This dissertation investigates the performance of two-class classi cation credit scoring data sets with low default ratios. The standard two-class parametric Gaussian and naive Bayes (NB), as well as the non-parametric Parzen classi ers are extended, using Bayes' rule, to include either a class imbalance or a Bernoulli prior. This is done with the aim of addressing the low default probability problem. Furthermore, the performance of Parzen classi cation with Silverman and Minimum Leave-one-out Entropy (MLE) Gaussian kernel bandwidth estimation is also investigated. It is shown that the non-parametric Parzen classi ers yield superior classi cation power. However, there is a longing for these non-parametric classi ers to posses a predictive power, such as exhibited by the odds ratio found in logistic regression (LR). The dissertation therefore dedicates a section to, amongst other things, study the paper entitled \Model-Free Objective Bayesian Prediction" (Bernardo 1999). Since this approach to Bayesian kernel density estimation is only developed for the univariate and the uncorrelated multivariate case, the section develops a theoretical multivariate approach to Bayesian kernel density estimation. This approach is theoretically capable of handling both correlated as well as uncorrelated features in data. This is done through the assumption of a multivariate Gaussian kernel function and the use of an inverse Wishart prior. / Dissertation (MSc)--University of Pretoria, 2017. / The financial assistance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed and conclusions arrived at, are those of the authors and are not necessarily to be attributed to the NRF. / Statistics / MSc / Unrestricted

Determining which students are at-risk of poorer outcomes -- such as dropping out, failing classes, or decreasing standardized examination scores -- has become an important area of both research and practice in K-12 education. The models produced from this type of predictive modeling research are increasingly used by high schools in Early Warning Systems to identify which students are at risk and intervene to support better outcomes. It has become common practice to re-build and validate these detectors, district-by-district, due to different data semantics and various risk factors for students in different districts. As these detectors become more widely used, however, a new challenge emerges in applying these detectors across a broad spectrum of school districts with varying availability of past student data. Some districts have insufficient high-quality past data for building an effective detector. Novel approaches that can address the complex data challenges a new district presents are critical for advancing the field. Using an ensemble-based algorithm, I develop a modeling approach that can generate a useful model for a previously unseen district. During the ensembling process, my approach, District Similarity Ensemble Extrapolation (DSEE), weights districts that are more similar to the Target district more strongly during ensembling than less similar districts. Using this approach, I can predict student-at-risk status effectively for unseen districts, across a range of grade ranges, and achieve prediction goodness but ultimately fails to perform better than the previously published Knowles (2015) and Bowers (2012) EWS models proposed for use across districts.

