Statistical Machine Learning for Automatic Assessment of Physical Activity Intensity Using Multi-axial Accelerometry and Heart Rate

This work explores the automatic recognition of physical activity intensity patterns from multi-axial accelerometry and heart rate signals. Data collection was carried out in free-living conditions and in three controlled gymnasium circuits, for a total amount of 179.80 h of data divided into: sedentary situations (65.5%), light-to-moderate activity (17.6%) and vigorous exercise (16.9%). The proposed machine learning algorithms comprise the following steps: time-domain feature definition, standardization and PCA projection, unsupervised clustering (by k-means and GMM) and a HMM to account for long-term temporal trends. Performance was evaluated by 30 runs of a 10-fold cross-validation. Both k-means and GMM-based approaches yielded high overall accuracy (86.97% and 85.03%, respectively) and, given the imbalance of the dataset, meritorious F-measures (up to 77.88%) for non-sedentary cases. Classification errors tended to be concentrated around transients, what constrains their practical impact. Hence, we consider our proposal to be suitable for 24 h-based monitoring of physical activity in ambulatory scenarios and a first step towards intensity-specific energy expenditure estimators

Regularization for sparsity in statistical analysis and machine learning

Pragmatism is the leading motivation of regularization. We can understand regularization as a modification of the maximum-likelihood estimator so that a reasonable answer could be given in an unstable or ill-posed situation. To mention some typical examples, this happens when fitting parametric or non-parametric models with more parameters than data or when estimating large covariance matrices. Regularization is usually used, in addition, to improve the bias-variance tradeoff of an estimation. Then, the definition of regularization is quite general, and, although the introduction of a penalty is probably the most popular type, it is just one out of multiple forms of regularization. In this dissertation, we focus on the applications of regularization for obtaining sparse or parsimonious representations, where only a subset of the inputs is used. A particular form of regularization, L1-regularization, plays a key role for reaching sparsity. Most of the contributions presented here revolve around L1-regularization, although other forms of regularization are explored (also pursuing sparsity in some sense). In addition to present a compact review of L1-regularization and its applications in statistical and machine learning, we devise methodology for regression, supervised classification and structure induction of graphical models. Within the regression paradigm, we focus on kernel smoothing learning, proposing techniques for kernel design that are suitable for high dimensional settings and sparse regression functions. We also present an application of regularized regression techniques for modeling the response of biological neurons. Supervised classification advances deal, on the one hand, with the application of regularization for obtaining a na¨ıve Bayes classifier and, on the other hand, with a novel algorithm for brain-computer interface design that uses group regularization in an efficient manner. Finally, we present a heuristic for inducing structures of Gaussian Bayesian networks using L1-regularization as a filter. El pragmatismo es la principal motivación de la regularización. Podemos entender la regularización como una modificación del estimador de máxima verosimilitud, de tal manera que se pueda dar una respuesta cuando la configuración del problema es inestable. A modo de ejemplo, podemos mencionar el ajuste de modelos paramétricos o no paramétricos cuando hay más parámetros que casos en el conjunto de datos, o la estimación de grandes matrices de covarianzas. Se suele recurrir a la regularización, además, para mejorar el compromiso sesgo-varianza en una estimación. Por tanto, la definición de regularización es muy general y, aunque la introducción de una función de penalización es probablemente el método más popular, éste es sólo uno de entre varias posibilidades. En esta tesis se ha trabajado en aplicaciones de regularización para obtener representaciones dispersas, donde sólo se usa un subconjunto de las entradas. En particular, la regularización L1 juega un papel clave en la búsqueda de dicha dispersión. La mayor parte de las contribuciones presentadas en la tesis giran alrededor de la regularización L1, aunque también se exploran otras formas de regularización (que igualmente persiguen un modelo disperso). Además de presentar una revisión de la regularización L1 y sus aplicaciones en estadística y aprendizaje de máquina, se ha desarrollado metodología para regresión, clasificación supervisada y aprendizaje de estructura en modelos gráficos. Dentro de la regresión, se ha trabajado principalmente en métodos de regresión local, proponiendo técnicas de diseño del kernel que sean adecuadas a configuraciones de alta dimensionalidad y funciones de regresión dispersas. También se presenta una aplicación de las técnicas de regresión regularizada para modelar la respuesta de neuronas reales. Los avances en clasificación supervisada tratan, por una parte, con el uso de regularización para obtener un clasificador naive Bayes y, por otra parte, con el desarrollo de un algoritmo que usa regularización por grupos de una manera eficiente y que se ha aplicado al diseño de interfaces cerebromáquina. Finalmente, se presenta una heurística para inducir la estructura de redes Bayesianas Gaussianas usando regularización L1 a modo de filtro.

