A partial dependence factorial analysis to deal with selection bias in observational studis

This thesis presents a creative and practical approach to dealing with the problem of selection bias. Selection bias may be the most important vexing problem in program evaluation or in any line of research that attempts to assert causality. Some of the greatest minds in economics and statistics have scrutinized the problem of selection bias, with the resulting approaches – Rubin’s Potential Outcome Approach(Rosenbaum and Rubin,1983; Rubin, 1991,2001,2004) or Heckman’s Selection model (Heckman, 1979) – being widely accepted and used as the best fixes. These solutions to the bias that arises in particular from self selection are imperfect, and many researchers, when feasible, reserve their strongest causal inference for data from experimental rather than observational studies. The innovative aspect of this thesis is to propose a data transformation that allows measuring and testing in an automatic and multivariate way the presence of selection bias. The approach involves the construction of a multi-dimensional conditional space of the X matrix in which the bias associated with the treatment assignment has been eliminated. Specifically, we propose the use of a partial dependence analysis of the X-space as a tool for investigating the dependence relationship between a set of observable pre-treatment categorical covariates X and a treatment indicator variable T, in order to obtain a measure of bias according to their dependence structure. The measure of selection bias is then expressed in terms of inertia due to the dependence between X and T that has been eliminated. Given the measure of selection bias, we propose a multivariate test of imbalance in order to check if the detected bias is significant, by using the asymptotical distribution of inertia due to T (Estadella et al. 2005) , and by preserving the multivariate nature of data. Further, we propose the use of a clustering procedure as a tool to find groups of comparable units on which estimate local causal effects, and the use of the multivariate test of imbalance as a stopping rule in choosing the best cluster solution set. The method is non parametric, it does not call for modeling the data, based on some underlying theory or assumption about the selection process, but instead it calls for using the existing variability within the data and letting the data to speak. The idea of proposing this multivariate approach to measure selection bias and test balance comes from the consideration that in applied research all aspects of multivariate balance, not represented in the univariate variable- by-variable summaries, are ignored. The first part contains an introduction to evaluation methods as part of public and private decision process and a review of the literature of evaluation methods. The attention is focused on Rubin Potential Outcome Approach, matching methods, and briefly on Heckman’s Selection Model. The second part focuses on some resulting limitations of conventional methods, with particular attention to the problem of how testing in the correct way balancing. The third part contains the original contribution proposed , a simulation study that allows to check the performance of the method for a given dependence setting and an application to a real data set. Finally, we discuss, conclude and explain our future perspectives.

Geotechnical characterization of mixed sandy and silty soils using piezocone tests: Analysis of partial drainage phenomena and rate effects on the experimental soil response

The cone penetration test (CPT), together with its recent variation (CPTU), has become the most widely used in-situ testing technique for soil profiling and geotechnical characterization. The knowledge gained over the last decades on the interpretation procedures in sands and clays is certainly wide, whilst very few contributions can be found as regards the analysis of CPT(u) data in intermediate soils. Indeed, it is widely accepted that at the standard rate of penetration (v = 20 mm/s), drained penetration occurs in sands while undrained penetration occurs in clays. However, a problem arise when the available interpretation approaches are applied to cone measurements in silts, sandy silts, silty or clayey sands, since such intermediate geomaterials are often characterized by permeability values within the range in which partial drainage is very likely to occur. Hence, the application of the available and well-established interpretation procedures, developed for ‘standard’ clays and sands, may result in invalid estimates of soil parameters. This study aims at providing a better understanding on the interpretation of CPTU data in natural sand and silt mixtures, by taking into account two main aspects, as specified below: 1)Investigating the effect of penetration rate on piezocone measurements, with the aim of identifying drainage conditions when cone penetration is performed at a standard rate. This part of the thesis has been carried out with reference to a specific CPTU database recently collected in a liquefaction-prone area (Emilia-Romagna Region, Italy). 2)Providing a better insight into the interpretation of piezocone tests in the widely studied silty sediments of the Venetian lagoon (Italy). Research has focused on the calibration and verification of some site-specific correlations, with special reference to the estimate of compressibility parameters for the assessment of long-term settlements of the Venetian coastal defences.

