Prior studies indicated that beam-to-column connections of reinforced concrete (RC) moment resisting frame structures experience considerable deformations under earthquake loading and these deformations have a major contribution to story drift of the building. In current analysis and design applications, however, the connection regions are generally modeled as rigid zones and the inelastic behavior of the joint is not taken into account. This assumption gives rise to an underestimation of the story drifts and hence to an improper assessment of the seismic performance of the structure. In order to implement the effect of these regions into the seismic design and analysis of buildings, a model that properly represents the seismic behavior of connection regions needs to be developed. In this study, a parametric model which predicts the joint shear strength versus strain relationship is generated by investigating the several prior experimental studies on RC beam-to-column connections subjected to cyclic loading and establishing an extensive database. Considering previous experimental research and employing statistical correlation method, parameters that significantly influence the joint behavior are determined and these parameters are combined together to form a joint model. This model is then verified by comparing the results obtained from the dynamic earthquake analysis by Perform 3D with the experimental ones. The main contribution of the developed model is taking into account parameters like the effect of eccentricity, column axial load, slab, wide beams and transverse beams on the seismic behavior of the connection region, besides the key parameters such as concrete compressive strength, reinforcement yield strength, joint width and joint transverse reinforcement ratio.

This thesis jointly models the 24 hour work trip timing and mode choice decisions of commuters in the Greater Toronto Area. A discrete-continuous specification, with a multinomial logit model for mode choice and an accelerated time hazard model for trip timing, is used to allow for unrestricted correlation between these two fundamental decisions. Statistically significant correlations are found between mode choice and trip timing for work journeys with expected differences between modes. Furthermore, the joint models have a wide range of policy sensitive statistically significant parameters of intuitive sign and magnitude, revealing expected differences between workers of different occupation groups. Furthermore, the estimated models have a high degree of fit to observed cumulative departure and arrival time distribution functions and to observed mode choices. Finally, sensitivity tests have demonstrated that the model is capable of capturing peak spreading in response to increasing auto congestion.

This thesis jointly models the 24 hour work trip timing and mode choice decisions of commuters in the Greater Toronto Area. A discrete-continuous specification, with a multinomial logit model for mode choice and an accelerated time hazard model for trip timing, is used to allow for unrestricted correlation between these two fundamental decisions. Statistically significant correlations are found between mode choice and trip timing for work journeys with expected differences between modes. Furthermore, the joint models have a wide range of policy sensitive statistically significant parameters of intuitive sign and magnitude, revealing expected differences between workers of different occupation groups. Furthermore, the estimated models have a high degree of fit to observed cumulative departure and arrival time distribution functions and to observed mode choices. Finally, sensitivity tests have demonstrated that the model is capable of capturing peak spreading in response to increasing auto congestion.

