Evgueni N. Smirnov

The ROC isometrics approach to construct reliable classifiers

We address the problem of applying machine-learning classifiers in domains where incorrect classifications have severe consequences. In these domains we propose to apply classifiers only when their performance can be defined by the domain expert prior to classification. The classifiers so obtained are called reliable classifiers. In the article we present three main contributions. First, we establish the effect on an ROC curve when ambiguous instances are left unclassified. Second, we propose the ROC isometrics approach to tune and transform a classifier in such a way that it becomes reliable. Third, we provide an empirical evaluation of the approach. From our analysis and experimental evaluation we may conclude that the ROC isometrics approach is an effective and efficient approach to construct reliable classifiers. In addition, a discussion about related work clearly shows the benefits of the approach when compared with existing approaches that also have the option to leave ambiguous instances unclassified.

Theoretical and Experimental Study of a Meta-Typicalness Approach for Reliable Classification

The authors propose a meta-typicalness approach to apply the typicalness framework for any type of classifiers. The approach can be used to construct classifiers with specified classification performance. Experiments show that the approach results in classifiers that can outperform an existing typicalness-based classifier

Inexpensive user tracking using Boltzmann machines

Inexpensive user tracking is an important problem in various application domains such as healthcare, human-computer interaction, energy savings, safety, robotics, security and so on. Yet, it cannot be easily solved due to its probabilistic nature, high level of abstraction and uncertainties, on the one side, and to the limitations of our current technologies and learning algorithms, on the other side. In this paper, we tackle this problem by using the Multi-integrated Sensor Technology, which comes at a low price. At the same time, we are aiming to address the lightweight learning requirements by investigating Factored Conditional Restricted Boltzmann Machines (FCRBMs), a form of Deep Learning, that has proven to be an efficient and effective machine learning framework. However, due to their construction properties, the conventional FCRBMs are only capable of performing predictions but are not capable of making classification. Herein, we are proposing extended FCRBMs (eFCRBMs), which incorporate a novel classification scheme, to solve this problem. Experiments performed on both artificially generated as well as real-world data demonstrate the effectiveness and efficiency of the proposed technique. We show that eFCRBMs outperform popular approaches including Support Vector Machines, Naive Bayes, AdaBoost, and Gaussian Mixture Models.

Reliable Knowledge Discovery

Reliable Knowledge Discovery focuses on theory, methods, and techniques for RKDD, a new sub-field of KDD. It studies the theory and methods to assure the reliability and trustworthiness of discovered knowledge and to maintain the stability and consistency of knowledge discovery processes. RKDD has a broad spectrum of applications, especially in critical domains like medicine, finance, and military. Reliable Knowledge Discovery also presents methods and techniques for designing robust knowledge-discovery processes. Approaches to assessing the reliability of the discovered knowledge are introduced. Particular attention is paid to methods for reliable feature selection, reliable graph discovery, reliable classification, and stream mining. Estimating the data trustworthiness is covered in this volume as well. Case studies are provided in many chapters. Reliable Knowledge Discovery is designed for researchers and advanced-level students focused on computer science and electrical engineering as a secondary text or reference. Professionals working in this related field and KDD application developers will also find this book useful.

Meta-conformity approach to reliable classification

The conformity framework has recently been proposed for the task of reliable classification. Given a classifier B, the framework allows to obtain p-values of the classifications assigned to individual instances. However, applying the framework is a difficult problem: we need to construct an instance non-conformity function for the classifier B. To avoid constructing such a function we propose a meta-conformity approach. If a conformity-based classifier M is available, the approach is to train M as a meta classifier that predicts the correctness of each classification of the classifier B. In this way the classification p-values of the classifier B are represented by the classification p-values of the classifier M. The meta-conformity approach can be used for constructing classifiers with predefined generalization performance. Experiments show that the approach results in classifiers that can outperform existing conformity-based classifiers.

Nonlinear filtering of asymmetric stochastic volatility models and Value-at-Risk estimation

This paper develops an efficient approach to analytical learning of Asymmetric Stochastic Volatility (ASV) models through nonlinear filtering, and shows that they are useful for practical risk management. This involves the derivation of a Nonlinear Quadrature Filter (NQF) that operates directly on the nonlinear ASV model. The NQF filter makes Gaussian approximations to the prior and posterior density of the latent volatility, but not in the observation space which makes possible easy handling of heavy-tailed data. Experiments in Value-at-Risk (VaR) assessment via an original bootsrtapping methodology are conducted with NQF and several ASV learning algorithms. The results indicate that our approach yields models with better statistical characteristics than the considered competitors, and slightly improved VaR forecasts.

Applying Preference Biases to Conjunctive and Disjunctive Version Spaces

The paper considers conjunctive and disjunctive version space learning as an incomplete search in complete hypotheses spaces. The incomplete search is guided by preference biases which are implemented by procedures based on the instance-based boundary sets representation of version spaces. The conditions for tractability of this representation are defined. As a result we propose to use instance-based boundary sets as a basis for the computationally feasible application of preference biases to version spaces.

A Bayesian analysis for the bivariate geometric distribution in the presence of covariates and censored data

Abstract In this paper, we introduce a Bayesian analysis for bivariate geometric distributions applied to lifetime data in the presence of covariates and censored data using Markov Chain Monte Carlo (MCMC) methods. We show that the use of a discrete bivariate geometric distribution could bring us some computational advantages when compared to standard existing bivariate exponential lifetime distributions introduced in the literature assuming continuous lifetime data as for example, the exponential Block and Basu bivariate distribution. Posterior summaries of interest are obtained using the popular OpenBUGS software. A numerical illustration is introduced considering a medical data set related to the recurrence times of infection for kidney patients.

Master Orientation Tool

This paper describes our ongoing work on developing a Master Orientation Tool for University College Maastricht (UCM). UCM Bachelor students use the tool to discover Master programs that fit their academic profiles. The tool includes a memory-based collaborative recommender system. The system memory contains data on academic profiles of UCM alumni students, labeled by the Master programs they have chosen. The tool operates as a collaborative system: given the academic profile of a Bachelor student, it recommends Master programs for that student based on the proximity of her profile to the profiles of the alumni. The Master Orientation Tool allows students to modify their own profiles and thus to explore alternatives in their study and how they influence their Master program possibilities. The tool is operational at UCM since September 2011 and is popular among the students.

Efficient AdaBoost Region Classification

The task of region classification is to construct class regions containing the correct classes of the objects being classified with an error probability *** *** [0,1]. To turn a point classifier into a region classifier, the conformal framework is employed [11,14]. However, to apply the framework we need to design a non-conformity function. This function has to estimate the instances non-conformity for the point classifier used. This paper introduces a new non-conformity function for AdaBoost. The function has two main advantages over the only existing non-conformity function for AdaBoost. First, it reduces the time complexity of computing class regions with a factor equal to the size of the training data. Second, it results in statistically better class regions.