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.

Avian haemosporidian persistence and co-infection in great tits at the individual level.

BACKGROUND: Many studies have tracked the distribution and persistence of avian haemosporidian communities across space and time at the population level, but few studies have investigated these aspects of infection at the individual level over time. Important aspects of parasite infection at the individual level can be missed if only trends at the population level are studied. This study aimed to determine how persistent Haemosporida are in great tit individuals recaptured over several years, whether parasitaemia differed by parasite lineage (mitochondrial cytochrome b haplotype) and how co-infection (i.e. concurrent infection with multiple genera of parasites) affects parasitaemia and body mass. METHODS: Parasite prevalence was determined by polymerase chain reaction (PCR), quantitative PCR were used to assess parasitaemia and sequencing was employed to determine the identity of the lineages using the MalAvi database. RESULTS: Haemosporidian prevalence was high over sampled years with 98% of 55 recaptured individuals showing infection in at least one year of capture. Eighty-two percent of all positive individuals suffered co-infection, with an overall haemosporidian lineage diversity of seventeen. Plasmodium and Haemoproteus parasites were found to be highly persistent, with lineages from these genera consistently found in individuals across years and with no differences in individual parasitaemia being recorded at subsequent captures. Conversely, Leucocytozoon parasites showed higher turnover with regard to lineage changes or transitions in infection status (infected vs non-infected) across years. Parasitaemia was found to be lineage specific and there was no relationship between Plasmodium parasitaemia or host body condition and the presence of Leucocytozoon parasites. CONCLUSIONS: The findings of this study suggest that different genera of haemosporidian parasites interact differently with their host and other co-infecting parasites, influencing parasite persistence most likely through inter-parasite competition or host-parasite immune interactions. Even-though co-infections do not seem to result in increased virulence (higher parasitaemia or poorer host body condition), further investigation into infection potential of these parasites, both individually and as co-infections, is necessary.

Pheomelanin-based coloration and the ability to cope with variation in food supply and parasitism.

Although gene by environment interactions may play a key role in the maintenance of genetic polymorphisms, little is known about the ecological factors involved in these interactions. We investigated whether food supply and parasites can mediate covariation between the degree of adult pheomelanin-based coloration, a heritable trait, and offspring body mass in the tawny owl (Strix aluco). We swapped clutches between nests to allocate genotypes randomly among environments. Three weeks after hatching, we challenged the immune system of 80 unrelated nestlings with either a phytohemagglutinin (PHA) or a lipopolysaccharide, surrogates of alternative parasites, and then fed them ad lib. or food-restricted them during the following 6 days in the laboratory. Whatever the immune challenge, nestlings fed ad lib. converted food more efficiently into body mass when their biological mother was dark pheomelanic. In contrast, food-restricted nestlings challenged with PHA lost less body mass when their biological mother was pale pheomelanic. Nestling tawny owls born from differently melanic mothers thus show differing reaction norms relative to food availability and parasitism. This suggests that dark and pale pheomelanic owls reflect alternative adaptations to food availability and parasites, factors known to vary in space and time.

Fast Track and Fast Track II - Cedar Rapids, Iowa, HR-544, Construction Report, 1989

Two lanes of a major four lane arterial street needed to be reconstructed in Cedar Rapids, Iowa. The traffic volumes and difficulty of detouring the traffic necessitated closure for construction be held to an absolute minimum. Closure of the intersections, even for one day, was not politically feasible. Therefore, Fast Track and Fast Track II was specified for the project. Fast Track concrete paving has been used successfully in Iowa since 1986. The mainline portion of the project was specified to be Fast Track and achieved the opening strength of 400 psi in less than twelve hours. The intersections were allowed to be closed between 6 PM and 6 AM. This could occur twice - once to remove the old pavement and place the base and temporary surface and the second time to pave and cure the new concrete. The contractor was able to meet these restrictions. The Fast Track II used in the intersections achieved the opening strength of 350 psi in six to seven hours. Two test sections were selected in the mainline Fast Track and two intersections were chosen to test the Fast Tract II. Both flexural and compression specimens were tested. Pulse velocity tests were conducted on the pavement and test specimens. Maturity curves were developed through monitoring of the temperatures. Correlations were performed between the maturity and pulse velocity and the flexural strengths. The project was successful in establishing the feasibility of construction at night, with no disruption of traffic in the daytime, using fast Track II. Both the Fast Track II pavements were performing well four years after construction.

