Machine learning for geospatial data : algorithms, software tools and case studies

Résumé Cette thèse est consacrée à l'analyse, la modélisation et la visualisation de données environnementales à référence spatiale à l'aide d'algorithmes d'apprentissage automatique (Machine Learning). L'apprentissage automatique peut être considéré au sens large comme une sous-catégorie de l'intelligence artificielle qui concerne particulièrement le développement de techniques et d'algorithmes permettant à une machine d'apprendre à partir de données. Dans cette thèse, les algorithmes d'apprentissage automatique sont adaptés pour être appliqués à des données environnementales et à la prédiction spatiale. Pourquoi l'apprentissage automatique ? Parce que la majorité des algorithmes d'apprentissage automatiques sont universels, adaptatifs, non-linéaires, robustes et efficaces pour la modélisation. Ils peuvent résoudre des problèmes de classification, de régression et de modélisation de densité de probabilités dans des espaces à haute dimension, composés de variables informatives spatialisées (« géo-features ») en plus des coordonnées géographiques. De plus, ils sont idéaux pour être implémentés en tant qu'outils d'aide à la décision pour des questions environnementales allant de la reconnaissance de pattern à la modélisation et la prédiction en passant par la cartographie automatique. Leur efficacité est comparable au modèles géostatistiques dans l'espace des coordonnées géographiques, mais ils sont indispensables pour des données à hautes dimensions incluant des géo-features. Les algorithmes d'apprentissage automatique les plus importants et les plus populaires sont présentés théoriquement et implémentés sous forme de logiciels pour les sciences environnementales. Les principaux algorithmes décrits sont le Perceptron multicouches (MultiLayer Perceptron, MLP) - l'algorithme le plus connu dans l'intelligence artificielle, le réseau de neurones de régression généralisée (General Regression Neural Networks, GRNN), le réseau de neurones probabiliste (Probabilistic Neural Networks, PNN), les cartes auto-organisées (SelfOrganized Maps, SOM), les modèles à mixture Gaussiennes (Gaussian Mixture Models, GMM), les réseaux à fonctions de base radiales (Radial Basis Functions Networks, RBF) et les réseaux à mixture de densité (Mixture Density Networks, MDN). Cette gamme d'algorithmes permet de couvrir des tâches variées telle que la classification, la régression ou l'estimation de densité de probabilité. L'analyse exploratoire des données (Exploratory Data Analysis, EDA) est le premier pas de toute analyse de données. Dans cette thèse les concepts d'analyse exploratoire de données spatiales (Exploratory Spatial Data Analysis, ESDA) sont traités selon l'approche traditionnelle de la géostatistique avec la variographie expérimentale et selon les principes de l'apprentissage automatique. La variographie expérimentale, qui étudie les relations entre pairs de points, est un outil de base pour l'analyse géostatistique de corrélations spatiales anisotropiques qui permet de détecter la présence de patterns spatiaux descriptible par une statistique. L'approche de l'apprentissage automatique pour l'ESDA est présentée à travers l'application de la méthode des k plus proches voisins qui est très simple et possède d'excellentes qualités d'interprétation et de visualisation. Une part importante de la thèse traite de sujets d'actualité comme la cartographie automatique de données spatiales. Le réseau de neurones de régression généralisée est proposé pour résoudre cette tâche efficacement. Les performances du GRNN sont démontrées par des données de Comparaison d'Interpolation Spatiale (SIC) de 2004 pour lesquelles le GRNN bat significativement toutes les autres méthodes, particulièrement lors de situations d'urgence. La thèse est composée de quatre chapitres : théorie, applications, outils logiciels et des exemples guidés. Une partie importante du travail consiste en une collection de logiciels : Machine Learning Office. Cette collection de logiciels a été développée durant les 15 dernières années et a été utilisée pour l'enseignement de nombreux cours, dont des workshops internationaux en Chine, France, Italie, Irlande et Suisse ainsi que dans des projets de recherche fondamentaux et appliqués. Les cas d'études considérés couvrent un vaste spectre de problèmes géoenvironnementaux réels à basse et haute dimensionnalité, tels que la pollution de l'air, du