Magister Scientiae Dentium - MSc(Dent) / The investigation was designed to evaluate the scoring systems of pitts (1984), and that of Murray and Majid(1978), when used to monitor caries progression in deciduous teeth. The evaluation.was based on the reproducibility and discrlininatory ability of the two systems. The Reproducibility Study was designed to compare the reproducibility of the two systems, and in addition, to illustrate, firstly the use of the subject as the sampling unit in measuring reproducibility, and secondly, a more sensitive method of measuring reproducibility when analysing caries progression data. The Progression Study was designed to discriminatory ability. In addition the compare use of the the effect on subject as the sampling unit in monitoring caries progression was illustrated in the analysis of this part of the investigation. A sub-sample of the posterior bitewing radiographs of 301, 5 year old children from a Duraphat clinical trial (Murray et al. 1977, Murray and Majid 1978) were re-examined. For the Reproducibility Study 150 sets of radiographs were examined a total of 4 times, (repeated examinations for each method). For the Progression Study three serial bitewing radiographs of 50 children were examined using the two methods. For the Reproducibility Study, Kendall's Tau-B was used as an approxlination of the weighted Kappa as a measure of reproducibility. While the pitts method appeared to be more reliable, the difference .between the tYK>methods was not significant( p~ 05). The surface cannot be used as an independent unit in measuring reproducibility. A method using the subject as the sampling unit was illustrated. Attention was drawn to the need to develop a measure of reproducibility for progression studies which would take into account the magnitude of the disagreement (instead of just disagreement) into the overall index of reproducibility. The use of weighted Kappa is suggested as a more appropriate measure of reproducibility. In the Progression Study Method 1 is more sensitive to the various stages of the disease process and provides a more complete overall picture of the carious process. The proportion of enamel lesions recorded for Method 1 were consistantly higher than that for Method 2. The behaviour of outer and inner enamel lesions differed considerably and Method 1 allowed the behaviour of these lesions to be considered separately. The progression rates were found to be faster with Method 2. With Method 1 30% of enamel lesions per subject had progressed to dentine or been filled 12 months later, the corresponding figure for Method 2 was 50%. Method 2 by excluding outer enamel lesions introduces two biases. The combination of these biases favour overestimating the proportion of lesions deemed to have progressed. The use of Method 2 may lead to the unnecessary loss of valuable data; more surfaces were excluded as being unreadable because of overlap. The average proportion of surfaces per subject recorded as unreadible due to overlap was 7% at baseline, 8% at 12 months and 8% at 24 months, the corresponding figures for Method 2 were 13%, 13% and 22% for Method 2. Method 1 thus appears to offer some advantages. The use of the subject as the sampling unit in analysing caries progression data offers a mnnber of advantages when canpared to the use of the surface as the sampling unit. The findings of the study indicate the proportions of high risk subjects (subjects in whom a large proportion of lesions progressed in a given time period) was low. With Method 1 in only 11% of the subjects did 80-100% of the enamel lesions progress after 12 months. The findings indicate that the Pitts system is the more useful scoring system in studies monitoring caries progression in deciduous teeth. / British Council

The diploma thesis is focused on the appreciation of clients' creditworthiness in obtaining loans at Komerční banka, a.s., which is one of the largest and most widely used banks in the Czech Republic. The work is divided into literary research and practical part. Literary research characterizes and explains the basic concepts of credit management and credit analysis. There are described different methods and models for assessing the creditworthiness of clients, which are further used in the practical part and are the basis for assessing the client's creditworthiness. The second part of the thesis analyzes Komerční banka, a.s., where its loan products are analyzed, and specific credit processes and credit analyzes are explained in model cases.

No description available.

No description available.

No description available.

Doctor of Philosophy / In my research, I developed protein model quality estimation methods aimed at evaluating the reliability of computationally predicted protein models in the absence of experimentally solved ground truth structures. These methods specifically focus on estimating errors within the protein models to quantify their structural accuracy. Recognizing that even the most advanced protein structure prediction techniques may produce models with errors, I also developed a complementary protein model refinement method. This refinement method iteratively optimizes the weakly modeled regions, guided by the error estimation module of my quality estimation approach. The development of these model quality estimation methods, therefore, not only offers valuable insights into the structural reliability of protein models but also contributes to optimizing the overall reliability of protein models generated by state-of-the-art computational methods.

