Master of Science / Department of Computing and Information Sciences / William H. Hsu / Logistic regression is a model for function estimation that measures the relationship between independent variables and a categorical dependent variable, and by approximating a conditional probabilistic density function using a logistic function, also known as a sigmoidal function. Multinomial logistic regression is used to predict categorical variables where there can be more than two categories or classes. The most common type of algorithm for optimizing the cost function for this model is gradient descent. In this project, I implemented logistic regression using conjugate gradient descent (CGD). I used the 20 Newsgroups data set collected by Ken Lang. I compared the results with those for existing implementations of gradient descent. The conjugate gradient optimization methodology outperforms existing implementations.

Master of Science / Department of Computing and Information Sciences / Doina Caragea / With the rising trend in social networking, more people express their opinions on the web. As a consequence, there has been an increase in the number of blogs where people write reviews about the products they buy or services they experience. These reviews can be very helpful to other potential customers who want to know the pros and cons of a product, and also to manufacturers who want to get feedback from customers about their products. Sentiment analysis of online data (such as review blogs) is a rapidly growing field of research in Machine Learning, which can leverage online reviews and quickly extract the sentiment of a whole blog. The accuracy of a sentiment analyzer relies heavily on correctly identifying associations between a sentiment (opinion) word and the targeted mention (token or object) in blog sentences. In this work, we focus on the task of automatically identifying sentiment-mention associations, in other words, we identify the target mention that is associated with a sentiment word in a sentence. Support Vector Machines (SVM), a supervised machine learning algorithm, was used to learn classifiers for this task. Syntactic and semantic features extracted from sentences were used as input to the SVM algorithm. The dataset used in the work has reviews from car and camera domain. The work is divided into two phases. In the first phase, we learned domain specific classifiers for the car and camera domains, respectively. To further improve the predictions of the domain specific classifiers we investigated the use of transfer learning techniques in the second phase. More precisely, the goal was to use knowledge from a source domain to improve predictions for a target domain. We considered two transfer learning approaches: a feature level fusion approach and a classifier level fusion approach. Experimental results show that transfer learning can help to improve the predictions made using the domain specific classifier approach. While both the feature level and classifier level fusion approaches were shown to improve the prediction accuracy, the classifier level fusion approach gave better results.

This thesis considers the automatic acquisition of knowledge about discourse connectives. It focuses in particular on their semantic properties, and on the relationships that hold between them. There is a considerable body of theoretical and empirical work on discourse connectives. For example, Knott (1996) motivates a taxonomy of discourse connectives based on relationships between them, such as HYPONYMY and EXCLUSIVE, which are defined in terms of substitution tests. Such work requires either great theoretical insight or manual analysis of large quantities of data. As a result, to date no manual classification of English discourse connectives has achieved complete coverage. For example, Knott gives relationships between only about 18% of pairs obtained from a list of 350 discourse connectives. This thesis explores the possibility of classifying discourse connectives automatically, based on their distributions in texts. This thesis demonstrates that state-of-the-art techniques in lexical acquisition can successfully be applied to acquiring information about discourse connectives. Central to this thesis is the hypothesis that distributional similarity correlates positively with semantic similarity. Support for this hypothesis has previously been found for word classes such as nouns and verbs (Miller and Charles, 1991; Resnik and Diab, 2000, for example), but there has been little exploration of the degree to which it also holds for discourse connectives. We investigate the hypothesis through a number of machine learning experiments. These experiments all use unsupervised learning techniques, in the sense that they do not require any manually annotated data, although they do make use of an automatic parser. First, we show that a range of semantic properties of discourse connectives, such as polarity and veridicality (whether or not the semantics of a connective involves some underlying negation, and whether the connective implies the truth of its arguments, respectively), can be acquired automatically with a high degree of accuracy. Second, we consider the tasks of predicting the similarity and substitutability of pairs of discourse connectives. To assist in this, we introduce a novel information theoretic function based on variance that, in combination with distributional similarity, is useful for learning such relationships. Third, we attempt to automatically construct taxonomies of discourse connectives capturing substitutability relationships. We introduce a probability model of taxonomies, and show that this can improve accuracy on learning substitutability relationships. Finally, we develop an algorithm for automatically constructing or extending such taxonomies which uses beam search to help find the optimal taxonomy.

As the real estate market plays a central role in a countries financial situation, as a life insurer, a bank and a property developer, Skandia wants a method for better assessing the risks connected to the real estate market. The goal of this paper is to increase the understanding of property market risk and its covariate risks and to conduct an analysis of how a fall in real estate prices could affect Skandia’s exposed assets.This paper explores a recurrent neural network model with the aim of quantifying identified risk factors using exogenous data. The recurrent neural network model is compared to a vector autoregressive model with exogenous inputs that represent economic conditions.The results of this paper are inconclusive as to which method that produces the most accurate model under the specified settings. The recurrent neural network approach produces what seem to be better results in out-of-sample validation but both the recurrent neural network model and the vector autoregressive model fail to capture the hypothesized relationship between the exogenous and modeled variables. However producing results that does not fit previous assumptions, further research into artificial neural networks and tests with additional variables and longer sample series for calibration is suggested as the model preconditions are promising.

