• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 12
  • 3
  • 2
  • 1
  • Tagged with
  • 19
  • 19
  • 19
  • 4
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 3
  • 3
  • 3
  • 3
  • 3
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Topics in the foundations of statistical inference and statistical mechanics /

Guszcza, James. January 2000 (has links)
Thesis (Ph. D.)--University of Chicago, Dept. of Philosophy. / Includes bibliographical references. Also available on the Internet.
2

Entropy based techniques with applications in data mining

Okafor, Anthony. January 2005 (has links)
Thesis (Ph. D.)--University of Florida, 2005. / Title from title page of source document. Document formatted into pages; contains 97 pages. Includes vita. Includes bibliographical references.
3

The Genetic Algorithm and Maximum Entropy Dice

Fellman, Laura Suzanne 29 January 1996 (has links)
The Brandeis dice problem, originally introduced in 1962 by Jaynes as an illustration of the principle of maximum entropy, was solved using the genetic algorithm, and the resulting solution was compared with that obtained analytically. The effect of varying the genetic algorithm parameters was observed, and the optimum values for population size, mutation rate, and mutation interval were determined for this problem. The optimum genetic algorithm program was then compared to a completely random method of search and optimization. Finally, the genetic algorithm approach was extended to several variations of the original problem for which an analytical approach would be impractical.
4

Generative models of similarity-based classification /

Cazzanti, Luca. January 2007 (has links)
Thesis (Ph. D.)--University of Washington, 2007. / Vita. Includes bibliographical references (p. 101-107).
5

Sentence Compression by Removing Recursive Structure from Parse Tree

Matsubara, Shigeki, Kato, Yoshihide, Egawa, Seiji 04 December 2008 (has links)
PRICAI 2008: Trends in Artificial Intelligence 10th Pacific Rim International Conference on Artificial Intelligence, Hanoi, Vietnam, December 15-19, 2008. Proceedings
6

Improving shared weight neural networks generalization using regularization theory and entropy maximization /

Khabou, Mohamed Ali, January 1999 (has links)
Thesis (Ph. D.)--University of Missouri-Columbia, 1999. / Typescript. Vita. Includes bibliographical references (leaves 114-121). Also available on the Internet.
7

Improving shared weight neural networks generalization using regularization theory and entropy maximization

Khabou, Mohamed Ali, January 1999 (has links)
Thesis (Ph. D.)--University of Missouri-Columbia, 1999. / Typescript. Vita. Includes bibliographical references (leaves 114-121). Also available on the Internet.
8

Exploiting non-redundant local patterns and probabilistic models for analyzing structured and semi-structured data

Wang, Chao, January 2008 (has links)
Thesis (Ph. D.)--Ohio State University, 2008. / Title from first page of PDF file. Includes bibliographical references (p. 140-150).
9

Investigations into the use of quantified Bayesian maximum entropy methods to generate improved distribution maps and biomass estimates from fisheries acoustic survey data /

Heywood, Ben, January 2008 (has links)
Thesis (M.Phil.) - University of St Andrews, April 2008.
10

Mesure de luminescence induite par faisceaux d'ions lourds rapides résolue à l'echelle picoseconde / Measurement of picosecond time-resolved, swift heavy ion induced luminescence

Durantel, Florent 13 December 2018 (has links)
Nous avons travaillé sur le développement d’un instrument de mesure de la luminescence induite par un faisceau d’ions lourds (nucléons  12) et d’énergie de l’ordre du MeV/nucléons. Basé sur une méthode de comptage de photons uniques obtenus par coïncidences, le dispositif permet d’obtenir sur 16 voies à la fois un spectre en énergie dans le domaine proche UV-visible-proche IR (185-920 nm) et la réponse temporelle sur la gamme ns-µs, avec un échantillonnage de 100 ps. Des mesures en température peuvent être réalisées depuis la température ambiante jusqu’à 30K.Ce travail met particulièrement l’accent sur les méthodes d’extraction des données : Une fois montrée la nécessité de déconvoluer les signaux, on s’intéresse dans un premier temps à évaluer différents profils instrumentaux modélisés et reconstruit à partir de mesures. A cet effet, un travail de caractérisation temporelle de chaque constituant du dispositif est mené. Puis ces profils instrumentaux sont utilisés dans deux méthodes de déconvolution par moindres carrés d’abord puis par maximum d’entropie ensuite.Deux matériaux types sont testés : Le Titanate de Strontium pour l’étude de la dynamique de l’excitation électronique, et un scintillateur plastique commercial, le BC400, pour l’étude du vieillissement et de la baisse des performances en fonction de la fluence. Dans les deux cas on a pu mettre en évidence la présence d’une composante ultra rapide de constante de temps subnanoseconde. / We developed an instrument for measuring the luminescence induced by a heavy ion beam (nucleons  12) and energy in the range of MeV / nucleon. Based on a single photon counting method obtained by coincidences, the device can provide in the same run a 16-channel energy spectrum in the UV-visible- IR region (185-920 nm) and a time-resolved response in the range of ns up to µs for each channel. Temperature measurements can be performed from room temperature down to 30K.This work places particular emphasis on data extraction methods: Once the need to deconvolve the signals demonstrated the evaluation of different instrument profiles (simulated and reconstructed from measurements) leads to a systematic temporal characterization of each component of the device. Then, these instrumental profiles are used in two deconvolution methods: least squares first followed by maximum entropy method.Two typical materials are tested: the Strontium Titanate for the study of the dynamics of the electronic excitation, and a commercial scintillator, the BC400, for the study of the aging and the decrease of performances with fluence. In both cases, we have been able to highlight the presence of an ultrafast component of subnanosecond time constant.

Page generated in 0.0921 seconds