sol et de l'eau par des produits radioactifs et des métaux lourds, la classification de types de sols et d'unités hydrogéologiques, la cartographie des incertitudes pour l'aide à la décision et l'estimation de risques naturels (glissements de terrain, avalanches). Des outils complémentaires pour l'analyse exploratoire des données et la visualisation ont également été développés en prenant soin de créer une interface conviviale et facile à l'utilisation. Machine Learning for geospatial data: algorithms, software tools and case studies Abstract The thesis is devoted to the analysis, modeling and visualisation of spatial environmental data using machine learning algorithms. In a broad sense machine learning can be considered as a subfield of artificial intelligence. It mainly concerns with the development of techniques and algorithms that allow computers to learn from data. In this thesis machine learning algorithms are adapted to learn from spatial environmental data and to make spatial predictions. Why machine learning? In few words most of machine learning algorithms are universal, adaptive, nonlinear, robust and efficient modeling tools. They can find solutions for the classification, regression, and probability density modeling problems in high-dimensional geo-feature spaces, composed of geographical space and additional relevant spatially referenced features. They are well-suited to be implemented as predictive engines in decision support systems, for the purposes of environmental data mining including pattern recognition, modeling and predictions as well as automatic data mapping. They have competitive efficiency to the geostatistical models in low dimensional geographical spaces but are indispensable in high-dimensional geo-feature spaces. The most important and popular machine learning algorithms and models interesting for geo- and environmental sciences are presented in details: from theoretical description of the concepts to the software implementation. The main algorithms and models considered are the following: multi-layer perceptron (a workhorse of machine learning), general regression neural networks, probabilistic neural networks, self-organising (Kohonen) maps, Gaussian mixture models, radial basis functions networks, mixture density networks. This set of models covers machine learning tasks such as classification, regression, and density estimation. Exploratory data analysis (EDA) is initial and very important part of data analysis. In this thesis the concepts of exploratory spatial data analysis (ESDA) is considered using both traditional geostatistical approach such as_experimental variography and machine learning. Experimental variography is a basic tool for geostatistical analysis of anisotropic spatial correlations which helps to understand the presence of spatial patterns, at least described by two-point statistics. A machine learning approach for ESDA is presented by applying the k-nearest neighbors (k-NN) method which is simple and has very good interpretation and visualization properties. Important part of the thesis deals with a hot topic of nowadays, namely, an automatic mapping of geospatial data. General regression neural networks (GRNN) is proposed as efficient model to solve this task. Performance of the GRNN model is demonstrated on Spatial Interpolation Comparison (SIC) 2004 data where GRNN model significantly outperformed all other approaches, especially in case of emergency conditions. The thesis consists of four chapters and has the following structure: theory, applications, software tools, and how-to-do-it examples. An important part of the work is a collection of software tools - Machine Learning Office. Machine Learning Office tools were developed during last 15 years and was used both for many teaching courses, including international workshops in China, France, Italy, Ireland, Switzerland and for realizing fundamental and applied research projects. Case studies considered cover wide spectrum of the real-life low and high-dimensional geo- and environmental problems, such as air, soil and water pollution by radionuclides and heavy metals, soil types and hydro-geological units classification, decision-oriented mapping with uncertainties, natural hazards (landslides, avalanches) assessments and susceptibility mapping. Complementary tools useful for the exploratory data analysis and visualisation were developed as well. The software is user friendly and easy to use.