Non-Markovian stochastic processes and their applications: from anomalous diffusion to time series analysis

This work provides a forward step in the study and comprehension of the relationships between stochastic processes and a certain class of integral-partial differential equation, which can be used in order to model anomalous diffusion and transport in statistical physics. In the first part, we brought the reader through the fundamental notions of probability and stochastic processes, stochastic integration and stochastic differential equations as well. In particular, within the study of H-sssi processes, we focused on fractional Brownian motion (fBm) and its discrete-time increment process, the fractional Gaussian noise (fGn), which provide examples of non-Markovian Gaussian processes. The fGn, together with stationary FARIMA processes, is widely used in the modeling and estimation of long-memory, or long-range dependence (LRD). Time series manifesting long-range dependence, are often observed in nature especially in physics, meteorology, climatology, but also in hydrology, geophysics, economy and many others. We deepely studied LRD, giving many real data examples, providing statistical analysis and introducing parametric methods of estimation. Then, we introduced the theory of fractional integrals and derivatives, which indeed turns out to be very appropriate for studying and modeling systems with long-memory properties. After having introduced the basics concepts, we provided many examples and applications. For instance, we investigated the relaxation equation with distributed order time-fractional derivatives, which describes models characterized by a strong memory component and can be used to model relaxation in complex systems, which deviates from the classical exponential Debye pattern. Then, we focused in the study of generalizations of the standard diffusion equation, by passing through the preliminary study of the fractional forward drift equation. Such generalizations have been obtained by using fractional integrals and derivatives of distributed orders. In order to find a connection between the anomalous diffusion described by these equations and the long-range dependence, we introduced and studied the generalized grey Brownian motion (ggBm), which is actually a parametric class of H-sssi processes, which have indeed marginal probability density function evolving in time according to a partial integro-differential equation of fractional type. The ggBm is of course Non-Markovian. All around the work, we have remarked many times that, starting from a master equation of a probability density function f(x,t), it is always possible to define an equivalence class of stochastic processes with the same marginal density function f(x,t). All these processes provide suitable stochastic models for the starting equation. Studying the ggBm, we just focused on a subclass made up of processes with stationary increments. The ggBm has been defined canonically in the so called grey noise space. However, we have been able to provide a characterization notwithstanding the underline probability space. We also pointed out that that the generalized grey Brownian motion is a direct generalization of a Gaussian process and in particular it generalizes Brownain motion and fractional Brownain motion as well. Finally, we introduced and analyzed a more general class of diffusion type equations related to certain non-Markovian stochastic processes. We started from the forward drift equation, which have been made non-local in time by the introduction of a suitable chosen memory kernel K(t). The resulting non-Markovian equation has been interpreted in a natural way as the evolution equation of the marginal density function of a random time process l(t). We then consider the subordinated process Y(t)=X(l(t)) where X(t) is a Markovian diffusion. The corresponding time-evolution of the marginal density function of Y(t) is governed by a non-Markovian Fokker-Planck equation which involves the same memory kernel K(t). We developed several applications and derived the exact solutions. Moreover, we considered different stochastic models for the given equations, providing path simulations.

Innovative analytical methods for Central Nervous System Drug analysis in biological fluids

During recent years a consistent number of central nervous system (CNS) drugs have been approved and introduced on the market for the treatment of many psychiatric and neurological disorders, including psychosis, depression, Parkinson disease and epilepsy. Despite the great advancements obtained in the treatment of CNS diseases/disorders, partial response to therapy or treatment failure are frequent, at least in part due to poor compliance, but also genetic variability in the metabolism of psychotropic agents or polypharmacy, which may lead to sub-therapeutic or toxic plasma levels of the drugs, and finally inefficacy of the treatment or adverse/toxic effects. With the aim of improving the treatment, reducing toxic/side effects and patient hospitalisation, Therapeutic Drug Monitoring (TDM) is certainly useful, allowing for a personalisation of the therapy. Reliable analytical methods are required to determine the plasma levels of psychotropic drugs, which are often present at low concentrations (tens or hundreds of nanograms per millilitre). The present PhD Thesis has focused on the development of analytical methods for the determination of CNS drugs in biological fluids, including antidepressants (sertraline and duloxetine), antipsychotics (aripiprazole), antiepileptics (vigabatrin and topiramate) and antiparkinsons (pramipexole). Innovative methods based on liquid chromatography or capillary electrophoresis coupled to diode-array or laser-induced fluorescence detectors have been developed, together with the suitable sample pre-treatment for interference removal and fluorescent labelling in case of LIF detection. All methods have been validated according to official guidelines and applied to the analysis of real samples obtained from patients, resulting suitable for the TDM of psychotropic drugs.

