Machine learning algorithms for spatial data. Case studies: Environmental pollution, natural hazards, renewable resources

This paper presents general problems and approaches for the spatial data analysis using machine learning algorithms. Machine learning is a very powerful approach to adaptive data analysis, modelling and visualisation. The key feature of the machine learning algorithms is that they learn from empirical data and can be used in cases when the modelled environmental phenomena are hidden, nonlinear, noisy and highly variable in space and in time. Most of the machines learning algorithms are universal and adaptive modelling tools developed to solve basic problems of learning from data: classification/pattern recognition, regression/mapping and probability density modelling. In the present report some of the widely used machine learning algorithms, namely artificial neural networks (ANN) of different architectures and Support Vector Machines (SVM), are adapted to the problems of the analysis and modelling of geo-spatial data. Machine learning algorithms have an important advantage over traditional models of spatial statistics when problems are considered in a high dimensional geo-feature spaces, when the dimension of space exceeds 5. Such features are usually generated, for example, from digital elevation models, remote sensing images, etc. An important extension of models concerns considering of real space constrains like geomorphology, networks, and other natural structures. Recent developments in semi-supervised learning can improve modelling of environmental phenomena taking into account on geo-manifolds. An important part of the study deals with the analysis of relevant variables and models' inputs. This problem is approached by using different feature selection/feature extraction nonlinear tools. To demonstrate the application of machine learning algorithms several interesting case studies are considered: digital soil mapping using SVM, automatic mapping of soil and water system pollution using ANN; natural hazards risk analysis (avalanches, landslides), assessments of renewable resources (wind fields) with SVM and ANN models, etc. The dimensionality of spaces considered varies from 2 to more than 30. Figures 1, 2, 3 demonstrate some results of the studies and their outputs. Finally, the results of environmental mapping are discussed and compared with traditional models of geostatistics.

Violences Domestiques : Prise en Charge et Prévention

Longtemps cachée, déniée, considérée comme honteuse par ses victimes, la violence domestique n'en est pas moins une réalité concrète qui tue, dans notre pays, une femme toutes les deux semaines et qui coûte des millions de francs par an à la société civile. Dès lors, ce type de violence ne peut plus être considéré comme relevant uniquement de la sphère privée. L'isolement des victimes, dû à l'incompréhension du phénomène, aux préjugés, à la peur et à l'ignorance de structures d'aides, n'est plus admissible. La violence domestique est un cas complexe et aucune discipline ne peut, à elle seule, y faire face. Les solutions sont donc l'affaire de tous. C'est pourquoi des professionnels romands créent le programme «C'est assez», au début des années 2000. S'ensuit un long cheminement pour mettre en lumière la problématique de la violence domestique et construire un réseau de soins, de prise en charge et de prévention cohérent, soutenu tant par le travail d'infirmiers, de travailleurs sociaux, de policiers, de médecins et de psychologues que de juges et de politiciens. Cet ouvrage parle de succès, d'échecs, d'espoirs, et il cherche avant tout à rendre sensible chaque citoyen à ce qui devrait être une évidence: prévenir la violence domestique et en faire une affaire de santé publique.

PHE-CRCE-010: Estimating Local Mortality Burdens associated with Particulate Air Pollution

The increase in mortality risk associated with long-term exposure to particulate air pollution is one of the most important, and best-characterised, effects of air pollution on health. This report presents estimates of the size of this effect on mortality in local authority areas in the UK, building upon the attributable fractions reported as an indicator in the public health outcomes framework for England. It discusses the concepts and assumptions underlying these calculations and gives information on how such estimates can be made. The estimates are expected to be useful to health and wellbeing boards when assessing local public health priorities, as well as to others working in the field of air quality and public health. The estimates of mortality burden are based on modelled annual average concentrations of fine particulate matter (PM2.5) in each local authority area originating from human activities. Local data on the adult population and adult mortality rates is also used. Central estimates of the fraction of mortality attributable to long-term exposure to current levels of anthropogenic (human-made) particulate air pollution range from around 2.5% in some local authorities in rural areas of Scotland and Northern Ireland and between 3 and 5% in Wales, to over 8% in some London boroughs. Because of uncertainty in the increase in mortality risk associated with ambient PM2.5, the actual burdens associated with these modelled concentrations could range from approximately one-sixth to about double these figures. Thus, current levels of particulate air pollution have a considerable impact on public health. Measures to reduce levels of particulate air pollution, or to reduce exposure of the population to such pollution, are regarded as an important public health initiative.