Pääomarakenteisiin vaikuttavat tekijät suomalaisissa pörssiyrityksissä

Tutkielman tavoitteena on selvittää lineaarisen regressioanalyysin avulla paneelidataa käyttäen suomalaisten pörssiyritysten pääomarakenteisiin vaikuttavat tekijät vuosina 1999-2004. Näiden tekijöiden avulla päätellään, mitä pääomarakenneteoriaa/-teorioita nämä yritykset noudattavat. Pääomarakenneteoriat voidaan jakaa kahteen luokkaan sen mukaan, pyritäänkö niissä optimaaliseen pääomarakenteeseen vai ei. Tradeoff- ja siihen liittyvässä agenttiteoriassa pyritään optimaaliseen pääomarakenteeseen. Tradeoff-teoriassa pääomarakenne valitaan punnitsemalla vieraan pääoman hyötyjä ja haittoja. Agenttiteoria on muuten samanlainen kuin tradeoff-teoria, mutta siinä otetaan lisäksi huomioon velan agenttikustannukset. Pecking order - ja ajoitusteoriassa ei pyritä optimaaliseen pääoma-rakenteeseen. Pecking order -teoriassa rahoitus valitaan hierarkian mukaan (tulorahoitus, vieras pääoma, välirahoitus, oma pääoma). Ajoitusteoriassa valitaan se rahoitusmuoto, jota on kannattavinta hankkia vallitsevassa markkinatilanteessa. Empiiristen tulosten mukaan velkaantumisaste riippuu positiivisesti riskistä, vakuudesta ja aineettomasta omaisuudesta. Velkaantumisaste riippuu negatiivisesti likviditeetistä, osaketuotoista ja kannattavuudesta. Osingoilla ei ole vaikutusta velkaantumisasteeseen. Toimialoista teollisuustuotteiden ja -palveluiden sekä perusteollisuuden aloilla on korkeammat velkaantumisasteet kuin muilla toimialoilla. Tulokset tukevat pääosin pecking order -teoriaa ja jonkin verran ajoitusteoriaa. Muut teoriat saavat vain vähäistä tukea.

Entendendo estrutura molecular com a molécula de hidrogênio ionizada

In this paper a general view about the modern molecular structure theory is developed discussing the ionized hydrogen molecule. We introduce some necessary approximation methods for the electronic and nuclear spectra study adopting a systematic approach. In addition though, we have performed calculations in order to illustrate these methods.

Síntese de membranas cerâmicas via método sol-gel utilizando TEOS e N,N-dimetilformamida

In this work we obtained microporous and mesoporous silica membranes by sol-gel processing. Tetraethylortosilicate (TEOS) was used as precursor. Nitric acid was used as catalyst. In order to study the affect of N,N-dimethylformamide (NDF) as drying additive, we used a molar ratio TEOS/NDF of 1/3. The performance of N,N-dimethylformamide was evaluated through monolithicity measurements. The structural evolutions occurring during the sol-gel transition and in the interconnected network of the membranes during thermal treatment were monitored by Fourier transform infrared spectroscopy (FTIR), thermogravimetric analyses and nitrogen sorption. We noted that in the presence of N,N-dimethylformamide, polymerization goes through a temporary stabilization of oligomers. The Si-O(H) bonds are stronger and belong to a more cross-linked structure for the N,N-dimethylformamide containing sol. The membranes obtained in the presence of N,N-dimethylformamide have larger surface area and its pore structure is in the range of mesoporous. The membranes obtained without additive have pore structure in the range of microporous.

Application of a Pid+fuzzy controller on the motion control system in machine tools

This work deals with an hybrid PID+fuzzy logic controller applied to control the machine tool biaxial table motions. The non-linear model includes backlash and the axis elasticity. Two PID controllers do the primary table control. A third PID+fuzzy controller has a cross coupled structure whose function is to minimise the trajectory contour errors. Once with the three PID controllers tuned, the system is simulated with and without the third controller. The responses results are plotted and compared to analyse the effectiveness of this hybrid controller over the system. They show that the proposed methodology reduces the contour error in a proportion of 70:1.

Modification of commercial poly-lactide for extrusion coated food packaging applications

