Environmental data mining and modeling based on machine learning algorithms and geostatistics

The paper presents some contemporary approaches to spatial environmental data analysis. The main topics are concentrated on the decision-oriented problems of environmental spatial data mining and modeling: valorization and representativity of data with the help of exploratory data analysis, spatial predictions, probabilistic and risk mapping, development and application of conditional stochastic simulation models. The innovative part of the paper presents integrated/hybrid model-machine learning (ML) residuals sequential simulations-MLRSS. The models are based on multilayer perceptron and support vector regression ML algorithms used for modeling long-range spatial trends and sequential simulations of the residuals. NIL algorithms deliver non-linear solution for the spatial non-stationary problems, which are difficult for geostatistical approach. Geostatistical tools (variography) are used to characterize performance of ML algorithms, by analyzing quality and quantity of the spatially structured information extracted from data with ML algorithms. Sequential simulations provide efficient assessment of uncertainty and spatial variability. Case study from the Chernobyl fallouts illustrates the performance of the proposed model. It is shown that probability mapping, provided by the combination of ML data driven and geostatistical model based approaches, can be efficiently used in decision-making process. (C) 2003 Elsevier Ltd. All rights reserved.

Population specific and up to date cardiovascular risk charts can be efficiently obtained with record linkage of routine and observational data.

BACKGROUND: Only few countries have cohorts enabling specific and up-to-date cardiovascular disease (CVD) risk estimation. Individual risk assessment based on study samples that differ too much from the target population could jeopardize the benefit of risk charts in general practice. Our aim was to provide up-to-date and valid CVD risk estimation for a Swiss population using a novel record linkage approach. METHODS: Anonymous record linkage was used to follow-up (for mortality, until 2008) 9,853 men and women aged 25-74 years who participated in the Swiss MONICA (MONItoring of trends and determinants in CVD) study of 1983-92. The linkage success was 97.8%, loss to follow-up 1990-2000 was 4.7%. Based on the ESC SCORE methodology (Weibull regression), we used age, sex, blood pressure, smoking, and cholesterol to generate three models. We compared the 1) original SCORE model with a 2) recalibrated and a 3) new model using the Brier score (BS) and cross-validation. RESULTS: Based on the cross-validated BS, the new model (BS = 14107×10(-6)) was somewhat more appropriate for risk estimation than the original (BS = 14190×10(-6)) and the recalibrated (BS = 14172×10(-6)) model. Particularly at younger age, derived absolute risks were consistently lower than those from the original and the recalibrated model which was mainly due to a smaller impact of total cholesterol. CONCLUSION: Using record linkage of observational and routine data is an efficient procedure to obtain valid and up-to-date CVD risk estimates for a specific population.

Investigation of discrete imaging models and iterative image reconstruction in differential X-ray phase-contrast tomography.

Differential X-ray phase-contrast tomography (DPCT) refers to a class of promising methods for reconstructing the X-ray refractive index distribution of materials that present weak X-ray absorption contrast. The tomographic projection data in DPCT, from which an estimate of the refractive index distribution is reconstructed, correspond to one-dimensional (1D) derivatives of the two-dimensional (2D) Radon transform of the refractive index distribution. There is an important need for the development of iterative image reconstruction methods for DPCT that can yield useful images from few-view projection data, thereby mitigating the long data-acquisition times and large radiation doses associated with use of analytic reconstruction methods. In this work, we analyze the numerical and statistical properties of two classes of discrete imaging models that form the basis for iterative image reconstruction in DPCT. We also investigate the use of one of the models with a modern image reconstruction algorithm for performing few-view image reconstruction of a tissue specimen.