Healthcare is an integral component in peoples lives, especially for the rising elderly population, and must be affordable. The United States Medicare program is vital in serving the needs of the elderly. The growing number of people enrolled in the Medicare program, along with the enormous volume of money involved, increases the appeal for, and risk of, fraudulent activities. For many real-world applications, including Medicare fraud, the interesting observations tend to be less frequent than the normative observations. This difference between the normal observations and those observations of interest can create highly imbalanced datasets. The problem of class imbalance, to include the classification of rare cases indicating extreme class imbalance, is an important and well-studied area in machine learning. The effects of class imbalance with big data in the real-world Medicare fraud application domain, however, is limited. In particular, the impact of detecting fraud in Medicare claims is critical in lessening the financial and personal impacts of these transgressions. Fortunately, the healthcare domain is one such area where the successful detection of fraud can garner meaningful positive results. The application of machine learning techniques, plus methods to mitigate the adverse effects of class imbalance and rarity, can be used to detect fraud and lessen the impacts for all Medicare beneficiaries. This dissertation presents the application of machine learning approaches to detect Medicare provider claims fraud in the United States. We discuss novel techniques to process three big Medicare datasets and create a new, combined dataset, which includes mapping fraud labels associated with known excluded providers. We investigate the ability of machine learning techniques, unsupervised and supervised, to detect Medicare claims fraud and leverage data sampling methods to lessen the impact of class imbalance and increase fraud detection performance. Additionally, we extend the study of class imbalance to assess the impacts of rare cases in big data for Medicare fraud detection. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2018. / FAU Electronic Theses and Dissertations Collection

Developments in advanced technologies, such as DNA microarrays, have generated tremendous amounts of data available to researchers in the field of bioinformatics. These state-of-the-art technologies present not only unprecedented opportunities to study biological phenomena of interest, but significant challenges in terms of processing the data. Furthermore, these datasets inherently exhibit a number of challenging characteristics, such as class imbalance, high dimensionality, small dataset size, noisy data, and complexity of data in terms of hard to distinguish decision boundaries between classes within the data. In recognition of the aforementioned challenges, this dissertation utilizes a variety of machine-learning and data-mining techniques, such as ensemble classification algorithms in conjunction with data sampling and feature selection techniques to alleviate these problems, while improving the classification results of models built on these datasets. However, in building classification models researchers and practitioners encounter the challenge that there is not a single classifier that performs relatively well in all cases. Thus, numerous classification approaches, such as ensemble learning methods, have been developed to address this problem successfully in a majority of circumstances. Ensemble learning is a promising technique that generates multiple classification models and then combines their decisions into a single final result. Ensemble learning often performs better than single-base classifiers in performing classification tasks. This dissertation conducts thorough empirical research by implementing a series of case studies to evaluate how ensemble learning techniques can be utilized to enhance overall classification performance, as well as improve the generalization ability of ensemble models. This dissertation investigates ensemble learning techniques of the boosting, bagging, and random forest algorithms, and proposes a number of modifications to the existing ensemble techniques in order to improve further the classification results. This dissertation examines the effectiveness of ensemble learning techniques on accounting for challenging characteristics of class imbalance and difficult-to-learn class decision boundaries. Next, it looks into ensemble methods that are relatively tolerant to class noise, and not only can account for the problem of class noise, but improves classification performance. This dissertation also examines the joint effects of data sampling along with ensemble techniques on whether sampling techniques can further improve classification performance of built ensemble models. / Includes bibliography. / Dissertation (Ph.D.)--Florida Atlantic University, 2016. / FAU Electronic Theses and Dissertations Collection

