisable, where it was very hard to customise the previous data analysis methods. It was found that the development of the customisable data analysis interface reduced man-hours spent on reporting with 70% to 80% for large energy consumers by automating the reports. The man-hours are estimated to have a value of R 20 000 to R 60 000 per month, depending on the salaries of the personnel and the volume of reports. It will help the Demand-Side Management (DSM) projects to become a continuous system to lower electricity consumption by providing information that is useful for energy management. / MIng (Computer and Electronic Engineering), North-West University, Potchefstroom Campus, 2015

In South Africa the main electricity supplier “Eskom” is struggling to meet the increasing demand. To lower the problematic electricity demand, demand side management projects are implemented by large electricity consumers. Measuring equipment is installed as part of these projects to monitor and manage the electricity consumption. Measured data is stored and can be analysed to produce information used for energy management. This, however, is a difficult and time-consuming task, because there are large volumes of data to filter through. It is repetitive work which can be automated. Various data analysis methods are available. These include plotting charts and tables using software packages or data management products. Manually analysing the data using different methods and software packages can be a long and painstaking process especially with large volumes of historical data. Information needs to be customised for different users. For example, managers need to view the end power usage and the amount of electrical energy that can be saved or was saved. Technical personnel need more detail about the electricity consumption by individual components in their system. To interpret the data in different ways a powerful tool is needed. There are many existing tools and software packages available, but most of them focus on buildings or factories. The software packages also have fixed reporting methods that are usually not customisable. In this study a customisable data analysis interface was developed to provide a solution for all the different needs. This interface is user friendly without limiting its customisable functionality. Data is received via emails, processed and then stored in a database hosted by a web server. Users access a website and configure custom charts and tables using the available data. The charts and tables are then displayed on the client’s own home page when the client logs onto the website. This interface was implemented on a website. The results of the interface were tested by automating existing reports using the customisable data analysis interface. Also when compared with the previous data analysis methods it was easily customisable, where it was very hard to customise the previous data analysis methods. It was found that the development of the customisable data analysis interface reduced man-hours spent on reporting with 70% to 80% for large energy consumers by automating the reports. The man-hours are estimated to have a value of R 20 000 to R 60 000 per month, depending on the salaries of the personnel and the volume of reports. It will help the Demand-Side Management (DSM) projects to become a continuous system to lower electricity consumption by providing information that is useful for energy management. / MIng (Computer and Electronic Engineering), North-West University, Potchefstroom Campus, 2015

Graph analysis uses graph data collected on a physical, biological, or social phenomena to shed light on the underlying dynamics and behavior of the agents in that system. Many fields contribute to this topic including graph theory, algorithms, statistics, machine learning, and linear algebra. This dissertation advances a novel framework for dynamic graph analysis that combines numerical, statistical, and streaming algorithms to provide deep understanding into evolving networks. For example, one can be interested in the changing influence structure over time. These disparate techniques each contribute a fragment to understanding the graph; however, their combination allows us to understand dynamic behavior and graph structure. Spectral partitioning methods rely on eigenvectors for solving data analysis problems such as clustering. Eigenvectors of large sparse systems must be approximated with iterative methods. This dissertation analyzes how data analysis accuracy depends on the numerical accuracy of the eigensolver. This leads to new bounds on the residual tolerance necessary to guarantee correct partitioning. We present a novel stopping criterion for spectral partitioning guaranteed to satisfy the Cheeger inequality along with an empirical study of the performance on real world networks such as web, social, and e-commerce networks. This work bridges the gap between numerical analysis and computational data analysis.

This thesis will explore how customers and bartenders accomplish the service encounter in a public house, or bar. Whilst there is a body of existing literature on service encounters, this mainly investigates customer satisfaction and ignores the mundane activities that comprise the service encounter itself. In an attempt to fill this gap, I will examine how the activities unfold sequentially by examining the spoken and embodied conduct of the participants, over the course of the encounter. The data comprise audio -and video- recorded, dyadic and multi-party interactions between customer(s) and bartender(s), occurring at the bar counter. The data were analyzed using conversation analysis (CA) to investigate the talk and embodied conduct of participants, as these unfold sequentially. The first analytic chapter investigates how interactions between customers and bartenders are opened. The analysis reveals practices for communicating availability to enter into a service encounter; with customers being found to do this primarily through embodied conduct, and bartenders primarily through spoken turns. The second analytic chapter investigates the role of objects in the ordering sequence. Specifically, the analysis reveals how the Cash Till and the seating tables in the bar are mobilized by participants to accomplish action. In the third analytic chapter, multi-party interactions are investigated, focusing on the organization of turn-taking when two or more customers interact with one or more bartenders. Here, customers are found to engage in activities where they align as a unit, with a lead speaker, who interacts with the bartender on behalf of the party. In the final analytic chapter, the payment sequence of the service encounter is explored to investigate at what sequential position in the interaction payment, as an action, is oriented to. Analysis reveals that a wallet, purse, or bag, may be displayed and money or a payment card retrieved, in a variety of sequential slots, with each contributing differentially to the efficiency of the interaction. I also find that payment may be prematurely proffered due to the preference for efficiency. Overall, the thesis makes innovative contributions to our understanding of customer and bartender practices for accomplishing core activities in what members come to recognize as a service encounter It also contributes substantially to basic conversation analytic research on openings , which has traditionally been founded on telephone interactions, as well as the action of requesting. I enhance our knowledge of face-to-face opening practices, by revealing that the canonical opening sequence (see Schegloff, 1968; 1979; 1986) is not present, at least in this context. From the findings, I also develop our understanding of how objects constrain, or further, progressivity in interaction; while arguing for the importance of analysing the participants semiotic field in aggregate with talk and embodied conduct. The thesis also contributes to existing literature on multi-party interactions, identifying a new turn-taking practice with a directional flow that works effectively to accomplish ordering. Finally, I contribute to knowledge on the provision of payment, an under-researched yet prominent action in the service encounter. This thesis will show the applicability of CA to service providers; by analysing the talk and embodied conduct in aggregate, effective practices for accomplishing a successful service encounter are revealed.