Exploring the space of histidine containing dipeptides in search of novel efficient RCS sequestering agents.

The study reports a set of forty proteinogenic histidine-containing dipeptides as potential carbonyl quenchers. The peptides were chosen to cover as exhaustively as possible the accessible chemical space, and their quenching activities toward 4-hydroxy-2-nonenal (HNE) and pyridoxal were evaluated by HPLC analyses. The peptides were capped at the C-terminus as methyl esters or amides to favor their resistance to proteolysis and diastereoisomeric pairs were considered to reveal the influence of configuration on quenching. On average, the examined dipeptides are less active than the parent compound carnosine (βAla + His) thus emphasizing the unfavorable effect of the shortening of the βAla residue as confirmed by the control dipeptide Gly-His. Nevertheless, some peptides show promising activities toward HNE combined with a remarkable selectivity. The results emphasize the beneficial role of aromatic and positively charged residues, while negatively charged and H-bonding side chains show a detrimental effect on quenching. As a trend, ester derivatives are slightly more active than amides while heterochiral peptides are more active than their homochiral diastereoisomer. Overall, the results reveal that quenching activity strongly depends on conformational effects and vicinal residues (as evidenced by the reported QSAR analysis), offering insightful clues for the design of improved carbonyl quenchers and to rationalize the specific reactivity of histidine residues within proteins.

Effect of growth regulators on 'Brookfield' apple gas diffusion and metabolism under controlled atmosphere storage

The objective of this work was to evaluate the effect of growth regulators on gas diffusion and on metabolism of 'Brookfield' apple, and to determine their correlation with quality characteristics of fruit stored in controlled atmosphere. A completely randomized design was used with four replicates. After eight months of storage, the effects of water (control), aminoethoxyvinylglycine (AVG), AVG + ethephon, AVG + naphthaleneacetic acid (NAA), ethephon + NAA, sole NAA, 1-MCP, ethylene absorption by potassium permanganate (ABS), AVG + ABS, and of AVG + 1-MCP - applied at different rates and periods - were evaluated on: gas diffusion rate, ethylene production, respiratory rate, internal ethylene concentration, internal CO2 content, mealiness, and intercellular space. Fruit from the control and sole NAA treatments had the highest mealiness occurrence. Growth regulators significantly changed the gaseous diffusion through the pulp of 'Brookfield' apple, mainly in the treatment AVG + ABS, which kept the highest gas diffusion rate. NAA spraying in the field, with or without another growth regulator, increased ripening metabolism by rising ethylene production and respiration rate, and reduced gas diffusion during shelf life. AVG spraying cannot avoid the ethephon effect during the ripening process, and reduces both the internal space and mealiness incidence, but it is not able to induce ethylene production or to increase respiration rates.

Palaeotectonic and palaeogeographic evolution of the western Tethys and PeriTethyan domain (IGCP Project 369)

Project No. 369 (''Comparative Evolution of PeriTethyan Rift Basins'') of the International Geological Correlation Program produced a new palaeotectonic-palaeo-geographic atlas of the western PeriTethyan domain. The atlas contains more than two hundred new maps and documents grouped in nine regional sets (Iberia, Polish Trough, Eastern European and Scythian Platforms, Moesian Platform, Levant, Arabian Platform,, Northern Africa, NE Africa-NW Arabia, Libya-Pelagian Shelf plus a set of reconstructions for the whole western Tethys. The area, considered in the atlas stretches. from west to east, from the eastern Atlantic shores to the Urals and, from north to south, from the Baltic shield to equatorial Africa; the time span covered extends from the Late. Carboniferous to the Present. The dataset, resulting from an extensive cooperation between industrial and academic sources, is accessible interactively on a CD-ROM (Stampfli et al., 2001a) and includes legend, timetable, short explanatory notes, full references and additional supporting data. This dataset provides information on the development of the Tethyan realm in space and time. In particular, the relation between the Variscan and Cimmerian cycles in the Mediterranean realm is illustrated by numerous palaeogeographic and palaeotectonic maps.

Urgences vitales au cabinet médical: implications pour la formation et l'equipement du médecin de premier recours [Life-threatening emergencies at the office: implications for medical education and equipment of the primary care physician].