Poly-L-lactide (PLLA) is a widely used sustainable and biodegradable alternative to replace synthetic non-degradable plastic materials in the packaging industry. Conversely, its processing properties are not always optimal, e.g. insufficient melt strength at higher temperatures (necessary in extrusion coating processes). This thesis reports on research to improve properties of commercial PLLA grade (3051D from NatureWorks), to satisfy and extend end-use applications, such as food packaging by blending with modified PLLA. Adjustment of the processability by chain branching of commercial poly-L-lactide initiated by peroxide was evaluated. Several well-defined branched structures with four arms (sPLLA) were synthesized using pentaerythritol as a tetra-functional initiator. Finally, several block copolymers consisting of polyethylene glycol and PLLA (i.e. PEGLA) were produced to obtain a well extruded material with improved heat sealing properties. Reactive extrusion of poly-L-lactide was carried out in the presence of 0.1, 0.3 and 0.5 wt% of various peroxides [tert-butyl-peroxybenzoate (TBPB), 2,5-dimethyl-2,5-(tert-butylperoxy)-hexane (Lupersol 101; LOL1) and benzoyl peroxide (BPO)] at 190C. The peroxide-treated PLLAs showed increased complex viscosity and storage modulus at lower frequencies, indicating the formation of branched/cross linked architectures. The material property changes were dependent on the peroxide, and the used peroxide concentration. Gel fraction analysis showed that the peroxides, afforded different gel contents, and especially 0.5 wt% peroxide, produced both an extremely high molar mass, and a cross linked structure, not perhaps well suited for e.g. further use in a blending step. The thermal behavior was somewhat unexpected as the materials prepared with 0.5 wt% peroxide showed the highest ability for crystallization and cold crystallization, despite substantial cross linking. The peroxide-modified PLLA, i.e. PLLA melt extruded with 0.3 wt% of TBPB and LOL1 and 0.5 wt% BPO was added to linear PLLA in ratios of 5, 15 and 30 wt%. All blends showed increased zero shear viscosity, elastic nature (storage modulus) and shear sensitivity. All blends remained amorphous, though the ability of annealing was improved slightly. Extrusion coating on paperboard was conducted with PLLA, and peroxide-modified PLLA blends (90:10). All blends were processable, but only PLLA with 0.3 wt% of LOL1 afforded a smooth high quality surface with improved line speed. Adhesion levels between fiber and plastic, as well as heat seal performance were marginally reduced compared with pure 3051D. The water vapor transmission measurements (WVTR) of the blends containing LOL1 showed acceptable levels, only slightly lower than for comparable PLLA 3051D. A series of four-arm star-shaped poly-L-lactide (sPLLA) with different branch length was synthesized by ring opening polymerization (ROP) of L-lactide using pentaerythritol as initiator and stannous octoate as catalyst. The star-shaped polymers were further blended with its linear resin and studied for their melt flow and thermal properties. Blends containing 30 wt% of sPLLA with low molecular weight (30 wt%; Mwtotal: 2500 g mol-1 and 15000 g mol-1) showed lower zero shear viscosity and significantly increased shear thinning, while at the same time slightly increased crystallization of the blend. However, the amount of crystallization increased significantly with the higher molecular weight sPLLA, therefore the star-shaped structure may play a role as nucleating agent. PLLA-polyethylene glycol–PLLA triblock copolymers (PEGLA) with different PLLA block length were synthesized and their applicability as blends with linear PLLA (3051D NatureWorks) was investigated with the intention of improving heat-seal and adhesion properties of extrusion-coated paperboard. PLLA-PEG-PLLA was obtained by ring opening polymerization (ROP) of L-lactide using PEG (molecular weight 6000 g mol-1) as an initiator, and stannous octoate as catalyst. The structures of the PEGLAs were characterized by proton nuclear magnetic resonance spectroscopy (1H-NMR). The melt flow and thermal properties of all PEGLAs and their blends were evaluated using dynamic rheology, and differential scanning calorimeter (DSC). All blends containing 30 wt% of PEGLAs showed slightly higher zero shear viscosity, higher shear thinning and increased melt elasticity (based on tan delta). Nevertheless, no significant changes in thermal properties were distinguished. High molecular weight PEGLAs were used in extrusion coating line with 3051D without problems.

