Objetivo: Apresentar e discutir métodos para estimar prevalências ajustadas em pesquisas clínicas e epidemiológicas, bem como desenvolver rotinas computacionais em SAS e R. Métodos: No contexto de estudo transversal, foi simulada uma amostra de 2.000 observações independentes, considerando o desfecho dicotômico diabetes, sexo como a variável de exposição e idade como variável de ajuste. As estimativas de prevalências ajustadas (IC 95%) foram estimadas pelos métodos de predição condicional e marginal, utilizando as rotinas desenvolvidas em SAS e R. O método Delta foi usado para construir os intervalos de confiança. Os resultados foram comparados com aqueles do SUDAAN (SAS-Callable), Stata e a macro %ADJ_PROP (SAS). Resultados: No exemplo simulado, 68,2% são do sexo feminino e a idade média (DP) foi 57,6 (5,0) anos, sendo 54,2 (3,9) anos em homens e 59,2 (4,6) anos em mulheres. A estimativa da prevalência global do desfecho foi de 25,3% (IC 95%:23,4-27,3); sendo 13,8% (IC 95%:11,7-16,7) e 30,7% (IC 95%:28,3-33,2), respectivamente para homens e mulheres. As estimativas de prevalências ajustadas por idade, por meio do método de predição condicional, foram de 19,6% (IC 95%:16,2-23,6) para homens, e 23,6% (IC 95%:21,2-26,1) para mulheres. Pelo método de predição marginal, as estimativas foram de 22,4% (IC 95%:18,7-26,5) para homens, e 26,3% (IC 95%:24,1-28,6) para mulheres. Conclusão: A discrepância entre as estimativas não ajustadas é devida ao confundimento pela idade. Estimativas livres de confundimento podem ser obtidas por meio das prevalências ajustadas pela idade. No entanto, a estimativa pelo método de predição condicional não engloba a prevalência global. Em virtude disso, o método de predição marginal é, geralmente, mais adequado. A rotina desenvolvida na versão para R é uma alternativa aos softwares comerciais. / Objective: To present and discuss methods to estimate adjusted prevalences for clinical and epidemiological research, and develop computational routines in SAS and R. Methods: In the context of cross-sectional study, it was simulated a sample of 2,000 independent observations, considering the dichotomous outcome diabetes, sex as the exposure variable and age as an adjustment variable. Adjusted prevalences were estimated by the conditional and marginal methods, using routines developed in SAS and R. Confidence intervals were constructed using the Delta method. The results were compared with those of the SUDAAN (SAS-callable), Stata and macro %ADJ_PROP (SAS). Results: In simulated example, 68.2% are female and the mean (SD) age was 57.6 (5.00) years old, being that 54.2 (3.94) years for men and 59.2 (4.60) years in women. The estimated global prevalence of outcome was 25.3% (CI 95%: 23.4-27.3) and 13.8% (CI 95%: 11.7-16.7) and 30.7% (CI 95%: 28.3-33.2), respectively for men and women. Estimates of adjusted prevalence for age, through the conditional method, were 19.6% (CI 95%: 16.2-23.6) for men, and 23.6% (CI 95%: 21,2-26.1) for women. For marginal method, the estimates were 22.4% (CI 95%: 18.7-26.5) for men and 26.3% (CI 95%: 24.1-28.6) for women. Conclusion: The observed discrepancy in estimates by sex, unadjusted, can be attributed to confounding due to difference in age distribution between sexes. Comparable estimates (without confounding) of the prevalences can be obtained through prevalence adjusted for age. However, the estimate for the conditional method does not comprise the global prevalence. As a result, the marginal method is in general more suitable. The developed routines can be useful for estimating adjusted prevalences, particularly the R version (an alternative to commercial software).