Underground excavations break the balance of the initial stress field and cause stress redistributions in the surrounding rock masses. Problems normally arise as the stress exceeds the rock mass strength. In addition, the rock mass contains preexisting defects, such as the fissures, fractures, joints, faults, shear zones, dikes, etc., which could significantly weaken the rock mass strength and make the rock mass behavior complicated. The stability of underground excavations is of great importance to an operating mine project since it ensures the safety of the working environment and the successful ore exploration. Due to the complex geological conditions and engineering disturbances, the assessment of rock mass stability for a practical engineering problem is extremely challenging and difficult, which needs to be solved by the modern numerical methods. In this dissertation, the rock mass stability around tunnels in an underground mine in the USA was investigated by performing three-dimensional modeling using the 3DEC 3-Dimensional Distinct Element Code. Comprehensive stress analyses were respectively carried out on a preliminary model and a more advanced model. In the preliminary study, the built model contains the inclined lithologies, a non-persistent fault, and a convoluted tunnel system. The geomechanical property values used for the rock masses and discontinuities in the numerical model were estimated using the available geotechnical information and the experience of the research group. The Mohr-Coulomb and strain softening constitutive relations were prescribed for the rock masses; the coulomb slip joint model was assigned for the discontinuities. The influence of the boundary conditions, block constitutive models, horizontal in situ stress and rock support system on the tunnel stability was investigated. The rock mass behavior was quantified using the results of stress, displacement, and yielded zones around the tunnels. It showed that the roller boundary conditions resulted in slightly different but comparable results with the combined boundary conditions (roller and stress combined) where K0 equals to 0.4 or 0.5. Whereas the in-situ stress field for a complex geological system can only be obtained by applying proper boundary stresses and then by performing stress analysis. The softening behavior of the rock masses caused more deformations and yielded zones around the tunnels; the rock masses around the tunnels were observed to reach the residual strength values, which can be treated as failed areas. In addition, the M-C and s-s rock masses reacted differently as the K0 value changed. At K0=1.0, the tunnels seemed to be the most stable; K0=1.5, however, provided the worst scenario with roof and floor problems. With respect to the effectiveness of the support system, a large amount of the bonds of the supports was failing, thus, the deformations and yielded zones around the tunnels were slightly improved. Finally, comparisons between the numerical modeling results and the field measurements implied the applicability of strain softening behavior and a K0 value between 0.5 and 1.0 for the mine. Based on the specific geological, geotechnical, and construction information, a numerical model incorporating accurate features was developed. It includes a non-planar, weak interlayer, the persistent and non-persistent faults, and the open and backfilled excavations. The mechanical property values used for the rock masses and faults were estimated based on the laboratory test results of the intact rock and smooth joints. The strain softening behavior was specified for the rock masses belonging to the average quality, and the rock masses that reached residual strengths were assumed to be failing. The linear relations between the fault stiffnesses and normal stress were described using the continuously yielding joint model. To simulate the mine construction process in the field, the sequential excavation, backfilling, and supporting procedures were numerically implemented; additionally, a novel routine was applied to account for the delayed installation of the supports. Results showed that the tunnels close to the fault and the backfilled area were less stable. Most of the displacements around the tunnels occurred within a distance of zero to 2 or 3 m from the tunnel surface. The varying K0 value caused great changes in the rock mass behavior and the shear behavior of the major fault; significant instability of the tunnels was triggered by the high horizontal in situ stress. Parametric studies on the rock mass condition, rock mass residual strengths, and fault property values showed that the tunnel stability was more sensitive to the former two factors than the last one. A systematic investigation was conducted to evaluate the current rock supports installed at the mine where the increasing stress relaxation was incorporated. The deformations and of the failure zone thicknesses around the tunnels were reduced up to 8% and 20% after applying the supports instantaneously, and the reductions were improved by the delayed installation of supports. Additionally, the safety of supports was evaluated by the bond shear and bolt tensile failures, which was also improved with incorporation of delayed supporting. It was found that the current rock supports are insufficient in length, bond and tensile strengths. Therefore, a stronger support system was suggested. The stronger supports worked better in stabilizing the tunnels. Based on the deformations and failures of the rock masses, the length of the bolts on walls was suggested to be 4-5 m. At the end, the horizontal convergence strain predicted by the numerical simulations were calculated at two locations where the tape extensometers were installed. Good agreements with the field measurements were obtained for the cases that have the average rock mass properties and K0 values in the range 0.5-1.25.

In this study we look at the problem of analysing survival data in the presence of longitudinally collected covariates. New methodology for analysing such data has been developed through the use of hidden Markov modeling. Special attention has been given to the case of large information volume, where a preliminary data reduction is necessary. Novel graphical diagnostics have been proposed to assess goodness of fit and significance of covariates. The methodology developed has been applied to the data collected on behaviors of Mexican fruit flies, which were monitored throughout their lives. It has been found that certain patterns in eating behavior may serve as an aging marker. In particular it has been established that the frequency of eating is positively correlated with survival times. / Thesis (Ph.D, Mathematics & Statistics) -- Queen's University, 2010-04-26 18:34:01.131

