J48Consolidated: an implementation of CTC algorithm for WEKA

The CTC algorithm, Consolidated Tree Construction algorithm, is a machine learning paradigm that was designed to solve a class imbalance problem, a fraud detection problem in the area of car insurance [1] where, besides, an explanation about the classification made was required. The algorithm is based on a decision tree construction algorithm, in this case the well-known C4.5, but it extracts knowledge from data using a set of samples instead of a single one as C4.5 does. In contrast to other methodologies based on several samples to build a classifier, such as bagging, the CTC builds a single tree and as a consequence, it obtains comprehensible classifiers. The main motivation of this implementation is to make public and available an implementation of the CTC algorithm. With this purpose we have implemented the algorithm within the well-known WEKA data mining environment http://www.cs.waikato.ac.nz/ml/weka/). WEKA is an open source project that contains a collection of machine learning algorithms written in Java for data mining tasks. J48 is the implementation of C4.5 algorithm within the WEKA package. We called J48Consolidated to the implementation of CTC algorithm based on the J48 Java class.

Analysis, optimization and development of an answer scoring system

The main contribution of this work is to analyze and describe the state of the art performance as regards answer scoring systems from the SemEval- 2013 task, as well as to continue with the development of an answer scoring system (EHU-ALM) developed in the University of the Basque Country. On the overall this master thesis focuses on finding any possible configuration that lets improve the results in the SemEval dataset by using attribute engineering techniques in order to find optimal feature subsets, along with trying different hierarchical configurations in order to analyze its performance against the traditional one versus all approach. Altogether, throughout the work we propose two alternative strategies: on the one hand, to improve the EHU-ALM system without changing the architecture, and, on the other hand, to improve the system adapting it to an hierarchical con- figuration. To build such new models we describe and use distinct attribute engineering, data preprocessing, and machine learning techniques.

Mapeamento de difusão no reconhecimento e reconstrução de sinais

Em muitas representações de objetos ou sistemas físicos se faz necessário a utilização de técnicas de redução de dimensionalidade que possibilitam a análise dos dados em baixas dimensões, capturando os parâmetros essenciais associados ao problema. No contexto de aprendizagem de máquina esta redução se destina primordialmente à clusterização, reconhecimento e reconstrução de sinais. Esta tese faz uma análise meticulosa destes tópicos e suas conexões que se encontram em verdadeira ebulição na literatura, sendo o mapeamento de difusão o foco principal deste trabalho. Tal método é construído a partir de um grafo onde os vértices são os sinais (dados do problema) e o peso das arestas é estabelecido a partir do núcleo gaussiano da equação do calor. Além disso, um processo de Markov é estabelecido o que permite a visualização do problema em diferentes escalas conforme variação de um determinado parâmetro t: Um outro parâmetro de escala, Є, para o núcleo gaussiano é avaliado com cuidado relacionando-o com a dinâmica de Markov de forma a poder aprender a variedade que eventualmente seja o suporte do dados. Nesta tese é proposto o reconhecimento de imagens digitais envolvendo transformações de rotação e variação de iluminação. Também o problema da reconstrução de sinais é atacado com a proposta de pré-imagem utilizando-se da otimização de uma função custo com um parâmetro regularizador, γ, que leva em conta também o conjunto de dados iniciais.

Síntese de árvores de padrões Fuzzy através de Programação Genética Cartesiana.

Esta dissertação apresenta um sistema de indução de classificadores fuzzy. Ao invés de utilizar a abordagem tradicional de sistemas fuzzy baseados em regras, foi utilizado o modelo de Árvore de Padrões Fuzzy(APF), que é um modelo hierárquico, com uma estrutura baseada em árvores que possuem como nós internos operadores lógicos fuzzy e as folhas são compostas pela associação de termos fuzzy com os atributos de entrada. O classificador foi obtido sintetizando uma árvore para cada classe, esta árvore será uma descrição lógica da classe o que permite analisar e interpretar como é feita a classificação. O método de aprendizado originalmente concebido para a APF foi substituído pela Programação Genética Cartesiana com o intuito de explorar melhor o espaço de busca. O classificador APF foi comparado com as Máquinas de Vetores de Suporte, K-Vizinhos mais próximos, florestas aleatórias e outros métodos Fuzzy-Genéticos em diversas bases de dados do UCI Machine Learning Repository e observou-se que o classificador APF apresenta resultados competitivos. Ele também foi comparado com o método de aprendizado original e obteve resultados comparáveis com árvores mais compactas e com um menor número de avaliações.