Objetivo: Apresentar e discutir métodos para estimar prevalências ajustadas em pesquisas clínicas e epidemiológicas, bem como desenvolver rotinas computacionais em SAS e R. Métodos: No contexto de estudo transversal, foi simulada uma amostra de 2.000 observações independentes, considerando o desfecho dicotômico diabetes, sexo como a variável de exposição e idade como variável de ajuste. As estimativas de prevalências ajustadas (IC 95%) foram estimadas pelos métodos de predição condicional e marginal, utilizando as rotinas desenvolvidas em SAS e R. O método Delta foi usado para construir os intervalos de confiança. Os resultados foram comparados com aqueles do SUDAAN (SAS-Callable), Stata e a macro %ADJ_PROP (SAS). Resultados: No exemplo simulado, 68,2% são do sexo feminino e a idade média (DP) foi 57,6 (5,0) anos, sendo 54,2 (3,9) anos em homens e 59,2 (4,6) anos em mulheres. A estimativa da prevalência global do desfecho foi de 25,3% (IC 95%:23,4-27,3); sendo 13,8% (IC 95%:11,7-16,7) e 30,7% (IC 95%:28,3-33,2), respectivamente para homens e mulheres. As estimativas de prevalências ajustadas por idade, por meio do método de predição condicional, foram de 19,6% (IC 95%:16,2-23,6) para homens, e 23,6% (IC 95%:21,2-26,1) para mulheres. Pelo método de predição marginal, as estimativas foram de 22,4% (IC 95%:18,7-26,5) para homens, e 26,3% (IC 95%:24,1-28,6) para mulheres. Conclusão: A discrepância entre as estimativas não ajustadas é devida ao confundimento pela idade. Estimativas livres de confundimento podem ser obtidas por meio das prevalências ajustadas pela idade. No entanto, a estimativa pelo método de predição condicional não engloba a prevalência global. Em virtude disso, o método de predição marginal é, geralmente, mais adequado. A rotina desenvolvida na versão para R é uma alternativa aos softwares comerciais. / Objective: To present and discuss methods to estimate adjusted prevalences for clinical and epidemiological research, and develop computational routines in SAS and R. Methods: In the context of cross-sectional study, it was simulated a sample of 2,000 independent observations, considering the dichotomous outcome diabetes, sex as the exposure variable and age as an adjustment variable. Adjusted prevalences were estimated by the conditional and marginal methods, using routines developed in SAS and R. Confidence intervals were constructed using the Delta method. The results were compared with those of the SUDAAN (SAS-callable), Stata and macro %ADJ_PROP (SAS). Results: In simulated example, 68.2% are female and the mean (SD) age was 57.6 (5.00) years old, being that 54.2 (3.94) years for men and 59.2 (4.60) years in women. The estimated global prevalence of outcome was 25.3% (CI 95%: 23.4-27.3) and 13.8% (CI 95%: 11.7-16.7) and 30.7% (CI 95%: 28.3-33.2), respectively for men and women. Estimates of adjusted prevalence for age, through the conditional method, were 19.6% (CI 95%: 16.2-23.6) for men, and 23.6% (CI 95%: 21,2-26.1) for women. For marginal method, the estimates were 22.4% (CI 95%: 18.7-26.5) for men and 26.3% (CI 95%: 24.1-28.6) for women. Conclusion: The observed discrepancy in estimates by sex, unadjusted, can be attributed to confounding due to difference in age distribution between sexes. Comparable estimates (without confounding) of the prevalences can be obtained through prevalence adjusted for age. However, the estimate for the conditional method does not comprise the global prevalence. As a result, the marginal method is in general more suitable. The developed routines can be useful for estimating adjusted prevalences, particularly the R version (an alternative to commercial software).