Recent technological advances underpinned by high throughput sequencing have given new insights into the three-dimensional structure of mammalian genomes. Chromatin conformation assays have been the critical development in this area, particularly the Hi-C method which ascertains genome-wide patterns of intra and inter-chromosomal contacts. However many open questions remain concerning the functional relevance of such higher order structure, the extent to which it varies, and how it relates to other features of the genomic and epigenomic landscape. Current knowledge of nuclear architecture describes a hierarchical organisation ranging from small loops between individual loci, to megabase-sized self-interacting topological domains (TADs), encompassed within large multimegabase chromosome compartments. In parallel with the discovery of these strata, the ENCODE project has generated vast amounts of data through ChIP-seq, RNA-seq and other assays applied to a wide variety of cell types, forming a comprehensive bioinformatics resource. In this work we combine Hi-C datasets describing physical genomic contacts with a large and diverse array of chromatin features derived at a much finer scale in the same mammalian cell types. These features include levels of bound transcription factors, histone modifications and expression data. These data are then integrated in a statistically rigorous way, through a predictive modelling framework from the machine learning field. These studies were extended, within a collaborative project, to encompass a dataset of matched Hi-C and expression data collected over a murine neural differentiation timecourse. We compare higher order chromatin organisation across a variety of human cell types and find pervasive conservation of chromatin organisation at multiple scales. We also identify structurally variable regions between cell types, that are rich in active enhancers and contain loci of known cell-type specific function. We show that broad aspects of higher order chromatin organisation, such as nuclear compartment domains, can be accurately predicted in a variety of human cell types, using models based upon underlying chromatin features. We dissect these quantitative models and find them to be generalisable to novel cell types, presumably reflecting fundamental biological rules linking compartments with key activating and repressive signals. These models describe the strong interconnectedness between locus-level patterns of local histone modifications and bound factors, on the order of hundreds or thousands of basepairs, with much broader compartmentalisation of large, multi-megabase chromosomal regions. Finally, boundary regions are investigated in terms of chromatin features and co-localisation with other known nuclear structures, such as association with the nuclear lamina. We find boundary complexity to vary between cell types and link TAD aggregations to previously described lamina-associated domains, as well as exploring the concept of meta-boundaries that span multiple levels of organisation. Together these analyses lend quantitative evidence to a model of higher order genome organisation that is largely stable between cell types, but can selectively vary locally, based on the activation or repression of key loci.

This research is describing a novel method to generate realism improved synthetic images while preserving annotation information and the eye gaze direction. Furthermore, it describes how the perceptual loss can be utilized while introducing basic features and techniques from adversarial networks for better results.