La prédiction du devenir du greffon et de sa survie permettrait d’optimiser la prise en charge des patients transplantés. Le suivi des patients transplantés rénaux inclue des mesures répétées de marqueurs longitudinaux tels que la créatinine sérique et l’exposition aux médicaments immunosuppresseurs. L’approche statistique récemment proposée des modèles conjoints permet d’analyser la relation entre un processus longitudinal et la survenue d’un événement clinique. Dans la première partie de ce travail de thèse, nous avons utilisé les modèles conjoints à classes latentes pour étudier l’impact du profil de créatinine sérique au cours des 18 premiers mois post-greffe sur la survie du greffon à long terme. Dans la cohorte étudiée, trois groupes homogènes caractérisés par une trajectoire spécifique de l’évolution de la créatinine sérique en fonction du temps et un risque d’échec de greffe spécifique ont été identifiés. Les probabilités individuelles de l’échec de greffe pendant les 10 premières années post-transplantation ont été calculées sur la base du modèle conjoint développé. Chez les patients qui n’avaient pas développé d’anticorps anti-HLA spécifiques du donneur, le risque d’échec de greffe en fonction du temps était prédit avec un niveau de performance satisfaisant en termes de spécificité, sensibilité et précision.L’utilité clinique de cet outil devra être évaluée avec une approche dynamique. Dans une seconde partie, les modèles non linéaires à effets mixtes combinés avec l’approche des modèles de mélange a été utilisée pour analyser (i) l’association entre la variabilité de l’exposition au tacrolimus au cours du temps et l’adhésion au traitement rapportée par le patient et (ii) l’impact de cette variabilité d’exposition sur le risque de rejet aigu. Ce modèle a montré un effet significatif de la variabilité de l’exposition au cours du temps du tacrolimus sur la survenu de rejet aigu au-delà de 3 mois post-transplantation. Au contraire, aucune association entre l’adhésion et la variabilité de l’exposition au tacrolimus d’une part, et le risque de rejet aigu d’autre part n’a été observée dans cette étude qui n’incluait que des patients modérément non-adhérents. Ce résultat pose la question de l’impact d’une non adhésion modérée sur le devenir du greffon. / Prediction of graft outcome would be useful to optimize patient care. Follow-up of kidneytransplant patients include repeated measurements of longitudinal markers, such as serum creatinine and immunosuppressive drug exposure. Recently proposed joint models areappropriate to analyze relationship between longitudinal processes and time-to-event data. In the first part of present work, we used the approach of joint latent class mixed models tostudy the impact of time-profiles of serum creatinine collected within the first 18 months after kidney transplantation on long-term graft survival. The studied cohort was parted into three homogenous classes with a specific time-evolution of serum creatinine and a specific risk of graft failure. The individual predicted probabilities of graft failure up to 10 years posttransplantation, calculated from this joint model were satisfying in terms of sensitivity, specificity and overall accuracy, for patients who had not developed de novo donor specificanti-HLA antibodies. The clinical usefulness of developed predictive tooI needs to beevaluated with a dynamic approach. In the second part, non-linear mixed effects models witha mixture of distribution for random effects were used to investigate (i) the associationbetween variability over time of tacrolimus exposure and self-reported drug adherence and(ii) the impact of this variability on the acute rejection risk. This model found a significantimpact of tacrolimus time-exposure variability on acute rejection onset beyond 3 months posttransplantation. On the contrary, no association between adherence and (i) variability oftacrolimus time-exposure and (ii) acute rejection was observed in our study which included moderate non-adherent patients only. This result questions the impact of moderate nonadherence on graft outcome.

