No description available.

Dans ce mémoire, nous nous intéressons à des algorithmes de sélection de modèles dans un contexte de régression linéaire et logistique. Nous expliquons premièrement les notions de régression linéaire et logistique et deux critères de sélection, AIC et BIC. Ensuite, nous faisons une revue des aspects théoriques des algorithmes les plus connus en détaillant deux d'entre eux, Leaps and Bounds et Occam’s Window. Pour ces deux derniers, nous présentons aussi les détails pratiques des logiciels qui font leur implantation. La partie finale est consacrée à l'étude des trois méthodes de sélection des modèles basées sur les algorithmes Leaps and Bounds, Occam’s Window et sur une combinaison entre les deux, en utilisant la technique du moyennage de modèles. Nous présentons les performances de prédiction calculées à l'aide de la technique de validation croisée et les temps d'exécution de ces trois méthodes pour plusieurs jeux de données. ______________________________________________________________________________ MOTS-CLÉS DE L’AUTEUR : sélection de modèles, moyennage de modèles, régression linéaire, régression logistique, AIC, BIC, algorithme Leaps and Bounds, algorithme Occam’s Window, validation croisée.

This thesis examines the characteristics of dusky dolphin (Lagenorhynchus obscurus) nursery groups and ontogeny of dusky dolphin calves. Data were collected via boat-based group focal follows of nurseries from October 2006-May 2007. A total of 87 nursery groups were encountered. Data were analyzed according to age category (infant or yearling) and season (early or late). Nursery group membership was lowest in the early season and when yearlings were present. The average number of yearlings in a nursery group was less than that of infants. The predominant activity of calves was rest. Early infants rested the most, while travel seemed most important for late infants, and early yearlings were most likely to forage. With the exception of early infants, all calves were more likely than adults to interact with boats. When taking month into account, yearlings were more social in general than infants. Infants showed a positive trend in sociality, while yearling sociality remained relatively stable. Nursery groups are markedly segregated by calf age, and 80% of nursery groups contained calves of only one age group. Dusky dolphin calves show a similar trend in preference for position in relation to the mother as that in bottlenose dolphins (Tursiops sp.), with echelon swim decreasing with age. However, all calves appear to prefer echelon swim when nursery groups are traveling. Calves were more likely to swim independently in the late part of the season and while foraging or socializing, and were more likely to be in close proximity to their mothers while resting or traveling. Calves learned noisy leaps, followed by clean, coordinated, and acrobatic leaps, in that order. There was no clear relationship between behavioral state and types of leaps performed by calves. Early infants leapt less often than older calves, but leap frequency did not differ among the older calves. The overall pattern in the ontogeny of dusky dolphin leaps indicates that the physical development of leaps is learned individually, while the context in which the leaps are performed is learned from conspecifics. These results indicate that nursery groups represent an important environment for healthy physical and social development of calves.

This thesis contains two parts: the detectability of convex sets and the study on regression models In the first part of this dissertation, we investigate the problem of the detectability of an inhomogeneous convex region in a Gaussian random field. The first proposed detection method relies on checking a constructed statistic on each convex set within an nn image, which is proven to be un-applicable. We then consider using h(v)-parallelograms as the surrogate, which leads to a multiscale strategy. We prove that 2/9 is the minimum proportion of the maximally embedded h(v)-parallelogram in a convex set. Such a constant indicates the effectiveness of the above mentioned multiscale detection method. In the second part, we study the robustness, the optimality, and the computing for regression models. Firstly, for robustness, M-estimators in a regression model where the residuals are of unknown but stochastically bounded distribution are analyzed. An asymptotic minimax M-estimator (RSBN) is derived. Simulations demonstrate the robustness and advantages. Secondly, for optimality, the analysis on the least angle regressions inspired us to consider the conditions under which a vector is the solution of two optimization problems. For these two problems, one can be solved by certain stepwise algorithms, the other is the objective function in many existing subset selection criteria (including Cp, AIC, BIC, MDL, RIC, etc). The latter is proven to be NP-hard. Several conditions are derived. They tell us when a vector is the common optimizer. At last, extending the above idea about finding conditions into exhaustive subset selection in regression, we improve the widely used leaps-and-bounds algorithm (Furnival and Wilson). The proposed method further reduces the number of subsets needed to be considered in the exhaustive subset search by considering not only the residuals, but also the model matrix, and the current coefficients.