Argentina’s Worker-Recovered Factories : strategies and survival

Les usines récupérées par les ouvriers en Argentine sont devenues un mouvement social emblématique symbolisant l'un des aspects de la révolte sociale entourant la crise économique de 2001-2002. Les usines récupérées sont des entreprises abandonnées par leurs propriétaires originaux ou déclarées faillite, laissant derrières elles des salaires et des dettes impayés. Par conséquence, les ouvriers ont commencé à récupérer leurs usines; reprenant la production sans leurs anciens patrons, sous, et au profit de la gestion collective des ouvriers. Le mouvement est remarquable pour sa rémunération égalitaire et sa gestion horizontale. Ce travail examine la continuité des usines récupérées et ceci à travers l'évolution sociale, politique et économique du paysage de l'Argentine. Il évalue également l'impact du mouvement en tant que défi aux modes économiques de production hégémoniques et orientés vers le marché. En supposant que l'avenir du mouvement dépend de deux ensembles de facteurs, le rapport analyse les facteurs internes à travers le prisme de la théorie de mobilisation des ressources, ainsi que les facteurs externes à travers la perspective de la théorie de la structure de l'opportunité politique. Le travail conclut que la situation actuelle se trouve dans une impasse dans laquelle le mouvement a gagné l'acceptation institutionnelle, mais a échoué d'effectuer le changement structurel favorisant ses pratiques et garantissant la sécurité à long terme. Il argumente que le mouvement doit consolider certains aspects combatifs. Il doit consolider sa nouvelle identité en tant que mouvement social et forger des alliances stratégiques et tactiques tout en préservant son autonomie.

X-Ray Crysatllographic and Vibrational Spectroscopic Studies of Thorium Bromate Hydrate

Th(BrO3)3·H2O single crystals were grown from its aqueous solution at room temperature. Single crystal XRD, Raman and FTIR techniques were used to investigate the crystal structure. The crystal structure was solved by Patterson method. The as grown crystals are in monoclinic system with space group P21/c. The unit cell parameters are a = 12.8555(18) Å, b = 7.8970(11) Å, c = 9.0716(10) Å,  = 90°,  = 131.568° and  = 90° and unit cell volume is 689.1(2) Å3. Z = 8, R factor is 5.9. The Raman and FTIR studies indicate the lowering of symmetry of bromate anion from C3V to C1. Hydrogen bonds with varying strengths are present in the crystal. The centrosymmetric space group P21/c of the crystal is confirmed by the non-coincidence of majority of Raman and IR bands

Cooperation or Confrontation?

With globalization and increased connectedness, migration has become a political issue. Nevertheless, without citizenship participation in the political system in the host society is limited. Based on a neo-institutionalist approach and referring to political opportunity structure theory, this paper analyzes differences in means of political participation by migrants in two cities, one in the U.S. and one in Germany. Specific focus is put on the welfare state as one factor potentially influencing forms of participation of migrants. From interviews with migrants, local governments, and organizations, this paper establishes that political participation in the German city, Essen, is more institutionalized than in the U.S. city, Newark, NJ, where demonstrations and rallies play a more significant role. Looking at these findings, this paper explains the differences with a variation in the political opportunity structure between the two cities. Whereas in the conservative-corporatist welfare state, ideas of collective bar-gaining and the conferral of social rights to migrants leads to government-created bodies for migrant participation, in the U.S. city, these bodies do not exist and, therefore, migrants use different means of political participation. Through the conferral of social rights on migrants as well, in Germany, the cleavage between migrants and majority society has been pacified. In the U.S. city, where this is not the case, demonstrations and rallies are more common.

La Ley de Tecnologías de la Información y las Comunicaciones (L1341/09). Actualización de un marco legal para un sector en constante evolución

La necesidad de preparar al sector de tecnologías de la información y las comunicaciones para hacer frente a los desafíos que trae consigo el desarrollo de la convergencia en todas sus dimensiones, implicaba plantear un nuevo equilibrio entre la promoción del desarrollo competitivo del sector y el cumplimiento de los compromisos sociales de cobertura derivados de la naturaleza de servicio público que ostentan las telecomunicaciones. En consecuencia, desde comienzos de 2007 se trabajó, con una permanente retroalimentación intragubernamental y sectorial, en la estructuración de los pilares del Proyecto de Ley 112/07 Cámara - 340/08 Senado, que culminó en la sanción presidencial de la Ley 1341 el30 de julio de 2009. Este nuevo marco legal para un sector en constante evolución constituye un hito sin precedentes, que rompe con la tradición de más de diez años y seis intentos fallidos de ajuste legislativo e institucional.