Sensitivity analysis, dominant factors, and robustness of the ECETOC TRA v3, Stoffenmanager 4.5, and ART 1.5 occupational exposure models

Occupational exposure modeling is widely used in the context of the E.U. regulation on the registration, evaluation, authorization, and restriction of chemicals (REACH). First tier tools, such as European Centre for Ecotoxicology and TOxicology of Chemicals (ECETOC) targeted risk assessment (TRA) or Stoffenmanager, are used to screen a wide range of substances. Those of concern are investigated further using second tier tools, e.g., Advanced REACH Tool (ART). Local sensitivity analysis (SA) methods are used here to determine dominant factors for three models commonly used within the REACH framework: ECETOC TRA v3, Stoffenmanager 4.5, and ART 1.5. Based on the results of the SA, the robustness of the models is assessed. For ECETOC, the process category (PROC) is the most important factor. A failure to identify the correct PROC has severe consequences for the exposure estimate. Stoffenmanager is the most balanced model and decision making uncertainties in one modifying factor are less severe in Stoffenmanager. ART requires a careful evaluation of the decisions in the source compartment since it constitutes ∼75% of the total exposure range, which corresponds to an exposure estimate of 20-22 orders of magnitude. Our results indicate that there is a trade off between accuracy and precision of the models. Previous studies suggested that ART may lead to more accurate results in well-documented exposure situations. However, the choice of the adequate model should ultimately be determined by the quality of the available exposure data: if the practitioner is uncertain concerning two or more decisions in the entry parameters, Stoffenmanager may be more robust than ART.

Evolutionary Games in Networked Populations: Models and Experiments

Cooperation and coordination are desirable behaviors that are fundamental for the harmonious development of society. People need to rely on cooperation with other individuals in many aspects of everyday life, such as teamwork and economic exchange in anonymous markets. However, cooperation may easily fall prey to exploitation by selfish individuals who only care about short- term gain. For cooperation to evolve, specific conditions and mechanisms are required, such as kinship, direct and indirect reciprocity through repeated interactions, or external interventions such as punishment. In this dissertation we investigate the effect of the network structure of the population on the evolution of cooperation and coordination. We consider several kinds of static and dynamical network topologies, such as Baraba´si-Albert, social network models and spatial networks. We perform numerical simulations and laboratory experiments using the Prisoner's Dilemma and co- ordination games in order to contrast human behavior with theoretical results. We show by numerical simulations that even a moderate amount of random noise on the Baraba´si-Albert scale-free network links causes a significant loss of cooperation, to the point that cooperation almost vanishes altogether in the Prisoner's Dilemma when the noise rate is high enough. Moreover, when we consider fixed social-like networks we find that current models of social networks may allow cooperation to emerge and to be robust at least as much as in scale-free networks. In the framework of spatial networks, we investigate whether cooperation can evolve and be stable when agents move randomly or performing Le´vy flights in a continuous space. We also consider discrete space adopting purposeful mobility and binary birth-death process to dis- cover emergent cooperative patterns. The fundamental result is that cooperation may be enhanced when this migration is opportunistic or even when agents follow very simple heuristics. In the experimental laboratory, we investigate the issue of social coordination between indi- viduals located on networks of contacts. In contrast to simulations, we find that human players dynamics do not converge to the efficient outcome more often in a social-like network than in a random network. In another experiment, we study the behavior of people who play a pure co- ordination game in a spatial environment in which they can move around and when changing convention is costly. We find that each convention forms homogeneous clusters and is adopted by approximately half of the individuals. When we provide them with global information, i.e., the number of subjects currently adopting one of the conventions, global consensus is reached in most, but not all, cases. Our results allow us to extract the heuristics used by the participants and to build a numerical simulation model that agrees very well with the experiments. Our findings have important implications for policymakers intending to promote specific, desired behaviors in a mobile population. Furthermore, we carry out an experiment with human subjects playing the Prisoner's Dilemma game in a diluted grid where people are able to move around. In contrast to previous results on purposeful rewiring in relational networks, we find no noticeable effect of mobility in space on the level of cooperation. Clusters of cooperators form momentarily but in a few rounds they dissolve as cooperators at the boundaries stop tolerating being cheated upon. Our results highlight the difficulties that mobile agents have to establish a cooperative environment in a spatial setting without a device such as reputation or the possibility of retaliation. i.e. punishment. Finally, we test experimentally the evolution of cooperation in social networks taking into ac- count a setting where we allow people to make or break links at their will. In this work we give particular attention to whether information on an individual's actions is freely available to poten- tial partners or not. Studying the role of information is relevant as information on other people's actions is often not available for free: a recruiting firm may need to call a job candidate's refer- ences, a bank may need to find out about the credit history of a new client, etc. We find that people cooperate almost fully when information on their actions is freely available to their potential part- ners. Cooperation is less likely, however, if people have to pay about half of what they gain from cooperating with a cooperator. Cooperation declines even further if people have to pay a cost that is almost equivalent to the gain from cooperating with a cooperator. Thus, costly information on potential neighbors' actions can undermine the incentive to cooperate in dynamical networks.