Every medical practitioner is confronted on a daily basis with emergencies. Among these, life-threatening emergencies can have disastrous consequences in term of morbidity and mortality; 22 cardiac arrests and 10 deaths were reported among the 1,650 Swiss practices during a 5 year period. The occurrence of life-threatening emergencies at the office necessitates, according to the type and place of the practice, the skills of the practitioner and the organization of his practice, the implementation of procedures, equipments (for example room equipped with a defibrillator, respiratory nebulizer, splints, emergency drugs) and specific continuous education programs that should be encouraged and made available to the whole medical corporation.

Forensic intelligence framework-Part I: induction of a transversal model by comparing illicit drugs and false identity documents monitoring

Forensic intelligence is a distinct dimension of forensic science. Forensic intelligence processes have mostly been developed to address either a specific type of trace or a specific problem. Even though these empirical developments have led to successes, they are trace-specific in nature and contribute to the generation of silos which hamper the establishment of a more general and transversal model. Forensic intelligence has shown some important perspectives but more general developments are required to address persistent challenges. This will ensure the progress of the discipline as well as its widespread implementation in the future. This paper demonstrates that the description of forensic intelligence processes, their architectures, and the methods for building them can, at a certain level, be abstracted from the type of traces considered. A comparative analysis is made between two forensic intelligence approaches developed independently in Australia and in Europe regarding the monitoring of apparently very different kind of problems: illicit drugs and false identity documents. An inductive effort is pursued to identify similarities and to outline a general model. Besides breaking barriers between apparently separate fields of study in forensic science and intelligence, this transversal model would assist in defining forensic intelligence, its role and place in policing, and in identifying its contributions and limitations. The model will facilitate the paradigm shift from the current case-by-case reactive attitude towards a proactive approach by serving as a guideline for the use of forensic case data in an intelligence-led perspective. A follow-up article will specifically address issues related to comparison processes, decision points and organisational issues regarding forensic intelligence (part II).

Divergent and narrower climatic niches characterize polyploid species of European primroses in Primula sect. Aleuritia

Aim It is hypothesized that the ecological niches of polyploids should be both distinct and broader than those of diploids - characteristics that might have allowed the successful colonization of open habitats by polyploids during the Pleistocene glacial cycles. Here, we test these hypotheses by quantifying and comparing the ecological niches and niche breadths of a group of European primroses. Location Europe. Methods We gathered georeferenced data of four related species in Primula sect. Aleuritia at different ploidy levels (diploid, tetraploid, hexaploid and octoploid) and used seven bioclimatic variables to quantify niche overlap between species by applying a series of univariate and multivariate analyses combined with modelling techniques. We also employed permutation-based tests to evaluate niche similarity between the four species. Niche breadth for each species was evaluated both in the multivariate environmental space and in geographical space. Results The four species differed significantly from each other in mono-dimensional comparisons of climatological variables and occupied distinct habitats in the multi-dimensional environmental space. The majority of the permutation-based tests either indicated that the four species differed significantly in their habitat preferences and ecological niches or did not support significant niche similarity. Furthermore, our results revealed narrower niche breadths and geographical ranges in species of P. sect. Aleuritia at higher ploidy levels. Main conclusions The detected ecological differentiation between the four species of P. sect. Aleuritia at different ploidy levels is consistent with the hypothesis that polyploids occupy distinct ecological niches that differ from those of their diploid relative. Contrary to expectations, we find that polyploid species of P. sect. Aleuritia occupy narrower environmental and geographical spaces than their diploid relative. These results on the ecological niches of closely related polyploid and diploid species highlight factors that potentially contribute to the evolution and distribution of polyploid species.

Detection of clusters using space-time scan statistics

This paper aims at detecting spatio-temporal clustering in fire sequences using space?time scan statistics, a powerful statistical framework for the analysis of point processes. The methodology is applied to active fire detection in the state of Florida (US) identified by MODIS (Moderate Resolution Imaging Spectroradiometer) during the period 2003?06. Results of the present study show that statistically significant clusters can be detected and localized in specific areas and periods of the year. Three out of the five most likely clusters detected for the entire frame period are localized in the north of the state, and they cover forest areas; the other two clusters cover a large zone in the south, corresponding to agricultural land and the prairies in the Everglades. In order to analyze if the wildfires recur each year during the same period, the analyses have been performed separately for the 4 years: it emerges that clusters of forest fires are more frequent in hot seasons (spring and summer), while in the southern areas, they are widely present during the whole year. The recognition of overdensities of events and the ability to locate them in space and in time can help in supporting fire management and focussing on prevention measures.

