Search Results (9)
Feature construction techniques in time-series analysis and single-objective optimization
Feature construction, encompassing both feature engineering, which involves the manual design of features by domain experts, and representation learning, which refers to the automated discovery of useful data representations during model construction, is a fundamental aspect of machine learning. Its goal is to transform raw data into a more suitable …
Evolutionary algorithms for optimizing pessimistic bilevel problems and min-max problems
This thesis investigates two distinct but related optimization problems: bilevel and minmax problems, in the context of evolutionary algorithms (EAs). Bilevel optimization involves hierarchical decision-making, where decisions at the upper level are subject to constraints defined by the solutions of an optimization problem at the lower level. These problems are …
Use of information and communication technologies, data and knowledge to increase the impact of digital environments on food choice
Food and eating environments play a crucial role in shaping consumers' food choices. As food decision-making shifts more into the digital environment, it is essential to understand the impact of this setting on consumer's dietary behaviours. Online platforms and mobile apps provide great opportunities for the promotion of healthier food …
Study of inconsistencies in health risk assessment: the role of assessment context in decision analysis aimed at their reduction
The area of risk assessment continues to be challenged by foundational issues that limit its potential to inform public health decisions. This doctoral work improves our understanding of the interaction between science and public and environmental health policy by addressing specific challenges in risk assessment within and for decision-making. The …
Considering autocorrelation in predictive models
Most machine learning, data mining and statistical methods rely on the assumption that the analyzed data are independent and identically distributed (i.i.d.). More specifically, the individual examples included in the training data are assumed to be drawn independently from each other from the same probability distribution. However, cases where this …
Algorithms for Learning Regression Trees and Ensembles on Evolving Data Streams
In this thesis we address the problem of learning various types of decision trees from timechanging data streams. In particular, we study online machine learning algorithms for learning regression trees, linear model trees, option trees for regression, multi-target model trees, and ensembles of model trees from data streams. These are …
Searching for Credible Relations in Machine Learning
Can a model constructed by machine learning or data mining programs be trusted? For example, it is known that a decision tree model can contain less-credible parts caused by pathologies in induction algorithms, noise and missing values in data, or simply because of the complexity of a domain. Such models …
Ensembles for predicting structured outputs
In this thesis, we address the task of learning models for predicting structured outputs, which take as input a tuple of attribute values and produce as output a structured object. In contrast to classification and regression, where the output is a single scalar value, in our case the output is …
Hydraulic load cells for measurement of small loads
This work deals with the design of a small kitchen appliance with a weighing function integrated into a control loop in order to prevent unbalanced movement of the appliance during operation. In this particular case, the weighing should be appropriate for larger amounts (up to 5 kg) of ingredients. The …