Les délits commis par les employés dans les secteurs commercial et financier : certaines formes de la criminalité économique à la lumière d'une étude de victimisation et de prévention dans les entreprises du canton de Genève

Summary: Victims of crime are not only physical persons. Firms also experience some different forms of victimization either from outsiders or front employees. Crimes committed inside the firm are various: theft, embezzlement, fraud, fraudulent management, money laundering, vandalism etc. Assessed damage could be considerable. But in general, most of the criminological studies relating to victimization have been interested in individuals as victims. In Switzerland, to date, research on firms'victimization is simply rare and traditional research methods such as victimization survey, self-reported delinquency or controlled experiments are not very much utilized in Swiss white-collar crime studies. Thus, in order to improve the situation, we have decided to start a victimization study on white-collar crime committed in firms of the canton of Geneva. To narrow the field of research, our project seeks to apprehend white-collar crime as a series of offences committed by employees inside the firm, by fraud or deceit, aiming at an illegitimate enrichment. The phenomenon of white-collar crime perpetrated has been evaluated by examining the presence of risk factors and prevention strategies. Furthermore, the correlation between the victimization rate and some aspects of vulnerability has also been investigated. Using data collected from the victimization survey carried out in 2005, some of the key research questions are the following: - which are the types of offences committed? - Which types of measure of prevention are the most efficient to prevent such offences ? - Which type of business is more vulnerable? - Which are the risk factors? - Which is the profile of the authors? - Which are the motivations?, etc. Findings show that companies which valorize their employees and pursue a transparent management style suffer a lower victimization rate. Further, the retail sector seems to be more affected than the financial sector and big firms are more vulnerable.

Evaluation des campagnes de prévention contre le SIDA en Suisse : Kantonale Aktivitaeten der Aidspraevention : November 1987

Die Untersuchung verfolgt mehrere Ziele: -> erstens soll der Informationsstand über die Aktivitäten zur Aidsbekämpfung in den Kantonen erhöht werden, -> zweitens sollen Lösungsmodelle bezüglich des Umgangs mit dem Problem vergleichen dargestellt werden, -> drittens sollen Erkenntnisse darüber gesammelt werden, welche Individuen sich als besonders gute Vermittler präventiver Botschaften eignen.

Prévention des troubles lymphatiques et de leurs complications [Lymphatic disorders and prevention of their complications].

The prevalence of lymphedema is clearly underestimated. Too few patients receive treatment. It requires several specifically trained participants and must be conceived in the long term given the chronic nature and the incurability of this pathology. Prevention is therefore of major importance. Successfully applied to operated women for breast cancer, other models of coverage deserve to be developed to reduce the incidence of lymphedema and its complications, particularly after oncologic, orthopedic and vascular surgery and for patients affected by venous insufficiency.

La mort subite: de l' épidémiologie à la prévention [Sudden cardiac death: epidemiology and modern therapy].

Sudden Cardiac Death (SCD) has become an important public health challenge in the Western World. In Switzerland near 10,000 people suffer each year from SCD. The survival from SCD to hospital discharge is discouraging (near 5%). Large majority of events occur unexpectedly in the out-of-hospital environment and are not predicted with great accuracy by risk profiling. Because the majority of SCD occur by the mechanism of ventricular fibrillation, community-based defibrillation strategies have emerged as one approach to SCD problem. Newer strategies of defibrillation designed to respond faster to out-of-hospital cardiac arrest, including public access defibrillation, as well as aggressive primary and secondary prevention of coronary artery disease appears as the best approach for successful management of SCD.

Evaluation des campagnes de prévention contre le SIDA en Suisse : Analyse der Medienreporte : November 1987

Die Analyse der Medienreporte untersucht vor allem Prozesse wie die Verbreitung der Botschaft der Kampagne STOP AIDS durch die Medien, die Verbreitung von informellen und institutionellen Aktionen (zum Beispiel die Aktivitäten zur Aufklärung in den Schulen), die Rolle der Multiplikatoren (Aids-Hilfe, Informanten an öffentlichen Veranstaltungen), den Meinungsbildungsprozess zur Aidsproblematik im allgemeinen und zur Präventionsstrategie im besonderen, soweit dieser in den Medien abgebildet, beziehungsweise von den Medien getragen wird. Im weiteren beachtet die Analyse Rückmeldungen über die Effekte (Resultate) der Kampagne (Beurteilung der Wirkung der Kampagne, zum Beispiel auf das Verhalten von Jugendlichen, Fixern). Die Analyse erfolgt mit Hilfe eines Kategorienschemas. [Autor, p. 2]

Evaluation des campagnes de prévention contre le SIDA en Suisse : "Les migrants" : rapport final 1988

Les auteurs ont estimé nécessaire de mieux comprendre, d'une part les différents états de connaissance, d'attitude et de comportement des populations d'immigrés en Suisse par rapport au phénomène du SIDA et à sa prévention, et d'autre part, d'évaluer si ces états peuvent être mis en relation aux cultures d'origine ainsi qu'à des phénomènes de migration. Les auteurs ont choisi pour cette étude les communautés turque et africaine.