992 research outputs found
Sequential Feature Classification in the Context of Redundancies
No full textPfannschmidt L, Hammer B. Sequential Feature Classification in the Context of Redundancies. Draft
Limitations and possibilities in painting: tracing the unrepresentable
No full textM.F.A.Includes bibliographical references (p. 13)by Viktor Lukas Witkowsk
Relevance learning for redundant features
No full textPfannschmidt L. Relevance learning for redundant features. Bielefeld: Universität Bielefeld; 2021.Feature selection is a widely used strategy in machine learning
for the reduction of feature sets to their relevant essence to
improve predictions and performance. It is also employed for
knowledge discovery in applied disciplines such as biology and
medicine to find potentially causal factors. But machine learning
models often do not represent a unique solution to a given problem,
especially in high dimensional settings where redundant
factors are likely and spurious correlations exist.
Basing decisions about causal elements on feature selection is
therefore inaccurate or wrong when not considering the presence
of redundant but also relevant features. Most existing selection
algorithms are specifically removing redundancies and not suitable
for the task of all-relevant feature selection, or they require
careful parametrization and are hard to interpret, which makes
them difficult to use.
This thesis is focused on feature selection methods for the
analytical use case to facilitate understanding of potential causal
factors, for linear and non-linear problems. We propose several
new algorithms and methods for all-relevant feature selection
to improve knowledge discovery, enabled by statistical methods
to improve the accuracy of existing solutions and allow the differentiation
between different types of relevance. Furthermore,
we offer a new heuristic to automatically group related features
together, and we analyse the definition of relevance in the context
of privileged information, where data is only available in
training.
We also introduce software implementations, which were
specifically designed to be modular, efficient and able to parallelize
for applications in high dimensional problems. The methods
and implementations were evaluated on a wide range of
synthetic and real datasets to show their performance in comparison
with existing algorithms
Economics of credence goods : the effect of monitoring and its costs in the health care market
No full textauthor: Lukas SchönerMasterarbeit University of Innsbruck 201
Economics of credence goods : the effect of monitoring and its costs in the health care market
No full textauthor: Lukas SchönerMasterarbeit University of Innsbruck 201
Tailoring comonomer incorporation in Ziegler-Natta based Linear Low Density Polyethylene
No full textAuthor DI Lukas GöpperlDissertation Johannes Kepler Universität Linz 202
Time-bin entangled photons from a quantum dot embedded in a nanowire
No full textauthor: Lukas Kirchner, BSc.Zusammenfassung in deutscher SpracheMasterarbeit University of Innsbruck 201
Time-bin entangled photons from a quantum dot embedded in a nanowire
No full textauthor: Lukas Kirchner, BSc.Zusammenfassung in deutscher SpracheMasterarbeit University of Innsbruck 201
Non-tax revenue and subnational democracy: evidence from Colombia
No full textAn influential group of political economists posits a negative relationship between the relative size of non-tax revenues in public balance sheets and the level of democracy in a given county. While empirical evidence for this proposition is largely based on cross-national studies, scholars have largely neglected subnational contexts as important domains of research. Addressing this disjuncture, the author studies the electoral impact of two important subnational non-tax revenues - natural resource royalties and central-government fiscal transfers - at the municipal level in Colombia. Employing a propensity score matching approach to attenuate problems of omitted variable bias, a quantitative analysis of Colombia's 1119 municipalities shows that higher levels of fiscal transfers and petroleum royalties had no discernible impact on the average level of competitiveness in the 2007 and 2011 municipal elections. If anything, there seems to be a positive impact of non-tax revenues on electoral contestation. Supported by qualitative evidence from theoretical outliers in the set of observations, the results suggest that non-tax revenue promotes electoral competition by raising the stakes of attaining political office. At the same time, abundance in fiscal transfers and resource royalties may undermine democratic governance through means that are not reflected in electoral margins.M.A.Includes bibliographical referencesIncludes vitaby Lukas Konstantin Kelle
Time-bin entangled photons from a quantum dot embedded in a nanowire
No full textauthor: Lukas Kirchner, BSc.Zusammenfassung in deutscher SpracheMasterarbeit University of Innsbruck 201
- …