KLAIR: A virtual infant for spoken language acquisition research

Recent research into the acquisition of spoken language has stressed the importance of learning through embodied linguistic interaction with caregivers rather than through passive observation. However the necessity of interaction makes experimental work into the simulation of infant speech acquisition difficult because of the technical complexity of building real-time embodied systems. In this paper we present KLAIR: a software toolkit for building simulations of spoken language acquisition through interactions with a virtual infant. The main part of KLAIR is a sensori-motor server that supplies a client machine learning application with a virtual infant on screen that can see, hear and speak. By encapsulating the real-time complexities of audio and video processing within a server that will run on a modern PC, we hope that KLAIR will encourage and facilitate more experimental research into spoken language acquisition through interaction. Copyright © 2009 ISCA.

Gaussian process change point models

We combine Bayesian online change point detection with Gaussian processes to create a nonparametric time series model which can handle change points. The model can be used to locate change points in an online manner; and, unlike other Bayesian online change point detection algorithms, is applicable when temporal correlations in a regime are expected. We show three variations on how to apply Gaussian processes in the change point context, each with their own advantages. We present methods to reduce the computational burden of these models and demonstrate it on several real world data sets. Copyright 2010 by the author(s)/owner(s).

Fast L0-based sparse signal recovery

This paper develops an algorithm for finding sparse signals from limited observations of a linear system. We assume an adaptive Gaussian model for sparse signals. This model results in a least square problem with an iteratively reweighted L2 penalty that approximates the L0-norm. We propose a fast algorithm to solve the problem within a continuation framework. In our examples, we show that the correct sparsity map and sparsity level are gradually learnt during the iterations even when the number of observations is reduced, or when observation noise is present. In addition, with the help of sophisticated interscale signal models, the algorithm is able to recover signals to a better accuracy and with reduced number of observations than typical L1-norm and reweighted L1 norm methods. ©2010 IEEE.

Construction of nonparametric Bayesian models from parametric Bayes equations

We consider the general problem of constructing nonparametric Bayesian models on infinite-dimensional random objects, such as functions, infinite graphs or infinite permutations. The problem has generated much interest in machine learning, where it is treated heuristically, but has not been studied in full generality in non-parametric Bayesian statistics, which tends to focus on models over probability distributions. Our approach applies a standard tool of stochastic process theory, the construction of stochastic processes from their finite-dimensional marginal distributions. The main contribution of the paper is a generalization of the classic Kolmogorov extension theorem to conditional probabilities. This extension allows a rigorous construction of nonparametric Bayesian models from systems of finite-dimensional, parametric Bayes equations. Using this approach, we show (i) how existence of a conjugate posterior for the nonparametric model can be guaranteed by choosing conjugate finite-dimensional models in the construction, (ii) how the mapping to the posterior parameters of the nonparametric model can be explicitly determined, and (iii) that the construction of conjugate models in essence requires the finite-dimensional models to be in the exponential family. As an application of our constructive framework, we derive a model on infinite permutations, the nonparametric Bayesian analogue of a model recently proposed for the analysis of rank data.

An Infinite Latent Attribute Model for Network Data

Latent variable models for network data extract a summary of the relational structure underlying an observed network. The simplest possible models subdivide nodes of the network into clusters; the probability of a link between any two nodes then depends only on their cluster assignment. Currently available models can be classified by whether clusters are disjoint or are allowed to overlap. These models can explain a "flat" clustering structure. Hierarchical Bayesian models provide a natural approach to capture more complex dependencies. We propose a model in which objects are characterised by a latent feature vector. Each feature is itself partitioned into disjoint groups (subclusters), corresponding to a second layer of hierarchy. In experimental comparisons, the model achieves significantly improved predictive performance on social and biological link prediction tasks. The results indicate that models with a single layer hierarchy over-simplify real networks.