This thesis examines the characteristics of dusky dolphin (Lagenorhynchus obscurus) nursery groups and ontogeny of dusky dolphin calves. Data were collected via boat-based group focal follows of nurseries from October 2006-May 2007. A total of 87 nursery groups were encountered. Data were analyzed according to age category (infant or yearling) and season (early or late). Nursery group membership was lowest in the early season and when yearlings were present. The average number of yearlings in a nursery group was less than that of infants. The predominant activity of calves was rest. Early infants rested the most, while travel seemed most important for late infants, and early yearlings were most likely to forage. With the exception of early infants, all calves were more likely than adults to interact with boats. When taking month into account, yearlings were more social in general than infants. Infants showed a positive trend in sociality, while yearling sociality remained relatively stable. Nursery groups are markedly segregated by calf age, and 80% of nursery groups contained calves of only one age group. Dusky dolphin calves show a similar trend in preference for position in relation to the mother as that in bottlenose dolphins (Tursiops sp.), with echelon swim decreasing with age. However, all calves appear to prefer echelon swim when nursery groups are traveling. Calves were more likely to swim independently in the late part of the season and while foraging or socializing, and were more likely to be in close proximity to their mothers while resting or traveling. Calves learned noisy leaps, followed by clean, coordinated, and acrobatic leaps, in that order. There was no clear relationship between behavioral state and types of leaps performed by calves. Early infants leapt less often than older calves, but leap frequency did not differ among the older calves. The overall pattern in the ontogeny of dusky dolphin leaps indicates that the physical development of leaps is learned individually, while the context in which the leaps are performed is learned from conspecifics. These results indicate that nursery groups represent an important environment for healthy physical and social development of calves.

S tím, jak se neustále vyvíjejí nové technologie pro energeticky náročná průmyslová odvětví, stávající zařízení postupně zaostávají v efektivitě a produktivitě. Tvrdá konkurence na trhu a legislativa v oblasti životního prostředí nutí tato tradiční zařízení k ukončení provozu a k odstavení. Zlepšování procesu a projekty modernizace jsou zásadní v udržování provozních výkonů těchto zařízení. Současné přístupy pro zlepšování procesů jsou hlavně: integrace procesů, optimalizace procesů a intenzifikace procesů. Obecně se v těchto oblastech využívá matematické optimalizace, zkušeností řešitele a provozní heuristiky. Tyto přístupy slouží jako základ pro zlepšování procesů. Avšak, jejich výkon lze dále zlepšit pomocí moderní výpočtové inteligence. Účelem této práce je tudíž aplikace pokročilých technik umělé inteligence a strojového učení za účelem zlepšování procesů v energeticky náročných průmyslových procesech. V této práci je využit přístup, který řeší tento problém simulací průmyslových systémů a přispívá následujícím: (i)Aplikace techniky strojového učení, která zahrnuje jednorázové učení a neuro-evoluci pro modelování a optimalizaci jednotlivých jednotek na základě dat. (ii) Aplikace redukce dimenze (např. Analýza hlavních komponent, autoendkodér) pro vícekriteriální optimalizaci procesu s více jednotkami. (iii) Návrh nového nástroje pro analýzu problematických částí systému za účelem jejich odstranění (bottleneck tree analysis – BOTA). Bylo také navrženo rozšíření nástroje, které umožňuje řešit vícerozměrné problémy pomocí přístupu založeného na datech. (iv) Prokázání účinnosti simulací Monte-Carlo, neuronové sítě a rozhodovacích stromů pro rozhodování při integraci nové technologie procesu do stávajících procesů. (v) Porovnání techniky HTM (Hierarchical Temporal Memory) a duální optimalizace s několika prediktivními nástroji pro podporu managementu provozu v reálném čase. (vi) Implementace umělé neuronové sítě v rámci rozhraní pro konvenční procesní graf (P-graf). (vii) Zdůraznění budoucnosti umělé inteligence a procesního inženýrství v biosystémech prostřednictvím komerčně založeného paradigmatu multi-omics.