The theory of on-line learning: a statistical physics approach

In this paper we review recent theoretical approaches for analysing the dynamics of on-line learning in multilayer neural networks using methods adopted from statistical physics. The analysis is based on monitoring a set of macroscopic variables from which the generalisation error can be calculated. A closed set of dynamical equations for the macroscopic variables is derived analytically and solved numerically. The theoretical framework is then employed for defining optimal learning parameters and for analysing the incorporation of second order information into the learning process using natural gradient descent and matrix-momentum based methods. We will also briefly explain an extension of the original framework for analysing the case where training examples are sampled with repetition.

Semi-supervised learning of statistical models for natural language understanding

Natural language understanding is to specify a computational model that maps sentences to their semantic mean representation. In this paper, we propose a novel framework to train the statistical models without using expensive fully annotated data. In particular, the input of our framework is a set of sentences labeled with abstract semantic annotations. These annotations encode the underlying embedded semantic structural relations without explicit word/semantic tag alignment. The proposed framework can automatically induce derivation rules that map sentences to their semantic meaning representations. The learning framework is applied on two statistical models, the conditional random fields (CRFs) and the hidden Markov support vector machines (HM-SVMs). Our experimental results on the DARPA communicator data show that both CRFs and HM-SVMs outperform the baseline approach, previously proposed hidden vector state (HVS) model which is also trained on abstract semantic annotations. In addition, the proposed framework shows superior performance than two other baseline approaches, a hybrid framework combining HVS and HM-SVMs and discriminative training of HVS, with a relative error reduction rate of about 25% and 15% being achieved in F-measure.

E-learning Skill and Use in EU Countries – A Statistical Analysis

The article investigates the division between member states of the European Union considering the aspect of their level of information and communication technology (ICT) development focusing on e-learning. With the help of discriminant analysis the countries are categorized into groups based on their ICT maturity and e-learning literacy level of development. Making a comparison with a benchmarking tool, the ITU (International Telecommunication Union)’s ICT Development Index (IDI) the results are confirmed partly correct. The article tries to find economical explanations for the re-grouping of the countries ranking. Finally the author examines the reliability of Hungary’s ranking results and the factors which may affect this divergence from the real picture.

New Advancements of Scalable Statistical Methods for Learning Latent Structures in Big Data

Constant technology advances have caused data explosion in recent years. Accord- ingly modern statistical and machine learning methods must be adapted to deal with complex and heterogeneous data types. This phenomenon is particularly true for an- alyzing biological data. For example DNA sequence data can be viewed as categorical variables with each nucleotide taking four different categories. The gene expression data, depending on the quantitative technology, could be continuous numbers or counts. With the advancement of high-throughput technology, the abundance of such data becomes unprecedentedly rich. Therefore efficient statistical approaches are crucial in this big data era.

Previous statistical methods for big data often aim to find low dimensional struc- tures in the observed data. For example in a factor analysis model a latent Gaussian distributed multivariate vector is assumed. With this assumption a factor model produces a low rank estimation of the covariance of the observed variables. Another example is the latent Dirichlet allocation model for documents. The mixture pro- portions of topics, represented by a Dirichlet distributed variable, is assumed. This dissertation proposes several novel extensions to the previous statistical methods that are developed to address challenges in big data. Those novel methods are applied in multiple real world applications including construction of condition specific gene co-expression networks, estimating shared topics among newsgroups, analysis of pro- moter sequences, analysis of political-economics risk data and estimating population structure from genotype data.

Evaluating statistical and machine learning methods to predict risk of in-hospital child mortality in Uganda

Thesis (Master's)--University of Washington, 2016-08

Convex Optimization Algorithms and Statistical Bounds for Learning Structured Models

Thesis (Ph.D.)--University of Washington, 2016-08

Assessment of Preclinical Learning on Oral Surgery Using Three Instructional Strategies

The purpose of this investigation was to evaluate three learning methods for teaching basic oral surgical skills Thirty predoctoral dental students without any surgical knowledge or previous surgical experience were divided Into three groups (n=10 each) according to instructional strategy Group 1, active learning Group 2, text reading only, and Group 3, text reading and video demonstration After instruction, the apprentices were allowed to practice incision dissection and suture maneuvers in a bench learning model During the students' performance, a structured practice evaluation test to account for correct or incorrect maneuvers was applied by trained observers Evaluation tests were repeated after thirty and sixty days Data from resulting scores between groups and periods were considered for statistical analysis (ANOVA and Tukey Kramer) with a significant level of a=0 05 Results showed that the active learning group presented the significantly best learning outcomes related to immediate assimilation of surgical procedures compared to other groups All groups results were similar after sixty days of the first practice Assessment tests were fundamental to evaluate teaching strategies and allowed theoretical and proficiency learning feedbacks Repetition and interactive practice promoted retention of knowledge on basic oral surgical skills

Visualization of learning in multilayer perceptron networks using principal component analysis

This paper is concerned with the use of scientific visualization methods for the analysis of feedforward neural networks (NNs). Inevitably, the kinds of data associated with the design and implementation of neural networks are of very high dimensionality, presenting a major challenge for visualization. A method is described using the well-known statistical technique of principal component analysis (PCA). This is found to be an effective and useful method of visualizing the learning trajectories of many learning algorithms such as back-propagation and can also be used to provide insight into the learning process and the nature of the error surface.