There is a great need in real world applications for estimating depth through electronic means without human intervention. There are many methods in the field which help in autonomously finding depth measurements. Some of which are using LiDAR, Radar, etc. One of the most researched topic in the field of depth measurements is Computer Vision which uses techniques on 2D images to achieve the desired result. Out of the many 3D vision techniques used, stereovision is a field where a lot of research is being done to solve this kind of problem. Human vision plays an important part behind the inspiration and research performed in this field. Stereovision gives a very high spatial resolution of depth estimates which is used for obstacle avoidance, path planning, object recognition, etc. Stereovision makes use of two images in the image pair. These images are taken with two cameras from different views and those two images are processed to get depth information. Processing stereo images has been one of the most intensively sought-after research topics in computer vision. Many factors affect the performance of this approach like computational efficiency, depth discontinuities, lighting changes, correspondence and correlation, electronic noise, etc. An algorithm is proposed which uses texture features obtained using Laws Energy Masks and multi-block approach to perform correspondence matching between stereo pair of images with high baseline. This is followed by forming disparity maps to get the relative depth of pixels in the image. An analysis is also made between this approach to the current state-of-the-art algorithms. A robust method to score and rank the stereo algorithms is also proposed. This approach provides a simple way for researchers to rank the algorithms according to their application needs. / Master of Science / There is a great need in real world applications for estimating depth through electronic means without human intervention. There are many methods in the field which help in autonomously finding depth measurements. Some of which are using LiDAR, Radar, etc. One of the most researched topic in the field of depth measurements is Computer Vision which uses techniques on 2D images to achieve the desired result. Out of the many 3D vision techniques used, stereovision is a field where a lot of research is being done to solve this kind of problem. Human vision plays a important part behind the inspiration and research performed in this field. A large variety of algorithms are being developed to find the measure of depth of ideally each and every point on the pictured scene giving us a very high spatial resolution as compared to other methods. Real world needs of depth estimation and the benefits provided by using stereo vision are the main driving force behind this approach. Stereovision gives a very high spatial resolution which is used for obstacle avoidance, path planning, object recognition, etc. Stereovision makes use of image pairs taken from two cameras with different perspective to estimate depth. The two images in the image pair are taken with two cameras from different views (translational change in view) and those two images are processed to get depth information. The software tool developed is a new approach to perform correspondence matching to find depth using stereo vision concepts. This software tool developed in this work is written in MATLAB. The tools efficiency was evaluated using standard techniques which have been described in detail. The evaluation was also performed by using the software tool with the images collected using a pair of stereo cameras and a tape measure to measure the depth of an object by hand. A scoring method has also been proposed to rank the algorithms in the field of stereo vision.

Frågeställning: Hur ser bankernas kreditbedömning för små och medelstora företag ut och hur kommer den att förändras i och med införandet av de nya regelverken K2 och K3? Hur resonerar banker vid kreditbedömning när företag redovisar enligt K2 och K3?Vill bankerna styra åt en viss tillämpning eller önskar de rapportering på basis av den löpande bokföringen? Syfte: Syftet med detta examensarbete är att undersöka och analysera bankers kreditbedömning för små och medelstora företag, samt vilken påverkan kreditbedömningen kommer få efter införandet av de nya redovisningsregelverken. Metod: Studien hade en deduktiv forskningsansats och metoden var kvalitativ. Semistrukturerade intervjuer genomfördes som datainsamlingsmetod. Empirin analyserades med hjälp av den referensram som blev insamlad under en litteraturstudie och en artikelsökning. Slutsats: Studien visade att kreditbedömningens utgångspunkt är väldigt lika för de undersökta bankerna. Bankernas kreditbedömning har ännu inte påverkats av de nya regelverken K2 och K3 men det kan ändras i framtiden. Bankerna vill framförallt ha tillräckligt med underlag från företagen så att de kan göra en bra kreditbedömning. / Research questions: How does the banks credit assessment for SMEs look like and how will it change with the introduction of the new regulations K2 and K3? How do banks reason at credit assessment when companies report according to K2 and K3? Would banks steer to a particular application or do they wish to get reports on the basis of the daily accounting? Purpose: The purpose of this study is to investigate and analyze banks’ credit assessment for small and medium-sized enterprises, as well as the impact of credit assessment due to the new accounting regulations. Method: The study had a deductive research approach and the methodology was qualitative. Semi-structured interviews were carried out as a data collection method. The empirical data were analyzed using the reference frame which was collected in a literature search and an article search. Conclusion: The study outcome shows that the credit assessment starting point is very alike from the examined banks. The banks credit assessment process has not been affected due to introduction of the new regulations K2 and K3 although it can change in future context. The banks want, above all, to have enough material from the companies so that they can make a good credit assessment.