International Telemetering Conference Proceedings / October 28-31, 1996 / Town and Country Hotel and Convention Center, San Diego, California / With the technological evolution of flying equipment, computing and store capacity we need to have a new view of the methods of acquisition, storage and archiving data.

We present a detailed study of the colours of late-type galaxy discs for ten of the EDisCS galaxy clusters with 0.5 less than or similar to z less than or similar to 0.8. Our cluster sample contains 172 spiral galaxies, and our control sample is composed of 96 field disc galaxies. We deconvolved their ground-based V and I images obtained with FORS2 at the VLT with initial spatial resolutions between 0.4 and 0.8 arcsec to achieve a final resolution of 0.1 arcsec with 0.05 arcsec pixels, which is close to the resolution of the ACS at the HST. After removing the central region of each galaxy to avoid pollution by the bulges, we measured the V - I colours of the discs. We find that 50% of cluster spiral galaxies have disc V - I colours redder by more than 1 sigma of the mean colours of their field counterparts. This is well above the 16% expected for a normal distribution centred on the field disc properties. The prominence of galaxies with red discs depends neither on the mass of their parent cluster nor on the distance of the galaxies to the cluster cores. Passive spiral galaxies constitute 20% of our sample. These systems are not abnormally dusty. They are are made of old stars and are located on the cluster red sequences. Another 24% of our sample is composed of galaxies that are still active and star forming, but less so than galaxies with similar morphologies in the field. These galaxies are naturally located in the blue sequence of their parent cluster colour-magnitude diagrams. The reddest of the discs in clusters must have stopped forming stars more than similar to 5 Gyr ago. Some of them are found among infalling galaxies, suggesting preprocessing. Our results confirm that galaxies are able to continue forming stars for some significant period of time after being accreted into clusters, and suggest that star formation can decline on seemingly long (1 to 5 Gyr) timescales.

We measure the cross-correlation between weak lensing of galaxy images and of the cosmic microwave background (CMB). The effects of gravitational lensing on different sources will be correlated if the lensing is caused by the same mass fluctuations. We use galaxy shape measurements from 139 deg(2) of the Dark Energy Survey (DES) Science Verification data and overlapping CMB lensing from the South Pole Telescope (SPT) and Planck. The DES source galaxies have a median redshift of z(med) similar to 0.7, while the CMB lensing kernel is broad and peaks at z similar to 2. The resulting cross-correlation is maximally sensitive to mass fluctuations at z similar to 0.44. Assuming the Planck 2015 best-fitting cosmology, the amplitude of the DESxSPT cross-power is found to be A(SPT) = 0.88 +/- 0.30 and that from DESxPlanck to be A(Planck) = 0.86 +/- 0.39, where A = 1 corresponds to the theoretical prediction. These are consistent with the expected signal and correspond to significances of 2.9 sigma and 2.2 sigma, respectively. We demonstrate that our results are robust to a number of important systematic effects including the shear measurement method, estimator choice, photo-z uncertainty and CMB lensing systematics. We calculate a value of A = 1.08 +/- 0.36 for DESxSPT when we correct the observations with a simple intrinsic alignment model. With three measurements of this cross-correlation now existing in the literature, there is not yet reliable evidence for any deviation from the expected LCDM level of cross-correlation. We provide forecasts for the expected signal-to-noise ratio of the combination of the five-year DES survey and SPT-3G.

In modern China, air pollution has become an essential environmental problem. Over the last 2 years, the air pollution problem, as measured by PM 2.5 (particulate matter) is getting worse. My report aims to carry out a longitudinal data analysis of the air quality index (AQI) in 3 main economic zones in China. Longitudinal data, or repeated measures data, can be viewed as multilevel data with repeated measurements nested within individuals. I arrive at some conclusions about why the 3 areas have different AQI, mainly attributed to factors like population, GDP, temperature, humidity, and other factors like whether the area is inland or by the sea. The residual variance is partitioned into a between-zone component (the variance of the zone-level residuals) and a within-zone component (the variance of the city-level residuals). The zone residuals represent unobserved zone characteristics that affect AQI. In this report, the model building is mainly according to the sequence described by West et al (2007) with respect to the bottom-up procedures and the reference by Singer, J. D., & Willett, J. B (2003) which includes the non-linear situations. This report also compares the quartic curve model with piecewise growth model with respect to this data. The final model I reached is a piece wise model with time-level and zone-level predictors and also with temperature by time interactions. / text