Spatial learning ability of rats following differing levels of exposure to alcohol during early postnatal life

Rats exposed to a relatively high dose (7.5 g/kg body weight) of alcohol on either the fifth or tenth postnatal day of age have been reported to have long-lasting deficits in spatial learning ability as tested on the Morris water maze task. The question arises concerning the level of alcohol required to achieve this effect. Wistar rats were exposed to either 2, 4 or 6 g/kg body weight of ethanol administered as a 10% solution. This ethanol was given over an 8-h period on the fifth postnatal day of age by means of an intragastric cannula. Gastrostomy controls received a 5% sucrose solution substituted isocalorically for the ethanol. Another set of pups raised by their mother were used as suckle controls. All surgical procedures were carried out under halothane vapour anaesthesia. After the artificial feeding regimes all pups were returned to lactating dams and weaned at 21 days of age. The spatial learning ability of these rats was tested in the Morris water maze when they were between 61-64 days of age. This task requires the rats to swim in a pool containing water made opaque and locate and climb onto a submerged platform. The time taken to accomplish this is known as the escape latency. Each rat was subjected to 24 trials over 3 days of the test period. Statistical analysis of the escape latency data revealed that the rats given 6 g/kg body weight of ethanol had significant deficits in their spatial learning ability compared with their control groups. However, there was no significant difference in spatial learning ability for the rats given either 2 or 4 g/kg body weight of ethanol compared with their respective gastrostomy or suckle control animals. We concluded that ethanol exposure greater than 4 g/kg over an 8-h period to 5-day-old rats is required for them to develop long-term deficits in spatial learning behaviour. (C) 1998 Elsevier Science Inc.

Input-driven Language Learning

Input-driven models provide an explicit and readily testable account of language learning. Although we share Ellis's view that the statistical structure of the linguistic environment is a crucial and, until recently, relatively neglected variable in language learning, we also recognize that the approach makes three assumptions about cognition and language learning that are not universally shared. The three assumptions concern (a) the language learner as an intuitive statistician, (b) the constraints on what constitute relevant surface cues, and (c) the redescription problem faced by any system that seeks to derive abstract grammatical relations from the frequency of co-occurring surface forms and functions. These are significant assumptions that must be established if input-driven models are to gain wider acceptance. We comment on these issues and briefly describe a distributed, instance-based approach that retains the key features of the input-driven account advocated by Ellis but that also addresses shortcomings of the current approaches.

Perceptions of kanji learning strategies: Do they differ among Chinese character and alphabetic background learners

This study investigates three important issues in kanji learning strategies; namely, strategy use, effectiveness of strategy and orthographic background. A questionnaire on kanji learning strategy use and perceived effectiveness was administered to 116 beginner level, undergraduate students of Japanese from alphabetic and character backgrounds in Australia. Both descriptive and statistical analyses of the questionnaire responses revealed that the strategies used most often are the most helpful. Repeated writing was reported as the most used strategy type although alphabetic background learners reported using repeated writing strategies significantly more often than character background learners. The importance of strategy training and explicit instruction of fundamental differences between character and alphabetic background learners of Japanese is discussed in relation to teaching strategies. [Author abstract]

Difficulties in the Learning of Statistics by the Students of the Primary Education

The present study, covering students from public schools and a private school on the island of São Miguel (Azores, Portugal), aims to meet the difficulties of the students of the 3rd and 4th years of the primary education in solving tasks involving construction, reading and interpreting tables and statistical graphs, in the context of Organization and Data Handling (ODH). We present the main results obtained from statistical methods, among which we highlight some non-parametric hypothesis tests and the Categorical Principal Component Analysis (CatPCA), given the nature of the variables included in the questionnaire (mostly nominal and ordinal variables).

The impact of firm-promoted accreditation of prior learning processes on the worker-organization relationship

The process of Competences Recognition, Validation and Certification , also known as Accreditation of Prior Learning (APL), is an innovative means of attaining school certificates for individuals without an academic background. The main objective of this process is to validate what people have learned in informal contexts, in order to attribute academic certificates. With the increasing interest of the qualification of workers and governmental support, more and more Portuguese organizations promote this process within their facilities and their work hours. This study explores the relationship between the promotion of this Human Resource Development Programme and employee’s attitudes (Job Satisfaction and Organizational Commitment) and behaviours (Extra-role Organizational Citizenship Behaviours) towards the organization they work for. Results of a cross-sectional survey of Portuguese Industrial Workers (N=135) showed that statistical significant results are in the higher levels of Voice Behaviours (a dimension of Extra-role Organizational Citizenship Behaviour in the groups of workers who were involved or had graduated from the firm promoted APL process.