Las organizaciones no gubernamentales como instrumentos de intervención de Estados Unidos en las revoluciones de colores. Estudio de caso: revolución naranja en Ucrania

El presente trabajo de investigación tiene por objetivo explicar la estrategia de intervención de Estados Unidos en las Revoluciones de Colores, teniendo como estudio de caso a la Revolución Naranja de Ucrania en 2004. En base a ello, esta monografía describe y analiza las principales estrategias de poder blando implementadas por ciertas ONGS, que por razones políticas e ideológicas, se convirtieron en la principal herramienta de intervención de EE.UU. en su intento por respaldar un cambio pacífico de régimen en Ucrania. Para lo cual, se hará uso del concepto de Poder Blando de Joseph Nye, y de manera subsidiaria, de la teoría de Estructura de Oportunidades Políticas de Sidney Tarrow. Con el objetivo de comprender el triunfo de la Revolución Naranja a partir del contexto político nacional, y sobre todo, en razón de los intereses de política exterior de EE.UU. en el espacio Post Soviético.

Functionalised amyloid fibrils for roles in cell adhesion

We describe experiments designed to explore the possibility of using amyloid fibrils as new nanoscale biomaterials for promoting and exploiting cell adhesion, migration and differentiation in vitro. We created peptides that add the biological cell adhesion sequence (RGD) or a control sequence (RAD) to the C-terminus of an 11-residue peptide corresponding to residues 105-115 of the amyloidogenic protein transthyretin. These peptides readily self-assemble in aqueous solution to form amyloid fibrils, and X-ray fibre diffraction shows that they possess the same strand and sheet spacing in the characteristic cross-beta structure as do fibrils formed by the parent peptide. We report that the fibrils containing the RGD sequence are bioactive and that these fibrils interact specifically with cells via the RGD group displayed on the fibril surface. As the design of such functionalized fibrils can be systematically altered, these findings suggest that it will be possible to generate nanomaterials based on amyloid fibrils that are tailored to promote interactions with a wide variety of cell types. (c) 2007 Elsevier Ltd. All rights reserved.

Local orientational disorder in peptide fibrils probed by a combination of residue-specific 13C-18O labelling, polarised infrared spectroscopy and molecular combing

A novel combination of site-specific isotope labelling, polarised infrared spectroscopy and molecular combing reveal local orientational ordering in the fibril-forming peptide YTIAALLSPYSGGRADS. Use of 13C-18O labelled alanine residues demonstrates that the Nterminal end of the peptide is incorporated into the cross-beta structure, while the C-terminal end shows orientational disorder

Can British multinational enterprises finance economic development in South East Asia?

Purpose – The purpose of this paper is to shed new light on the debate about the role of foreign direct investment (FDI) and public policy in fostering economic development. Specifically, can the capital inflow of multinational enterprises (MNEs) and the ability of the subsidiaries to raise funds locally help promote development? This paper addresses this issue by examining the capital structure and financing sources of foreign subsidiaries of MNEs. Design/methodology/approach – This paper integrates the capital structure theories in finance with internalization theory in international business. It uses an original primary dataset collected by a survey of 101 foreign subsidiaries of British MNEs in six emerging economies in the ASEAN region. Findings – There are three significant findings. First, these subsidiaries rely heavily on internal funds generated within the MNEs and less on external debts raised in the host countries. Second, the foreign subsidiary's capital structure is influenced by the home country of origin of the parent firm and the parent firm's financing sources. Third, these subsidiaries have used the financial resources to develop business networks with local small and medium enterprises (SMEs) which contribute to economic development of the host countries. Originality/value – This paper examines the internal capital market within the MNE. It provides theoretical and empirical support for the capital structure theory of the hierarchy financing approach and also for internalization theory by addressing FDI inflows by MNEs and the raising of funds locally. These findings have important implications for public policy, namely the facilitation of MNE entry to encourage economic development.

Galois orders in skew monoid rings

We introduce a new class of noncommutative rings - Galois orders, realized as certain subrings of invariants in skew semigroup rings, and develop their structure theory. The class of Calms orders generalizes classical orders in noncommutative rings and contains many important examples, such as the Generalized Weyl algebras, the universal enveloping algebra of the general linear Lie algebra, associated Yangians and finite W-algebras (C) 2010 Elsevier Inc All rights reserved