Privacy-friendly Business Models for Location-Based Mobile Services

This paper presents a theoretical model to analyze the privacy issues around location based mobile business models. We report the results of an exploratory field experiment in Switzerland that assessed the factors driving user payoff in mobile business. We found that (1) the personal data disclosed has a negative effect on user payoff; (2) the amount of personalization available has a direct and positive effect, as well as a moderating effect on user payoff; (3) the amount of control over user's personal data has a direct and positive effect, as well as a moderating effect on user payoff. The results suggest that privacy protection could be the main value proposition in the B2C mobile market. From our theoretical model we derive a set of guidelines to design a privacy-friendly business model pattern for third-party services. We discuss four examples to show the mobile platform can play a key role in the implementation of these new business models.

Data Envelopment Analysis and non-discretionary inputs : How to select the most suitable model using multi-criteria decision analysis

Within Data Envelopment Analysis, several alternative models allow for an environmental adjustment. The majority of them deliver divergent results. Decision makers face the difficult task of selecting the most suitable model. This study is performed to overcome this difficulty. By doing so, it fills a research gap. First, a two-step web-based survey is conducted. It aims (1) to identify the selection criteria, (2) to prioritize and weight the selection criteria with respect to the goal of selecting the most suitable model and (3) to collect the preferences about which model is preferable to fulfil each selection criterion. Second, Analytic Hierarchy Process is used to quantify the preferences expressed in the survey. Results show that the understandability, the applicability and the acceptability of the alternative models are valid selection criteria. The selection of the most suitable model depends on the preferences of the decision makers with regards to these criteria.

Two-dimensional probabilistic inversion of plane-wave electromagnetic data: Methodology, model constraints and joint inversion with electrical resistivity data

Probabilistic inversion methods based on Markov chain Monte Carlo (MCMC) simulation are well suited to quantify parameter and model uncertainty of nonlinear inverse problems. Yet, application of such methods to CPU-intensive forward models can be a daunting task, particularly if the parameter space is high dimensional. Here, we present a 2-D pixel-based MCMC inversion of plane-wave electromagnetic (EM) data. Using synthetic data, we investigate how model parameter uncertainty depends on model structure constraints using different norms of the likelihood function and the model constraints, and study the added benefits of joint inversion of EM and electrical resistivity tomography (ERT) data. Our results demonstrate that model structure constraints are necessary to stabilize the MCMC inversion results of a highly discretized model. These constraints decrease model parameter uncertainty and facilitate model interpretation. A drawback is that these constraints may lead to posterior distributions that do not fully include the true underlying model, because some of its features exhibit a low sensitivity to the EM data, and hence are difficult to resolve. This problem can be partly mitigated if the plane-wave EM data is augmented with ERT observations. The hierarchical Bayesian inverse formulation introduced and used herein is able to successfully recover the probabilistic properties of the measurement data errors and a model regularization weight. Application of the proposed inversion methodology to field data from an aquifer demonstrates that the posterior mean model realization is very similar to that derived from a deterministic inversion with similar model constraints.

Age and adverse drug reactions from psychopharmacological treatment: data from the AMSP drug surveillance programme in Switzerland.