Objetivo: Apresentar e discutir métodos para estimar prevalências ajustadas em pesquisas clínicas e epidemiológicas, bem como desenvolver rotinas computacionais em SAS e R. Métodos: No contexto de estudo transversal, foi simulada uma amostra de 2.000 observações independentes, considerando o desfecho dicotômico diabetes, sexo como a variável de exposição e idade como variável de ajuste. As estimativas de prevalências ajustadas (IC 95%) foram estimadas pelos métodos de predição condicional e marginal, utilizando as rotinas desenvolvidas em SAS e R. O método Delta foi usado para construir os intervalos de confiança. Os resultados foram comparados com aqueles do SUDAAN (SAS-Callable), Stata e a macro %ADJ_PROP (SAS). Resultados: No exemplo simulado, 68,2% são do sexo feminino e a idade média (DP) foi 57,6 (5,0) anos, sendo 54,2 (3,9) anos em homens e 59,2 (4,6) anos em mulheres. A estimativa da prevalência global do desfecho foi de 25,3% (IC 95%:23,4-27,3); sendo 13,8% (IC 95%:11,7-16,7) e 30,7% (IC 95%:28,3-33,2), respectivamente para homens e mulheres. As estimativas de prevalências ajustadas por idade, por meio do método de predição condicional, foram de 19,6% (IC 95%:16,2-23,6) para homens, e 23,6% (IC 95%:21,2-26,1) para mulheres. Pelo método de predição marginal, as estimativas foram de 22,4% (IC 95%:18,7-26,5) para homens, e 26,3% (IC 95%:24,1-28,6) para mulheres. Conclusão: A discrepância entre as estimativas não ajustadas é devida ao confundimento pela idade. Estimativas livres de confundimento podem ser obtidas por meio das prevalências ajustadas pela idade. No entanto, a estimativa pelo método de predição condicional não engloba a prevalência global. Em virtude disso, o método de predição marginal é, geralmente, mais adequado. A rotina desenvolvida na versão para R é uma alternativa aos softwares comerciais. / Objective: To present and discuss methods to estimate adjusted prevalences for clinical and epidemiological research, and develop computational routines in SAS and R. Methods: In the context of cross-sectional study, it was simulated a sample of 2,000 independent observations, considering the dichotomous outcome diabetes, sex as the exposure variable and age as an adjustment variable. Adjusted prevalences were estimated by the conditional and marginal methods, using routines developed in SAS and R. Confidence intervals were constructed using the Delta method. The results were compared with those of the SUDAAN (SAS-callable), Stata and macro %ADJ_PROP (SAS). Results: In simulated example, 68.2% are female and the mean (SD) age was 57.6 (5.00) years old, being that 54.2 (3.94) years for men and 59.2 (4.60) years in women. The estimated global prevalence of outcome was 25.3% (CI 95%: 23.4-27.3) and 13.8% (CI 95%: 11.7-16.7) and 30.7% (CI 95%: 28.3-33.2), respectively for men and women. Estimates of adjusted prevalence for age, through the conditional method, were 19.6% (CI 95%: 16.2-23.6) for men, and 23.6% (CI 95%: 21,2-26.1) for women. For marginal method, the estimates were 22.4% (CI 95%: 18.7-26.5) for men and 26.3% (CI 95%: 24.1-28.6) for women. Conclusion: The observed discrepancy in estimates by sex, unadjusted, can be attributed to confounding due to difference in age distribution between sexes. Comparable estimates (without confounding) of the prevalences can be obtained through prevalence adjusted for age. However, the estimate for the conditional method does not comprise the global prevalence. As a result, the marginal method is in general more suitable. The developed routines can be useful for estimating adjusted prevalences, particularly the R version (an alternative to commercial software).

Bakalauro baigiamajame darbe nagrinėjami bankroto prognozavimo metodai. Teorinėje darbo dalyje aptarta bankroto samprata, išanalizuoti įmonės bankrotą sąlygojantys veiksniai ir priežastys, apžvelgti finansiniai ir nefinansiniai tyrimo metodai įmonių bankrotui prognozuoti. Praktinė darbo dalis skirta bankroto grėsmės įvertinimui pagal pasirinktus finansinius bankroto prognozavimo metodus konkrečiose gamybinėse įmonėse ir naudojamų tyrimo metodų tinkamumo nustatymui. Tyrimas parodė, kad tinkamiausi Lietuvos gamybinių įmonių bankrotui prognozuoti yra tiesinės diskriminantinės analizės modeliai – Altman, Springate, R. Liss ir Ca-Score. Be bankroto prognozavimo modelių, norint nustatyti kylančią bankroto grėsmę, reikia taikyti ir finansinius santykinius rodiklius bei pinigų srautų santykinius rodiklius. Atlikus tyrimą, pasitvirtino iškelta hipotezė, kad nėra vieno universalaus, tinkančio visoms įmonėms bankroto prognozavimo metodo. / The Bachelor‘s Thesis research methods of bankruptcy prediction. The theoretical part of Bachelor‘s Thesis discusses a conception of a bankruptcy, analyzes classification of factors and reasons that determine bankruptcy of a company and reviews fiscal research methods to analyze the reasons of companies’ bankruptcies. The practical part is aimed at evaluating the threat of a bankruptcy according to the chosen fiscal methods in specific companies and at determining suitability of the applied research models. The research has proved the linear discriminant models (Altman, Springate, R. Liss, Ca-Score) to be the most suitable for predicting bankruptcy in Lithuanian manufacturing companies. As well as to determine the threat of bankruptcy, it is necessary to apply financial ratios and ratios of cash flow. After analyzing, the raised hypothesis has proved out, that there is not a single particular method of the bankruptcy forecasting, which would suit all companies.