Biomass Gasification - Process analysis and dimensioning aspects for downdraft units and gas cleaning lines

In such territories where food production is mostly scattered in several small / medium size or even domestic farms, a lot of heterogeneous residues are produced yearly, since farmers usually carry out different activities in their properties. The amount and composition of farm residues, therefore, widely change during year, according to the single production process periodically achieved. Coupling high efficiency micro-cogeneration energy units with easy handling biomass conversion equipments, suitable to treat different materials, would provide many important advantages to the farmers and to the community as well, so that the increase in feedstock flexibility of gasification units is nowadays seen as a further paramount step towards their wide spreading in rural areas and as a real necessity for their utilization at small scale. Two main research topics were thought to be of main concern at this purpose, and they were therefore discussed in this work: the investigation of fuels properties impact on gasification process development and the technical feasibility of small scale gasification units integration with cogeneration systems. According to these two main aspects, the present work was thus divided in two main parts. The first one is focused on the biomass gasification process, that was investigated in its theoretical aspects and then analytically modelled in order to simulate thermo-chemical conversion of different biomass fuels, such as wood (park waste wood and softwood), wheat straw, sewage sludge and refuse derived fuels. The main idea is to correlate the results of reactor design procedures with the physical properties of biomasses and the corresponding working conditions of gasifiers (temperature profile, above all), in order to point out the main differences which prevent the use of the same conversion unit for different materials. At this scope, a gasification kinetic free model was initially developed in Excel sheets, considering different values of air to biomass ratio and the downdraft gasification technology as particular examined application. The differences in syngas production and working conditions (process temperatures, above all) among the considered fuels were tried to be connected to some biomass properties, such elementary composition, ash and water contents. The novelty of this analytical approach was the use of kinetic constants ratio in order to determine oxygen distribution among the different oxidation reactions (regarding volatile matter only) while equilibrium of water gas shift reaction was considered in gasification zone, by which the energy and mass balances involved in the process algorithm were linked together, as well. Moreover, the main advantage of this analytical tool is the easiness by which the input data corresponding to the particular biomass materials can be inserted into the model, so that a rapid evaluation on their own thermo-chemical conversion properties is possible to be obtained, mainly based on their chemical composition A good conformity of the model results with the other literature and experimental data was detected for almost all the considered materials (except for refuse derived fuels, because of their unfitting chemical composition with the model assumptions). Successively, a dimensioning procedure for open core downdraft gasifiers was set up, by the analysis on the fundamental thermo-physical and thermo-chemical mechanisms which are supposed to regulate the main solid conversion steps involved in the gasification process. Gasification units were schematically subdivided in four reaction zones, respectively corresponding to biomass heating, solids drying, pyrolysis and char gasification processes, and the time required for the full development of each of these steps was correlated to the kinetics rates (for pyrolysis and char gasification processes only) and to the heat and mass transfer phenomena from gas to solid phase. On the basis of this analysis and according to the kinetic free model results and biomass physical properties (particles size, above all) it was achieved that for all the considered materials char gasification step is kinetically limited and therefore temperature is the main working parameter controlling this step. Solids drying is mainly regulated by heat transfer from bulk gas to the inner layers of particles and the corresponding time especially depends on particle size. Biomass heating is almost totally achieved by the radiative heat transfer from the hot walls of reactor to the bed of material. For pyrolysis, instead, working temperature, particles size and the same nature of biomass (through its own pyrolysis heat) have all comparable weights on the process development, so that the corresponding time can be differently depending on one of these factors according to the particular fuel is gasified and the particular conditions are established inside the gasifier. The same analysis also led to the estimation of reaction zone volumes for each biomass fuel, so as a comparison among the dimensions of the differently fed gasification units was finally accomplished. Each biomass material showed a different volumes distribution, so that any dimensioned gasification unit does not seem to be suitable for more than one biomass species. Nevertheless, since reactors diameters were found out quite similar for all the examined materials, it could be envisaged to design a single units for all of them by adopting the largest diameter and by combining together the maximum heights of each reaction zone, as they were calculated for the different biomasses. A total height of gasifier as around 2400mm