An embedded real-time red peach detection system based on an OV7670 camera, ARM Cortex-M4 processor and 3D Look-Up Tables

This work proposes the development of an embedded real-time fruit detection system for future automatic fruit harvesting. The proposed embedded system is based on an ARM Cortex-M4 (STM32F407VGT6) processor and an Omnivision OV7670 color camera. The future goal of this embedded vision system will be to control a robotized arm to automatically select and pick some fruit directly from the tree. The complete embedded system has been designed to be placed directly in the gripper tool of the future robotized harvesting arm. The embedded system will be able to perform real-time fruit detection and tracking by using a three-dimensional look-up-table (LUT) defined in the RGB color space and optimized for fruit picking. Additionally, two different methodologies for creating optimized 3D LUTs based on existing linear color models and fruit histograms were implemented in this work and compared for the case of red peaches. The resulting system is able to acquire general and zoomed orchard images and to update the relative tracking information of a red peach in the tree ten times per second.

The First World War and Coal Trade Geography in Latin America and the Caribbean, 1890-1930

This paper aims to illustrate the dynamics of coal trade between Latin America and its main trade partners, i.e., the USA, Great Britain, and Germany, before and after the enormous disruption caused by the First World War. The coal trade was used as an indicator of modernization for Latin American countries, given that oil was at that time of secondary importance. Energy imports have determined the possibilities of each Latin American country in its process of development. Here, we address this question and place special emphasis on supply channels, concluding that the trade link with main suppliers was of key significance. Although this was very clear by the end of the period, the process had started well before the First World War, at least for the majority of LA&C countries. These points are developed through a gravity model applied to the bilateral coal trade. The importance of the market supplier share is addressed through cluster methodologies.

Effects of P300-based BCI use on reported presence in a virtual environment

Brain-computer interfaces (BCIs) are becoming more and more popular as an input device for virtual worlds and computer games. Depending on their function, a major drawback is the mental workload associated with their use and there is significant effort and training required to effectively control them. In this paper, we present two studies assessing how mental workload of a P300-based BCI affects participants" reported sense of presence in a virtual environment (VE). In the first study, we employ a BCI exploiting the P300 event-related potential (ERP) that allows control of over 200 items in a virtual apartment. In the second study, the BCI is replaced by a gaze-based selection method coupled with wand navigation. In both studies, overall performance is measured and individual presence scores are assessed by means of a short questionnaire. The results suggest that there is no immediate benefit for visualizing events in the VE triggered by the BCI and that no learning about the layout of the virtual space takes place. In order to alleviate this, we propose that future P300-based BCIs in VR are set up so as require users to make some inference about the virtual space so that they become aware of it,which is likely to lead to higher reported presence.

Characteristics of 2D convective structures in Catalonia (NE Spain): an analysis using radar data and GIS

Flood simulation studies use spatial-temporal rainfall data input into distributed hydrological models. A correct description of rainfall in space and in time contributes to improvements on hydrological modelling and design. This work is focused on the analysis of 2-D convective structures (rain cells), whose contribution is especially significant in most flood events. The objective of this paper is to provide statistical descriptors and distribution functions for convective structure characteristics of precipitation systems producing floods in Catalonia (NE Spain). To achieve this purpose heavy rainfall events recorded between 1996 and 2000 have been analysed. By means of weather radar, and applying 2-D radar algorithms a distinction between convective and stratiform precipitation is made. These data are introduced and analyzed with a GIS. In a first step different groups of connected pixels with convective precipitation are identified. Only convective structures with an area greater than 32 km2 are selected. Then, geometric characteristics (area, perimeter, orientation and dimensions of the ellipse), and rainfall statistics (maximum, mean, minimum, range, standard deviation, and sum) of these structures are obtained and stored in a database. Finally, descriptive statistics for selected characteristics are calculated and statistical distributions are fitted to the observed frequency distributions. Statistical analyses reveal that the Generalized Pareto distribution for the area and the Generalized Extreme Value distribution for the perimeter, dimensions, orientation and mean areal precipitation are the statistical distributions that best fit the observed ones of these parameters. The statistical descriptors and the probability distribution functions obtained are of direct use as an input in spatial rainfall generators.