Expansive soils are widely distributed in arid and semi-arid regions around the world and are typically found in a state of unsaturated condition. These soils are constituted of the clay mineral montmorillonite that is highly active and contributes significantly to volume changes of soils due to variations in the natural water content conditions. The volume changes of expansive soils often cause damage to lightly loaded structures. The costs associated with the damage to lightly loaded structures constructed on expansive soils in the United States alone were estimated as $2.3 billion per year in 1973, which increased to $13 billion per year by 2009. In other words, these damages have increased more than five fold during the last four decades. Similar trends in damages were also reported in other countries (e.g., Australia, China, France, Saudi Arabia, United Kingdom, etc.). Numerous methods have been proposed in the literature over the past 50 years for the prediction of the volume change movement of expansive soils. However, the focus of these methods has been towards estimating the maximum potential heave, which occurs when soils attain the saturation condition. The results of heave estimation considering saturated soil conditions are not always useful in engineering practice. This is because most of damages due to expansive soils often occur prior to reaching the saturation condition. A reliable design of structures on expansive soils is likely if the anticipated soil movements in the field can be reliably estimated over time, taking into account the influence of environmental factors. Limited studies are reported in the literature during the past decade in this direction to estimate/predict the expansive soil movements over time. The existing methods, however, suffer from the need to run expensive and time consuming tests. In addition, verification of these studies for different natural expansive soils has been rather limited. A simple approach, which is referred to as a modulus of elasticity based method (MEBM), is proposed in this study for the prediction of the heave/shrinkage movements of natural expansive soils over time. The proposed MEBM is based on a simplified constitutive relationship used for the first time to estimate the vertical soil movements with respect to time in terms of the matric suction variations and the corresponding values of the modulus of elasticity. The finite element program VADOSE/W (Geo-Slope 2007) for simulating the soil-atmospheric interactions is used as a tool to estimate the changes in matric suction over time. A semi-empirical model that was originally proposed by Vanapalli and Oh (2010) for fine-grained soils has been investigated and extended for unsaturated expansive soils to estimate the variation of the modulus of elasticity with respect to matric suction in the constitutive relationship of the proposed method. The MEBM has been tested for its validity in five case studies from the literature for a wide variety of site and environmental conditions, from Canada, China, and the United States. For each case study, factors influencing the volume change behavior of soils, such as climate conditions, soil cracks, lawn irrigation, and cover type (pavement, vegetation), are successfully modeled over the period of each simulation. The proposed MEBM provides good predictions of soil movements with respect to time for all the case studies. The MEBM is simple and efficient for the prediction of vertical movements of natural expansive soils underlying lightly loaded structures. In addition, a new dimensionless model is also proposed, based on the dimensional analysis approach, for the estimation of the modulus of elasticity which can also be used in the constitutive relationship of the MEBM. The dimensional model is rigorous and takes into account the most significant influencing parameters such as matric suction, net confining stress, initial void ratio, and degree of saturation. This model provides a comprehensive characterization of the modulus of elasticity of expansive soils under unsaturated conditions for different scenarios of loading conditions (i.e., both lightly and heavily loaded structures). The results of the present study are encouraging for proposing guidelines based on further investigations and research studies for the rational design of pavements, shallow and deep foundations placed on/in expansive soils using the mechanics of unsaturated soils.