QUESTION UNDER STUDY: The frequency of severe adverse drug reactions (ADRs) from psychotropic drugs was investigated in hospitalised psychiatric patients in relation to their age. Specifically, the incidence of ADRs in patients up to 60 years was compared to that of patients older than 60 years. METHODS: Prescription rates of psychotropic drugs and reports of severe ADRs were collected in psychiatric hospitals in Switzerland between 2001 and 2010. The data stem from the drug surveillance programme AMSP. RESULTS: A total of 699 patients exhibited severe ADRs: 517 out of 28,282 patients up to 60 years (1.8%); 182 out of 11,446 elderly patients (1.6%, ns). Logistic regression analyses showed a significantly negative relationship between the incidence of ADRs and patients' age in general and in particular for weight gain, extrapyramidal motor system (EPMS) symptoms, increased liver enzymes and galactorrhoea. A significantly negative relationship was observed for age and the dosages of olanzapine, quetiapine, risperidone, valproic acid and lamotrigine. When comparing age groups, frequency of ADRs was lower in general for antipsychotic drugs and anticonvulsants, in particular for valproic acid in the elderly. Weight gain was found to be lower in the elderly for antipsychotic drugs, in particular for olanzapine. For the group of mood-stabilising anticonvulsants (carbamazepine, lamotrigine and valproic acid) the elderly exhibited a lower incidence of reported allergic skin reactions. CONCLUSION: The results suggest that for psychiatric inpatients the incidence of common severe ADRs (e.g., weight gain or EPMS symptoms) arising from psychotropic medication decreases with the age of patients.

Reply to 'Sources of uncertainties in cod distribution models'

Ingvaldsen et al. comment on our study assessing global fish interchanges between the North Atlantic and Pacific oceans for more than 500 species during the entire 21st century. They propose that discrepancies between our model projections and observed data for cod in the Barents Sea are the result of the choice of Atmosphere-Ocean General Circulation Models (AOGCMs). We address this assertion here, re-running the cod model with additional observation data from the Barents Sea1, 3, and show that the lack of open-access, archived data for the Barents Sea was the primary cause of local prediction mismatch. This finding recalls the importance of systematic deposit of biodiversity data in global databases

Biotic interactions boost spatial models of species richness

Biotic interactions are known to affect the composition of species assemblages via several mechanisms, such as competition and facilitation. However, most spatial models of species richness do not explicitly consider inter-specific interactions. Here, we test whether incorporating biotic interactions into high-resolution models alters predictions of species richness as hypothesised. We included key biotic variables (cover of three dominant arctic-alpine plant species) into two methodologically divergent species richness modelling frameworks - stacked species distribution models (SSDM) and macroecological models (MEM) - for three ecologically and evolutionary distinct taxonomic groups (vascular plants, bryophytes and lichens). Predictions from models including biotic interactions were compared to the predictions of models based on climatic and abiotic data only. Including plant-plant interactions consistently and significantly lowered bias in species richness predictions and increased predictive power for independent evaluation data when compared to the conventional climatic and abiotic data based models. Improvements in predictions were constant irrespective of the modelling framework or taxonomic group used. The global biodiversity crisis necessitates accurate predictions of how changes in biotic and abiotic conditions will potentially affect species richness patterns. Here, we demonstrate that models of the spatial distribution of species richness can be improved by incorporating biotic interactions, and thus that these key predictor factors must be accounted for in biodiversity forecasts

Assessing the use of very high resolution data to predict species distribution in a mountain environment