would be obtained in this case. Besides, by arranging air injecting nozzles at different levels along the reactor, gasification zone could be properly set up according to the particular material is in turn gasified. Finally, since gasification and pyrolysis times were found to considerably change according to even short temperature variations, it could be also envisaged to regulate air feeding rate for each gasified material (which process temperatures depend on), so as the available reactor volumes would be suitable for the complete development of solid conversion in each case, without even changing fluid dynamics behaviour of the unit as well as air/biomass ratio in noticeable measure. The second part of this work dealt with the gas cleaning systems to be adopted downstream the gasifiers in order to run high efficiency CHP units (i.e. internal engines and micro-turbines). Especially in the case multi–fuel gasifiers are assumed to be used, weightier gas cleaning lines need to be envisaged in order to reach the standard gas quality degree required to fuel cogeneration units. Indeed, as the more heterogeneous feed to the gasification unit, several contaminant species can simultaneously be present in the exit gas stream and, as a consequence, suitable gas cleaning systems have to be designed. In this work, an overall study on gas cleaning lines assessment is carried out. Differently from the other research efforts carried out in the same field, the main scope is to define general arrangements for gas cleaning lines suitable to remove several contaminants from the gas stream, independently on the feedstock material and the energy plant size The gas contaminant species taken into account in this analysis were: particulate, tars, sulphur (in H2S form), alkali metals, nitrogen (in NH3 form) and acid gases (in HCl form). For each of these species, alternative cleaning devices were designed according to three different plant sizes, respectively corresponding with 8Nm3/h, 125Nm3/h and 350Nm3/h gas flows. Their performances were examined on the basis of their optimal working conditions (efficiency, temperature and pressure drops, above all) and their own consumption of energy and materials. Successively, the designed units were combined together in different overall gas cleaning line arrangements, paths, by following some technical constraints which were mainly determined from the same performance analysis on the cleaning units and from the presumable synergic effects by contaminants on the right working of some of them (filters clogging, catalysts deactivation, etc.). One of the main issues to be stated in paths design accomplishment was the tars removal from the gas stream, preventing filters plugging and/or line pipes clogging At this scope, a catalytic tars cracking unit was envisaged as the only solution to be adopted, and, therefore, a catalytic material which is able to work at relatively low temperatures was chosen. Nevertheless, a rapid drop in tars cracking efficiency was also estimated for this same material, so that an high frequency of catalysts regeneration and a consequent relevant air consumption for this operation were calculated in all of the cases. Other difficulties had to be overcome in the abatement of alkali metals, which condense at temperatures lower than tars, but they also need to be removed in the first sections of gas cleaning line in order to avoid corrosion of materials. In this case a dry scrubber technology was envisaged, by using the same fine particles filter units and by choosing for them corrosion resistant materials, like ceramic ones. Besides these two solutions which seem to be unavoidable in gas cleaning line design, high temperature gas cleaning lines were not possible to be achieved for the two larger plant sizes, as well. Indeed, as the use of temperature control devices was precluded in the adopted design procedure, ammonia partial oxidation units (as the only considered methods for the abatement of ammonia at high temperature) were not suitable for the large scale units, because of the high increase of reactors temperature by the exothermic reactions involved in the process. In spite of these limitations, yet, overall arrangements for each considered plant size were finally designed, so that the possibility to clean the gas up to the required standard degree was technically demonstrated, even in the case several contaminants are simultaneously present in the gas stream. Moreover, all the possible paths defined for the different plant sizes were compared each others on the basis of some defined operational parameters, among which total pressure drops, total energy losses, number of units and secondary materials consumption. On the basis of this analysis, dry gas cleaning methods proved preferable to the ones including water scrubber technology in al of the cases, especially because of the high water consumption provided by water scrubber units in ammonia adsorption process. This result is yet connected to the possibility to use activated carbon units for ammonia removal and Nahcolite adsorber for chloride acid. The very high efficiency of this latter material is also remarkable. Finally, as an estimation of the overall energy loss pertaining the gas cleaning process, the total enthalpy losses estimated for the three plant sizes were compared with the respective gas streams energy contents, these latter obtained on the basis of low heating value of gas only. This overall study on gas cleaning systems is thus proposed as an analytical tool by which different gas cleaning line configurations can be evaluated, according to the particular practical application they are adopted for and the size of cogeneration unit they are connected to.

