Analysis, modelling and classification of geospatial data using machine learning

The research considers the problem of spatial data classification using machine learning algorithms: probabilistic neural networks (PNN) and support vector machines (SVM). As a benchmark model simple k-nearest neighbor algorithm is considered. PNN is a neural network reformulation of well known nonparametric principles of probability density modeling using kernel density estimator and Bayesian optimal or maximum a posteriori decision rules. PNN is well suited to problems where not only predictions but also quantification of accuracy and integration of prior information are necessary. An important property of PNN is that they can be easily used in decision support systems dealing with problems of automatic classification. Support vector machine is an implementation of the principles of statistical learning theory for the classification tasks. Recently they were successfully applied for different environmental topics: classification of soil types and hydro-geological units, optimization of monitoring networks, susceptibility mapping of natural hazards. In the present paper both simulated and real data case studies (low and high dimensional) are considered. The main attention is paid to the detection and learning of spatial patterns by the algorithms applied.

Seizure semiology: an overview of the 'inverse problem'.

In clinical practice, a classification of seizures based on clinical signs and symptoms leads to an improved understanding of epilepsy-related issues and therefore strongly contributes to a better patient care. The inverse problem involves inferring the anatomical brain localization of a seizure from the scalp surface EEG, a concept we apply here to correlate seizure origin with seizure semiology. The spheres of sensorium, motor features, consciousness changes and autonomic alterations during ictal and postictal manifestations are reviewed, including several subdivisions used to better categorize particular features. Particular attention is given to behavioral features, as well as to features occurring in idiopathic generalized epileptic syndromes and psychogenic nonepileptic spells.

External dichotomous noise: The problem of the mean-first-passage time

A retarded backward equation for a non-Markovian process induced by dichotomous noise (the random telegraphic signal) is deduced. The mean-first-passage time of this process is exactly obtained. The Gaussian white noise and the white shot noise limits are studied. Explicit physical results in first approximation are evaluated.

MADAP, a flexible clustering tool for the interpretation of one-dimensional genome annotation data.

A recurring task in the analysis of mass genome annotation data from high-throughput technologies is the identification of peaks or clusters in a noisy signal profile. Examples of such applications are the definition of promoters on the basis of transcription start site profiles, the mapping of transcription factor binding sites based on ChIP-chip data and the identification of quantitative trait loci (QTL) from whole genome SNP profiles. Input to such an analysis is a set of genome coordinates associated with counts or intensities. The output consists of a discrete number of peaks with respective volumes, extensions and center positions. We have developed for this purpose a flexible one-dimensional clustering tool, called MADAP, which we make available as a web server and as standalone program. A set of parameters enables the user to customize the procedure to a specific problem. The web server, which returns results in textual and graphical form, is useful for small to medium-scale applications, as well as for evaluation and parameter tuning in view of large-scale applications, requiring a local installation. The program written in C++ can be freely downloaded from ftp://ftp.epd.unil.ch/pub/software/unix/madap. The MADAP web server can be accessed at http://www.isrec.isb-sib.ch/madap/.

Zenith angle distributions at Super-Kamiokande and SNO and the solution of the solar neutrino problem

We have performed a detailed study of the zenith angle dependence of the regeneration factor and distributions of events at SNO and SK for different solutions of the solar neutrino problem. In particular, we discuss the oscillatory behavior and the synchronization effect in the distribution for the LMA solution, the parametric peak for the LOW solution, etc. A physical interpretation of the effects is given. We suggest a new binning of events which emphasizes the distinctive features of the zenith angle distributions for the different solutions. We also find the correlations between the integrated day-night asymmetry and the rates of events in different zenith angle bins. The study of these correlations strengthens the identification power of the analysis.