Nowadays, Species Distribution Models (SDMs) are a widely used tool. Using different statistical approaches these models reconstruct the realized niche of a species using presence data and a set of variables, often topoclimatic. There utilization range is quite large from understanding single species requirements, to the creation of nature reserve based on species hotspots, or modeling of climate change impact, etc... Most of the time these models are using variables at a resolution of 50km x 50km or 1 km x 1 km. However in some cases these models are used with resolutions below the kilometer scale and thus called high resolution models (100 m x 100 m or 25 m x 25 m). Quite recently a new kind of data has emerged enabling precision up to lm x lm and thus allowing very high resolution modeling. However these new variables are very costly and need an important amount of time to be processed. This is especially the case when these variables are used in complex calculation like models projections over large areas. Moreover the importance of very high resolution data in SDMs has not been assessed yet and is not well understood. Some basic knowledge on what drive species presence-absences is still missing. Indeed, it is not clear whether in mountain areas like the Alps coarse topoclimatic gradients are driving species distributions or if fine scale temperature or topography are more important or if their importance can be neglected when balance to competition or stochasticity. In this thesis I investigated the importance of very high resolution data (2-5m) in species distribution models using either very high resolution topographic, climatic or edaphic variables over a 2000m elevation gradient in the Western Swiss Alps. I also investigated more local responses of these variables for a subset of species living in this area at two precise elvation belts. During this thesis I showed that high resolution data necessitates very good datasets (species and variables for the models) to produce satisfactory results. Indeed, in mountain areas, temperature is the most important factor driving species distribution and needs to be modeled at very fine resolution instead of being interpolated over large surface to produce satisfactory results. Despite the instinctive idea that topographic should be very important at high resolution, results are mitigated. However looking at the importance of variables over a large gradient buffers the importance of the variables. Indeed topographic factors have been shown to be highly important at the subalpine level but their importance decrease at lower elevations. Wether at the mountane level edaphic and land use factors are more important high resolution topographic data is more imporatant at the subalpine level. Finally the biggest improvement in the models happens when edaphic variables are added. Indeed, adding soil variables is of high importance and variables like pH are overpassing the usual topographic variables in SDMs in term of importance in the models. To conclude high resolution is very important in modeling but necessitate very good datasets. Only increasing the resolution of the usual topoclimatic predictors is not sufficient and the use of edaphic predictors has been highlighted as fundamental to produce significantly better models. This is of primary importance, especially if these models are used to reconstruct communities or as basis for biodiversity assessments. -- Ces dernières années, l'utilisation des modèles de distribution d'espèces (SDMs) a continuellement augmenté. Ces modèles utilisent différents outils statistiques afin de reconstruire la niche réalisée d'une espèce à l'aide de variables, notamment climatiques ou topographiques, et de données de présence récoltées sur le terrain. Leur utilisation couvre de nombreux domaines allant de l'étude de l'écologie d'une espèce à la reconstruction de communautés ou à l'impact du réchauffement climatique. La plupart du temps, ces modèles utilisent des occur-rences issues des bases de données mondiales à une résolution plutôt large (1 km ou même 50 km). Certaines bases de données permettent cependant de travailler à haute résolution, par conséquent de descendre en dessous de l'échelle du kilomètre et de travailler avec des résolutions de 100 m x 100 m ou de 25 m x 25 m. Récemment, une nouvelle génération de données à très haute résolution est apparue et permet de travailler à l'échelle du mètre. Les variables qui peuvent être générées sur la base de ces nouvelles données sont cependant très coûteuses et nécessitent un temps conséquent quant à leur traitement. En effet, tout calcul statistique complexe, comme des projections de distribution d'espèces sur de larges surfaces, demande des calculateurs puissants et beaucoup de temps. De plus, les facteurs régissant la distribution des espèces à fine échelle sont encore mal connus et l'importance de variables à haute résolution comme la microtopographie ou la température dans les modèles n'est pas certaine. D'autres facteurs comme la compétition ou la stochasticité naturelle pourraient avoir une influence toute aussi forte. C'est dans ce contexte que se situe mon travail de thèse. J'ai cherché à comprendre l'importance de la haute résolution dans les modèles de distribution d'espèces, que ce soit pour la température, la microtopographie ou les variables édaphiques le long d'un important gradient d'altitude dans les Préalpes vaudoises. J'ai également cherché à comprendre l'impact local de certaines variables potentiellement négligées en raison d'effets confondants le long du gradient altitudinal. Durant cette thèse, j'ai pu monter que les variables à haute résolution, qu'elles soient liées à la température ou à la microtopographie, ne permettent qu'une amélioration substantielle des modèles. Afin de distinguer une amélioration conséquente, il est nécessaire de travailler avec des jeux de données plus importants, tant au niveau des espèces que des variables utilisées. Par exemple, les couches climatiques habituellement interpolées doivent être remplacées par des couches de température modélisées à haute résolution sur la base de données de terrain. Le fait de travailler le long d'un gradient de température de 2000m rend naturellement la température très importante au niveau des modèles. L'importance de la microtopographie est négligeable par rapport à la topographie à une résolution de 25m. Cependant, lorsque l'on regarde à une échelle plus locale, la haute résolution est une variable extrêmement importante dans le milieu subalpin. À l'étage montagnard par contre, les variables liées aux sols et à l'utilisation du sol sont très importantes. Finalement, les modèles de distribution d'espèces ont été particulièrement améliorés par l'addition de variables édaphiques, principalement le pH, dont l'importance supplante ou égale les variables topographique lors de leur ajout aux modèles de distribution d'espèces habituels.

