Sökning: "Feature Selection"
Visar resultat 21 - 25 av 116 avhandlingar innehållade orden Feature Selection.
21. On the Selection of Methods and Tools for Analysis of Heat and Power Plants
Sammanfattning : The business climate of the energy market has been eventful over the last two decades, particularly with the introduction of the deregulated electricity market and promotion of environmental issues. This has subsequently lead to new analysis criteria, influencing both technology requirements and production costs. LÄS MER
22. Measurement accuracy. Applications in X-ray computed tomography (CT) dose reduction and magnetic resonance spectroscopy (MRS) volume selection
Sammanfattning : X-ray computed tomography (CT) images of transaxial cross sections of the body can be used to determine tissue volumes. However, the considerable radiation dose associated with high-quality clinical images has greatly limited the use of CT for this application. LÄS MER
23. Entrepreneurial Dynamics and Acquisitions of New Technology-Based Firms
Sammanfattning : The aim of this thesis is to explore the evolution of new technology-based firms (NTBFs) based on solid empirical material for the Swedish economy. NTBFs are argued to be one important type of entrepreneurial firms that contribute to technology change, knowledge spillovers, and economic growth. LÄS MER
24. Embedded in a context : the adaptation of immigrant youth
Sammanfattning : With rising levels of immigration comes a need to know what fosters positive adaptation for the youth growing up in a new culture of settlement.The issue is increasingly studied; however, little of the research conducted has combined a developmental with a contextual approach. LÄS MER
25. A study on selecting and optimizing perceptually relevant features for automatic speech recognition
Sammanfattning : The performance of an automatic speech recognition (ASR) system strongly depends on the representation used for the front-end. If the extracted features do not include all relevant information, the performance of the classification stage is inherently suboptimal. LÄS MER