Komik, Ernst und Mise en page. Zum Problem der Farblinien in Heinrich Wittenwilers "Der Ring"

Der Aufsatz nähert sich dem Interpretationsproblem des ,Ring', eines seit Forschungsbeginn ausserordentlich konträr diskutierten Textes, aus wissenschaftsanalytischer Perspektive. Er führt die Zersplitterung und die Extremisierung der Deutungsansätze, die die breite Forschung zu Wittenwilers Werk kennzeichnet, auf einen vergleichsweise simplen Befund zurück: die ungeklärte Frage nach dem Verhältnis von Komik und Didaktik und der Ernsthaftigkeit des lehrhaften Angebots im ,Ring'. Die so auf grundlegendster Ebene offen gebliebene Erfassung des Textes führt in einen interpretatorischen Leerraum, in dem die widersprüchlichsten, oft wissenschaftsideologisch begründeten Positionen Platz finden. Ihre Analyse zeigt, dass sie sich in wiederkehrende Gruppierungen ordnen. Als Schlüsselglied sowohl der Befunddaten als auch der konträren Vernetzungen von Befund und Deutung wird die wechselnd rote und grüne Initialenlinie der Münchner Handschrift identifiziert. Als im Prolog eingeführte Markierung von Ernst auf der einen und Komik auf der anderen Seite ist sie nicht nur visueller Ausdruck des Deutungsproblems des ,Ring', sondern sie hat es wegen der scheinbaren Unstimmigkeit ihrer Zuweisungen auch wesentlich bedingt. Der Aufsatz zeigt ihren bisher übersehenen Aufschluss für das Verständnis des Werks. Nimmt man die Linie als graphisches Verfahren ernst, stellt sie sich als strukturelle Markierung zum Auffinden von Text im diskontinuierlichen Zugriff dar - nicht aber, wie man sie bisher ausnahmslos verstand, als semantische Kommentierung eines laufenden, kontinuierlich zu lesenden Textes. Die strukturierende Funktion des Farbwechsels ist auf zwei Ebenen nachzuweisen: Auf Makroebene trennen seine Zuweisungen vorwiegend narrative und vorwiegend wissensvermittelnde Grosspartien des ,Ring' durch eine jeweilige Grundfarbe. Auf Mikroebene wird diese Grundfarbe durch die jeweils andere Farbe durchbrochen, um formale Einschnitte wie etwa Sprecherwechsel, Ortswechsel, Handlungsneueinsätze, Beginn und Ende eines eingeschalteten Binnentextes oder einer Sentenz anzuzeigen. Dem Benutzer der autornahen Handschrift sollte so in einem ersten Schritt (makrostrukturell) ermöglicht werden, gezielt auf gewünschte Stellen des Textes zuzugreifen und ihn in einem zweiten Schritt (mikrostrukturell) schneller zu erfassen. Nicht in seiner Umsetzung, sehr wohl aber in seiner Funktion steht dieses ungewöhnliche Layout zeitgenössischen Techniken der Buchgliederung durchaus nahe. Auch im Profil der Zusammenstellung seiner Binnentexte rückt der ,Ring' damit in die Nähe der im Spätmittelalter beliebten Sammelhandschriften mit Ziel einer Wissenssumme, die hier narrativ verbundenen wird. Es scheint, als möchte der ,Ring' in einem ungewöhnlichen Experiment beides sein: Kompilation und Werkganzes, Wissenssammlung und Erzählwerk - Texte und Text. Als Ergebnis einer Neuuntersuchung des Münchner codex unicus des ,Ring' weist der Beitrag schliesslich darauf hin, dass auch zahlreiche Markierungszeichen am Spaltenrand formale und strukturelle Texteinschnitte anzeigen und somit die gleiche Funktion haben wie der Farbwechsel auf Mikroebene. Ein kausaler, produktionstechnischer Zusammenhang zwischen diesen Zeichen und der Verteilung des Farbwechsels ist nicht auszuschliessen, zumal beide Verfahren etwa im letzten Fünftel des Textes zunehmend zusammenfallen. Möglicherweise war nur die makrostrukturelle Funktion des Farbwechsel ursprünglich indendiert und mit der Erläuterung im Prolog bezeichnet, während seine mikrostrukturelle Funktion erst Resultat des Abschreibprozesses ist.