Probabilistic graphical models to deal with age estimation of living persons

Due to the rise of criminal, civil and administrative judicial situations involving people lacking valid identity documents, age estimation of living persons has become an important operational procedure for numerous forensic and medicolegal services worldwide. The chronological age of a given person is generally estimated from the observed degree of maturity of some selected physical attributes by means of statistical methods. However, their application in the forensic framework suffers from some conceptual and practical drawbacks, as recently claimed in the specialised literature. The aim of this paper is therefore to offer an alternative solution for overcoming these limits, by reiterating the utility of a probabilistic Bayesian approach for age estimation. This approach allows one to deal in a transparent way with the uncertainty surrounding the age estimation process and to produce all the relevant information in the form of posterior probability distribution about the chronological age of the person under investigation. Furthermore, this probability distribution can also be used for evaluating in a coherent way the possibility that the examined individual is younger or older than a given legal age threshold having a particular legal interest. The main novelty introduced by this work is the development of a probabilistic graphical model, i.e. a Bayesian network, for dealing with the problem at hand. The use of this kind of probabilistic tool can significantly facilitate the application of the proposed methodology: examples are presented based on data related to the ossification status of the medial clavicular epiphysis. The reliability and the advantages of this probabilistic tool are presented and discussed.

Evaluation of drug-induced neurotoxicity based on metabolomics, proteomics and electrical activity measurements in complementary CNS in vitro models.

The present study was performed in an attempt to develop an in vitro integrated testing strategy (ITS) to evaluate drug-induced neurotoxicity. A number of endpoints were analyzed using two complementary brain cell culture models and an in vitro blood-brain barrier (BBB) model after single and repeated exposure treatments with selected drugs that covered the major biological, pharmacological and neuro-toxicological responses. Furthermore, four drugs (diazepam, cyclosporine A, chlorpromazine and amiodarone) were tested more in depth as representatives of different classes of neurotoxicants, inducing toxicity through different pathways of toxicity. The developed in vitro BBB model allowed detection of toxic effects at the level of BBB and evaluation of drug transport through the barrier for predicting free brain concentrations of the studied drugs. The measurement of neuronal electrical activity was found to be a sensitive tool to predict the neuroactivity and neurotoxicity of drugs after acute exposure. The histotypic 3D re-aggregating brain cell cultures, containing all brain cell types, were found to be well suited for OMICs analyses after both acute and long term treatment. The obtained data suggest that an in vitro ITS based on the information obtained from BBB studies and combined with metabolomics, proteomics and neuronal electrical activity measurements performed in stable in vitro neuronal cell culture systems, has high potential to improve current in vitro drug-induced neurotoxicity evaluation.

Digital evidence, 'absence' of data and ambiguous patterns of reasoning

In this paper we discuss the use of digital data by the Swiss Federal Criminal Court in a recent case of attempted homicide. We use this case to examine drawbacks for the defense when the presentation of scientific evidence is partial, especially when the only perspective mentioned is that of the prosecution. We tackle this discussion at two distinct levels. First, we pursue an essentially non-technical presentation of the topic by drawing parallels between the court's summing up of the case and flawed patterns of reasoning commonly seen in other forensic disciplines, such as DNA and particle traces (e.g., gunshot residues). Then, we propose a formal analysis of the case, using elements of probability and graphical probability models, to justify our main claim that the partial presentation of digital evidence poses a risk to the administration of justice in that it keeps vital information from the defense. We will argue that such practice constitutes a violation of general principles of forensic interpretation as established by forensic science literature and current recommendations by forensic science interest groups (e.g., the European Network of Forensic Science Institutes). Finally, we posit that argument construction and analysis using formal methods can help replace digital evidence appropriately into context and thus support a sound evaluation of the evidence.