<div> <div> <div> <p>Proteins are drivers of almost all biological processes in the cell. The functions of a protein are dependent on their three-dimensional structure and elucidating the structure and function of proteins is key to understanding how a biological system operates. In this research, we developed computational methods using machine learning techniques to predicts the structure and function of proteins. Protein 3D structure prediction has advanced significantly in recent years, largely due to deep learning approaches that predict inter-residue contacts and, more recently, distances using multiple sequence alignments (MSAs). The performance of these models depends on the number of similar protein sequences to the query protein, wherein some cases similar sequences are few but dissimilar sequences with local similarities are more and can be helpful. We have developed a novel deep learning-based approach AttentiveDist which further improves over the previous state of art. We added an attention mechanism where dis-similar sequences are also used (increasing number of sequences) and the model itself determines which information from such sequences it should attend to. We showed that the improvement of distance predictions was successfully transferred to achieve better protein tertiary structure modeling. We also show that structure prediction from a predicted distance map can be further enhanced by using predicted inter-residue sidechain center distances and main-chain hydrogen-bonds. Protein function prediction is another avenue we explored where we want to predict the function that a protein will perform. The crux of the approach is to predict the function of protein based on the function of similar sequences. Here, we developed a method where we use dissimilar sequences to extract additional information and improve performance over the previous approaches. We used phylogenetic analysis to determine if a dissimilar sequence can be close to the query sequence and thus can provide functional information. Our method was ranked highly in worldwide protein function prediction competition CAFA3 (2016-2019). Further, we expanded the method with a neural network to predict protein toxicity that can be used as a safety check for human-designed protein sequences.</p></div></div></div>

Railway noise is a very important and growing health hazard in today´s society.Railway systems pass through towns and cities and create noise. When trainsride through or over railway bridges, switches and crossings the noise increasessubstantially, causing great annoyance to the residents in the area.At the present time, acoustic regulations exist in most countries and are set to achievea good environmental quality in residential areas, schools, hospitals, offices andhotels.A few calculation models exist for railway traffic noise, such as The NPM 1996,NORD2000 and Cnossos-EU. The NPM 1996 is currently used in Sweden to calculatenoise propagation from railway traffic. To uphold the regulations set, it is importantthat the method used is as precise as possible. All of these calculation modelsare based on several correction factors. Today, the current corrections for trackconditions, that is railway bridges, switches and crossings, are not very accurate andneed to be reconsidered.The aim of this project was to investigate and quantify the error of the NPM correctionfactors and give some indication of how they should be adjusted. This is done to makethe noise prediction from railway traffic more accurate and thus protect the residentsfrom these health risks.The specific objective was to perform significant amount of field measurements ofnoise from trains on different steel bridges, switches and crossings, as well as on afew concrete bridge according to the measurement standard Nordtest NT 098. Thefield work was carried out over the period March to May, when weather conditionsmet the criteria for field measurements, in and around the Stockholm area.The results obtained revealed that the correction factors for steel bridges andcrossings are considerably lower in the NPM than measured in this project. However,the correction factors for switches and concrete bridges are similar to the ones inthe NPM. In this thesis, a part of the correction factors have only been invalidatedto a degree. It has been shown that each bridge is unique and perhaps there is apossibility of finding similarities between some type of bridges. However, much moreiv |measurements are needed to see any correlation between each bridge type. Thus,further and more comprehensive measurements have to be carried out in order toestablish new accurate correction factors for track conditions in the Nordic PredictionMethod.