Mapping, organic matter mass and water volume of a peatland in Serra do Espinhaço Meridional

Peatlands form in areas where net primary of organic matter production exceeds losses due to the decomposition, leaching or disturbance. Due to their chemical and physical characteristics, bogs can influence water dynamics because they can store large volumes of water in the rainy season and gradually release this water during the other months of the year. In Diamantina, Minas Gerais, Brazil, a peatland in the environmental protection area of Pau-de-Fruta ensures the water supply of 40,000 inhabitants. The hypothesis of this study is that the peat bogs in Pau-de-Fruta act as an environment for carbon storage and a regulator of water flow in the Córrego das Pedras basin. The objective of this study was to estimate the water volume and organic matter mass in this peatland and to study the influence of this environment on the water flow in the Córrego das Pedras basin. The peatland was mapped using 57 transects, at intervals of 100 m. Along all transects, the depth of the peat bog, the Universal Transverse Mercator (UTM) coordinates and altitude were recorded every 20 m and used to calculate the area and volume of the peatland. The water volume was estimated, using a method developed in this study, and the mass of organic matter based on samples from 106 profiles. The peatland covered 81.7 hectares (ha), and stored 497,767 m³ of water, representing 83.7 % of the total volume of the peat bog. The total amount of organic matter (OM) was 45,148 t, corresponding to 552 t ha-1 of OM. The peat bog occupies 11.9 % of the area covered by the Córrego das Pedras basin and stores 77.6 % of the annual water surplus, thus controlling the water flow in the basin and consequently regulating the water course.

A test of Jessor's problem behavior theory in a Eurasian and a Western European developmental context.

PURPOSE: The current study tested the applicability of Jessor's problem behavior theory (PBT) in national probability samples from Georgia and Switzerland. Comparisons focused on (1) the applicability of the problem behavior syndrome (PBS) in both developmental contexts, and (2) on the applicability of employing a set of theory-driven risk and protective factors in the prediction of problem behaviors. METHODS: School-based questionnaire data were collected from n = 18,239 adolescents in Georgia (n = 9499) and Switzerland (n = 8740) following the same protocol. Participants rated five measures of problem behaviors (alcohol and drug use, problems because of alcohol and drug use, and deviance), three risk factors (future uncertainty, depression, and stress), and three protective factors (family, peer, and school attachment). Final study samples included n = 9043 Georgian youth (mean age = 15.57; 58.8% females) and n = 8348 Swiss youth (mean age = 17.95; 48.5% females). Data analyses were completed using structural equation modeling, path analyses, and post hoc z-tests for comparisons of regression coefficients. RESULTS: Findings indicated that the PBS replicated in both samples, and that theory-driven risk and protective factors accounted for 13% and 10% in Georgian and Swiss samples, respectively in the PBS, net the effects by demographic variables. Follow-up z-tests provided evidence of some differences in the magnitude, but not direction, in five of six individual paths by country. CONCLUSION: PBT and the PBS find empirical support in these Eurasian and Western European samples; thus, Jessor's theory holds value and promise in understanding the etiology of adolescent problem behaviors outside of the United States.

Reply to Comment on 'Two-finger selection theory in the Saffman-Taylor problem

We clarify the meaning of the results of Phys. Rev. E 60, R5013 (1999). We discuss the use and implications of periodic boundary conditions, as opposed to rigid-wall ones. We briefly argue that the solutions of the paper above are physically relevant as part of a more general issue, namely the possible generalization to dynamics, of the microscopic solvability scenario of selection.

Digital soil mapping: strategy for data pre-processing

The region of greatest variability on soil maps is along the edge of their polygons, causing disagreement among pedologists about the appropriate description of soil classes at these locations. The objective of this work was to propose a strategy for data pre-processing applied to digital soil mapping (DSM). Soil polygons on a training map were shrunk by 100 and 160 m. This strategy prevented the use of covariates located near the edge of the soil classes for the Decision Tree (DT) models. Three DT models derived from eight predictive covariates, related to relief and organism factors sampled on the original polygons of a soil map and on polygons shrunk by 100 and 160 m were used to predict soil classes. The DT model derived from observations 160 m away from the edge of the polygons on the original map is less complex and has a better predictive performance.