Chow Kai Tik = 支援矢量回歸法之影像表示式及其影像處理與分析 / 周啓迪. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 380-383). / Text in English; abstracts in English and Chinese. / Chow Kai Tik = Zhi yuan shi liang hui gui fa zhi ying xiang biao shi shi ji qi ying xiang chu li yu fen xi / Zhou Qidi. / Abstract in English / Abstract in Chinese / Acknowledgement / Content / List of figures / Chapter Chapter 1 --- Introduction --- p.1-11 / Chapter 1.1 --- Introduction --- p.2 / Chapter 1.2 --- Road Map --- p.9 / Chapter Chapter 2 --- Review of Support Vector Machine --- p.12-124 / Chapter 2.1 --- Structural Risk Minimization (SRM) --- p.13 / Chapter 2.1.1 --- Introduction / Chapter 2.1.2 --- Structural Risk Minimization / Chapter 2.2 --- Review of Support Vector Machine --- p.21 / Chapter 2.2.1 --- Review of Support Vector Classification / Chapter 2.2.2 --- Review of Support Vector Regression / Chapter 2.2.3 --- Review of Support Vector Clustering / Chapter 2.2.4 --- Summary of Support Vector Machines / Chapter 2.3 --- Implementation of Support Vector Machines --- p.60 / Chapter 2.3.1 --- Kernel Adatron for Support Vector Classification (KA-SVC) / Chapter 2.3.2 --- Kernel Adatron for Support Vector Regression (KA-SVR) / Chapter 2.3.3 --- Sequential Minimal Optimization for Support Vector Classification (SMO-SVC) / Chapter 2.3.4 --- Sequential Minimal Optimization for Support Vector Regression (SMO-SVR) / Chapter 2.3.5 --- Lagrangian Support Vector Classification (LSVC) / Chapter 2.3.6 --- Lagrangian Support Vector Regression (LSVR) / Chapter 2.4 --- Applications of Support Vector Machines --- p.117 / Chapter 2.4.1 --- Applications of Support Vector Classification / Chapter 2.4.2 --- Applications of Support Vector Regression / Chapter Chapter 3 --- Image Representation by Support Vector Regression --- p.125-183 / Chapter 3.1 --- Introduction of SVR Representation --- p.116 / Chapter 3.1.1 --- Image Representation by SVR / Chapter 3.1.2 --- Implicit Smoothing of SVR representation / Chapter 3.1.3 --- "Different Insensitivity, C value, Kernel and Kernel Parameters" / Chapter 3.2 --- Variation on Encoding Method [Training Process] --- p.154 / Chapter 3.2.1 --- Training SVR with Missing Data / Chapter 3.2.2 --- Training SVR with Image Blocks / Chapter 3.2.3 --- Training SVR with Other Variations / Chapter 3.3 --- Variation on Decoding Method [Testing pr Reconstruction Process] --- p.171 / Chapter 3.3.1 --- Reconstruction with Different Portion of Support Vectors / Chapter 3.3.2 --- Reconstruction with Different Support Vector Locations and Lagrange Multiplier Values / Chapter 3.3.3 --- Reconstruction with Different Kernels / Chapter 3.4 --- Feature Extraction --- p.177 / Chapter 3.4.1 --- Features on Simple Shape / Chapter 3.4.2 --- Invariant of Support Vector Features / Chapter Chapter 4 --- Mathematical and Physical Properties of SYR Representation --- p.184-243 / Chapter 4.1 --- Introduction of RBF Kernel --- p.185 / Chapter 4.2 --- Mathematical Properties: Integral Properties --- p.187 / Chapter 4.2.1 --- Integration of an SVR Image / Chapter 4.2.2 --- Fourier Transform of SVR Image (Hankel Transform of Kernel) / Chapter 4.2.3 --- Cross Correlation between SVR Images / Chapter 4.2.4 --- Convolution of SVR Images / Chapter 4.3 --- Mathematical Properties: Differential Properties --- p.219 / Chapter 4.3.1 --- Review of Differential Geometry / Chapter 4.3.2 --- Gradient of SVR Image / Chapter 4.3.3 --- Laplacian of SVR Image / Chapter 4.4 --- Physical Properties --- p.228 / Chapter 4.4.1 --- 7Transformation between Reconstructed Image and Lagrange Multipliers / Chapter 4.4.2 --- Relation between Original Image and SVR Approximation / Chapter 4.5 --- Appendix --- p.234 / Chapter 4.5.1 --- Hankel Transform for Common Functions / Chapter 4.5.2 --- Hankel Transform for RBF / Chapter 4.5.3 --- Integration of Gaussian / Chapter 4.5.4 --- Chain Rules for Differential Geometry / Chapter 4.5.5 --- Derivation of Gradient of RBF / Chapter 4.5.6 --- Derivation of Laplacian of RBF / Chapter Chapter 5 --- Image Processing in SVR Representation --- p.244-293 / Chapter 5.1 --- Introduction --- p.245 / Chapter 5.2 --- Geometric Transformation --- p.241 / Chapter 5.2.1 --- "Brightness, Contrast and Image Addition" / Chapter 5.2.2 --- Interpolation or Resampling / Chapter 5.2.3 --- Translation and Rotation / Chapter 5.2.4 --- Affine Transformation / Chapter 5.2.5 --- Transformation with Given Optical Flow / Chapter 5.2.6 --- A Brief Summary / Chapter 5.3 --- SVR Image Filtering --- p.261 / Chapter 5.3.1 --- Discrete Filtering in SVR Representation / Chapter 5.3.2 --- Continuous Filtering in SVR Representation / Chapter Chapter 6 --- Image Analysis in SVR Representation --- p.294-370 / Chapter 6.1 --- Contour Extraction --- p.295 / Chapter 6.1.1 --- Contour Tracing by Equi-potential Line [using Gradient] / Chapter 6.1.2 --- Contour Smoothing and Contour Feature Extraction / Chapter 6.2 --- Registration --- p.304 / Chapter 6.2.1 --- Registration using Cross Correlation / Chapter 6.2.2 --- Registration using Phase Correlation [Phase Shift in Fourier Transform] / Chapter 6.2.3 --- Analysis of the Two Methods for Registrationin SVR Domain / Chapter 6.3 --- Segmentation --- p.347 / Chapter 6.3.1 --- Segmentation by Contour Tracing / Chapter 6.3.2 --- Segmentation by Thresholding on Smoothed or Sharpened SVR Image / Chapter 6.3.3 --- Segmentation by Thresholding on SVR Approximation / Chapter 6.4 --- Appendix --- p.368 / Chapter Chapter 7 --- Conclusion --- p.371-379 / Chapter 7.1 --- Conclusion and contribution --- p.372 / Chapter 7.2 --- Future work --- p.378 / Reference --- p.380-383

This thesis investigates the problem of visual learning using a robotic platform. Given a set of objects the robots task is to autonomously manipulate, observe, and learn. This allows the robot to recognise objects in a novel scene and pose, or separate them into distinct visual categories. The main focus of the work is in autonomously acquiring object models using robotic manipulation. Autonomous learning is important for robotic systems. In the context of vision, it allows a robot to adapt to new and uncertain environments, updating its internal model of the world. It also reduces the amount of human supervision needed for building visual models. This leads to machines which can operate in environments with rich and complicated visual information, such as the home or industrial workspace; also, in environments which are potentially hazardous for humans. The hypothesis claims that inducing robot motion on objects aids the learning process. It is shown that extra information from the robot sensors provides enough information to localise an object and distinguish it from the background. Also, that decisive planning allows the object to be separated and observed from a variety of dierent poses, giving a good foundation to build a robust classication model. Contributions include a new segmentation algorithm, a new classication model for object learning, and a method for allowing a robot to supervise its own learning in cluttered and dynamic environments.