This dissertation addresses the challenge of predicting human-induced subsidence in tectonic settings. The study focuses on the non-symmetric and shape-defying nature of subsidence troughs in tectonic regions, which deviates from conventional symmetric models. The aim of the dissertation is to improve the accuracy of subsidence prediction by incorporating horizontal stress effects into empirical methods. Through a combination of numerical investigations and empirical modelling, the research reveals stress-induced patterns in subsidence profiles. The developed model, based on various concepts, successfully incorporates asymmetry and shape deviation, resulting in significantly improved prediction accuracy. Application of the model to a real subsidence case in a salt cavern shows a 30% improvement in prediction (based on mean squared error comparison with classical solution). This new solution covers subsidence profile patterns not previously considered by empirical models.:Inhalt 1 Introduction 2 State of the art 2.1 Subsidence prediction methods 2.1.1 Empirical subsidence prediction method overview 2.1.2 Numerical methods for subsidence prediction 2.2 Subsidence monitoring methods 2.2.1 Observation methods 2.2.2 Interplay and evolution of techniques 2.3 Subsidence anomalies 2.4 In-situ-stress field 2.5 Subsidence prediction methods for anomalies 2.6 Conclusions 3 Goals and objectives 4 Foundations 4.1 Empirical subsidence prediction methods 4.1.1 Convergence 4.1.2 Transmission coefficient 4.1.2 Influence factor 4.2 Numerical models for subsidence case 4.2.1 Grid size for subsidence case 4.2.2 Boundary conditions 4.2.3 Constitutive models 4.3 Validation 4.3.1 Observation methods 4.3.2 Parameter estimation 4.3.3 Global parameter estimation 4.3.4 Local parameter estimation 4.3.5 Quality measures for result valuation and validation 5 Methodology 6 Numerical investigation 6.1 Preliminary investigation 6.1.1 Method 6.1.2 Choice of constitutive model 6.1.3 Model and input data 6.1.4 Preliminary investigation results 6.2 Design of the main experiment: non-uniform stress distribution 6.2.1 Constitutive model and input data 6.2.2 Model simplification 6.2.3 Output data 6.3 Contribution of asymmetrical stress distribution 6.3.1 Discussion of the basic distribution form 6.3.2 Discussion of maximum subsidence 6.3.3 Discussion of assymetry 6.3.4 Discussion of influence angle 6.4 Conclusions 7 Adaptation of an empirical model to the discovered features 7.1 Subsidence asymmetry 7.2 Subsidence shape flexibility 7.3 Unifying solution 7.4 Conclusion and outlook 8 Application to a full scale 8.1 General information for a salt cavern storage field 8.2 Estimation of the observed subsidence surface as reference 8.3 Model implementation 8.3.1 Parameter estimation results 8.4 Statistical validation of models 8.5 Conclusions 9 Conclusion 9.1 Limitations 9.2 Outlook References Appendix

The Swedish law requires a set of noise limit for residential buildings for health and safety. Conventionally, the Nordic Prediction Method (NMT) is used to predict the noise. However Nord2000 and CNOSSOS is going to be introduced to replace NMT. An investigation was made to determine which is more accurate in predicting railway noises, particularly reflected railway noises due to the uncharacteristic result in preliminary test. Compromises were made to recreate the measured scenarios in SoundPLAN 8.1. CNOSSOS features an alternative source model requiring new data on the trains and tracks and therefore unable to be compared. The conversion between NMT and Nord2000 was not successful due to the poor documentation of the method. An equivalent value in octave band was used instead. The result were not conclusive as no correlation was found. This might be partly due to the lacking of source model data in the form of track roughness. Result in reflected sound was not conclusive as well since 2 of 3 cases favour Nord2000 and the remaining one favours NMT. The uncertainty introduced in using of the standards and measurements were also rather large, which is also one of the factor in non correlating results. / Svensk lag kräver en viss uppsättning bullergränser gällande boende i bostadshus för deras hälsa och säkerhet. Konventionellt används den nordiska beräkningsmodellen för tågbuller (NMT) för att förutsäga bullret. Nord2000 och CNOSSOS kommer dock att införas för att ersätta NMT. En undersökning gjordes för att avgöra vilken eller vilka modeller som är mest exakta för att kunna förutsäga järnvägsljud, i synnerhet reflekterade järnvägsljud på grund av tidigare osäkerhet i resultatet i det preliminära testet. Kompromisser gjordes för att återskapa de uppmätta scenarierna i SoundPLAN 8.1. CNOSSOS erbjuder en alternativ modellering men kräver ny data från både tåg och räls och omöjliggjorde därför att kunna jämföras. Konverteringen mellan NMT och Nord2000 lyckades inte på grund av den bristfälliga dokumentationen av metoden. Ett ekvivalent värde i oktavband användes istället. Resultatet var inte definitivt eftersom ingen korrelation upptäcktes. Detta kan delvis bero på bristande källmodelldata i form av spårgrovhet. Resultatet av reflekterat ljud var inte heller definitivt eftersom 2 av 3 fall förespråkar Nord2000 och det återstående förespråkar NMT. Osäkerheten som implementeras vid användning av standarder och mätningar är också av betydande storlek, vilket också är en av faktorerna som härleder korrelerande resultat.