Wavelet analysis of heart rate variability related to nocturnal frontal lobe epilepsy seizures

Introduction: Nocturnal frontal lobe epilepsy (NFLE) is a distinct syndrome of partial epilepsy whose clinical features comprise a spectrum of paroxysmal motor manifestations of variable duration and complexity, arising from sleep. Cardiovascular changes during NFLE seizures have previously been observed, however the extent of these modifications and their relationship with seizure onset has not been analyzed in detail. Objective: Aim of present study is to evaluate NFLE seizure related changes in heart rate (HR) and in sympathetic/parasympathetic balance through wavelet analysis of HR variability (HRV). Methods: We evaluated the whole night digitally recorded video-polysomnography (VPSG) of 9 patients diagnosed with NFLE with no history of cardiac disorders and normal cardiac examinations. Events with features of NFLE seizures were selected independently by three examiners and included in the study only if a consensus was reached. Heart rate was evaluated by measuring the interval between two consecutive R-waves of QRS complexes (RRi). RRi series were digitally calculated for a period of 20 minutes, including the seizures and resampled at 10 Hz using cubic spline interpolation. A multiresolution analysis was performed (Daubechies-16 form), and the squared level specific amplitude coefficients were summed across appropriate decomposition levels in order to compute total band powers in bands of interest (LF: 0.039062 - 0.156248, HF: 0.156248 - 0.624992). A general linear model was then applied to estimate changes in RRi, LF and HF powers during three different period (Basal) (30 sec, at least 30 sec before seizure onset, during which no movements occurred and autonomic conditions resulted stationary); pre-seizure period (preSP) (10 sec preceding seizure onset) and seizure period (SP) corresponding to the clinical manifestations. For one of the patients (patient 9) three seizures associated with ictal asystole were recorded, hence he was treated separately. Results: Group analysis performed on 8 patients (41 seizures) showed that RRi remained unchanged during the preSP, while a significant tachycardia was observed in the SP. A significant increase in the LF component was instead observed during both the preSP and the SP (p<0.001) while HF component decreased only in the SP (p<0.001). For patient 9 during the preSP and in the first part of SP a significant tachycardia was observed associated with an increased sympathetic activity (increased LF absolute values and LF%). In the second part of the SP a progressive decrease in HR that gradually exceeded basal values occurred before IA. Bradycardia was associated with an increase in parasympathetic activity (increased HF absolute values and HF%) contrasted by a further increase in LF until the occurrence of IA. Conclusions: These data suggest that changes in autonomic balance toward a sympathetic prevalence always preceded clinical seizure onset in NFLE, even when HR changes were not yet evident, confirming that wavelet analysis is a sensitive technique to detect sudden variations of autonomic balance occurring during transient phenomena. Finally we demonstrated that epileptic asystole is associated with a parasympathetic hypertonus counteracted by a marked sympathetic activation.

Partial Discharge Phenomena in Converter and Traction Transformers: Identification and Reliability

After the development of power electronics converters, the number of transformers subjected to non-sinusoidal stresses (including DC) has increased in applications such as HVDC links and traction (electric train power cars). The effects of non-sinusoidal voltages on transformer insulation have been investigated by many researchers, but still now, there are some issues that must be understood. Some of those issues are tackled in this Thesis, studying PD phenomena behavior in Kraft paper, pressboard and mineral oil at different voltage conditions like AC, DC, AC+DC, notched AC and square waveforms. From the point of view of converter transformers, it was found that the combined effect of AC and DC voltages produces higher stresses in the pressboard that those that are present under pure DC voltages. The electrical conductivity of the dielectric systems in DC and AC+DC conditions has demonstrated to be a critical parameter, so, its measurement and analysis was also taken into account during all the experiments. Regarding notched voltages, the RMS reduction caused by notches (depending on firing and overlap angles) seems to increase the PDIV. However, the experimental results show that once PD activity has incepted, the notches increase PD repetition rate and magnitude, producing a higher degradation rate of paper. On the other hand, the reduction of mineral oil stocks, their relatively low flash point as well as environmental issues, are factors that are pushing towards the use of esters as transformer insulating fluids. This PhD Thesis also covers the study of two different esters with the scope to validate their use in traction transformers. Mineral oil was used as benchmark. The complete set of dielectric tests performed in the three fluids, show that esters behave better than mineral oil in practically all the investigated conditions, so, their application in traction transformers is possible and encouraged.

Potential Analysis for Hypoelliptic Second Order PDEs with Nonnegative Characteristic Form

In this Thesis we consider a class of second order partial differential operators with non-negative characteristic form and with smooth coefficients. Main assumptions on the relevant operators are hypoellipticity and existence of a well-behaved global fundamental solution. We first make a deep analysis of the L-Green function for arbitrary open sets and of its applications to the Representation Theorems of Riesz-type for L-subharmonic and L-superharmonic functions. Then, we prove an Inverse Mean value Theorem characterizing the superlevel sets of the fundamental solution by means of L-harmonic functions. Furthermore, we establish a Lebesgue-type result showing the role of the mean-integal operator in solving the homogeneus Dirichlet problem related to L in the Perron-Wiener sense. Finally, we compare Perron-Wiener and weak variational solutions of the homogeneous Dirichlet problem, under specific hypothesis on the boundary datum.