Une attention particulière doit être accordée aux jeunes patients transplantés du rein, qui sont prioritaires en France dans l’attribution de greffon rénal. Les objectifs de cette thèse étaient 1) d’étudier la dynamique du risque instantané de la perte du greffon en fonction de l’âge courant après la transplantation chez les jeunes ; 2) de réaliser un revue systématique des modèles de prédiction de perte du greffon rénal tout âge confondu; et 3) de développer et valider un modèle de prédiction adapté à cette jeune population. Pour répondre aux objectifs 1 et 3, nous avons utilisé les données des registres nationaux français REIN et CRITAL qui incluent de manière exhaustive tous les patients transplantés rénaux pédiatriques. Pour l’Objectif 1, une méthode statistique en deux étapes a permis de mettre en évidence une augmentation accrue du risque instantané de perte du greffon au moment de l’adolescence. Pour l’Objectif 2, une revue systématique des articles publiés entre 2005 et 2015 a montré qu’aucun outil prédictif de la perte du greffon n’avait été spécifiquement proposé pour les patients pédiatriques, ni aucun outil de prédiction dynamique tout âge confondu. Pour l’Objectif 3, nous avons développé et validé par validation croisée interne un modèle de prédiction dynamique de perte du greffon pour les jeunes transplantés, à partir d’un modèle conjoint à effets aléatoires partagés. Ce modèle incluait des prédicteurs classiques à l’inclusion défini par le 90ième jour après la transplantation (des caractéristiques du receveur (sexe, âge à la transplantation, durée de dialyse pré-greffe, maladie rénale initiale, nadir du DFGe entre la transplantation et J90), du donneur (âge, type), et de la transplantation (durée d’ischémie froide, nombre d’incompatibilités HLA, statut donneur/receveur pour sérologie CMV)). Le modèle incluait également la trajectoire du DFG estimé après la transplantation, en supposant que le risque instantané de perte du greffon dépendait à la fois du niveau courant du DFG mais aussi de sa pente courante. Nos résultats indiquent que ce modèle avait de bonnes performances prédictives (AUC à 5 ans variant de 0.75 à 0.86 selon les temps de prédiction après la transplantation), bien meilleures que le modèle de Cox classique ne tenant compte que des prédicteurs à l’inclusion (AUC à 5 ans variant de 0.56 à 0.62). Ce modèle permettant la mise à jour à chaque visite clinique après la transplantation, du risque futur de la perte du greffon en fonction de toutes les valeurs observées précédentes du DFG, devra être validé sur d’autres populations que la population française. Nous pensons en effet qu’un tel outil pourrait à terme être utile dans le suivi clinique des jeunes patients transplantés rénaux. / Particular attention should be paid to young patients transplanted from the kidney, which have priority in France in the assignment of renal graft. The objectives of this thesis were 1) to study the dynamics of the hazard of graft failure by current age after transplantation in young people; 2) to carry out a systematic review of prediction models for renal graft failure at all ages; and (3) to develop and validate a prediction model for this young population. To achieve Objectives 1 and 3, we used data from the French national registries REIN and CRISTAL, which included all pediatric renal transplant patients. For Objective 1, a two-stage statistical method revealed an increase in the hazard of graft failure during adolescence. For Objective 2, a systematic review of articles published between 2005 and 2015 showed that no predictive tool for graft failure has been specifically proposed for pediatric patients, as well as no dynamic predictive model for any age. For Objective 3, we developed and validated using internal cross-validation a dynamic prediction model of graft failure for young transplanted patients, using a joint model with shared random effects. This model included standard baseline predictors at the 90th day after transplantation (characteristics of the recipient (sex, age at transplantation, pretrasplant dialysis duration, primary renal disease, nadir of eGFR at J90), the donor (age and type), and transplantation (duration of cold ischemia, number of HLA incompatibilities, donor/recipient cytomegalovirus (CMV) serology status). The model also included the trajectory of GFR estimated after transplantation, assuming that the hazard of graft failure depended on both the current value of eGFR and its current slope. Our results indicate that this model had good predictive performances (AUC at 5 years ranging from 0.75 to 0.86 according to the time at prediction after transplantation), which were much better than the standard Cox model accounting for baseline predictors only (5-year AUC variant from 0.56 to 0.62). This model which allows the prediction of graft failure to be updated at each clinical visit after transplantation based on all previous observed values of eGFR, should be validated on populations other than the French population. We believe that such a tool could ultimately be useful in the clinical follow-up of young kidney transplanted patients.

Title: Joint Models for Longitudinal and Time-to-Event Data Author: Jana Vorlíčková Department: Department of Probability and Mathematical Statistics Supervisor: doc. RNDr. Arnošt Komárek, Ph.D., Department of Probability and Mathematical Statistics Abstract: The joint model of longitudinal data and time-to-event data creates a framework to analyze longitudinal and survival outcomes simultaneously. A commonly used approach is an interconnection of the linear mixed effects model and the Cox model through a latent variable. Two special examples of this model are presented, namely, a joint model with shared random effects and a joint latent class model. In the thesis we focus on the joint latent class model. This model assumes an existence of latent classes in the population that we are not able to observe. Consequently, it is assumed that the longitudinal part and the survival part of the model are independent within one class. The main intention of this work is to transfer the model to the Bayesian framework and to discuss an estimation procedure of parameters using a Bayesian statistic. It consists of a definition of the model in the Bayesian framework, a discussion of prior distributions and the derivation of the full conditional distributions for all parameters of the model. The model's ability to...

Submitted Diploma thesis deals with stress-strain analysis of deformation proximal end of femur with applied total hip joint endoprosthesis (replacement) – shortcut type. To identify deformation and tensity (stress) was used computational simulation by method of final elements. Have been created two computational models TEP- type Santori and type DePuy Proxima. Geometry model Santori was created on low level model geometry through the use of X-ray photograph. Principle of geometry model type DePuy Proxima was real Femoral stem endoprosthesis which was scanned on scanner ATOS. Geometry of both these replacements were set up in program Rhinoceros 4.0 and then execute in program CatiaV5R17. Data for geometry model of femur were gained from CT chains. Material model of femur have been crated in two variants. The first one looks at structure bone tissues and the second one were created by Gruen´s zones. Femoral Stem was weighted by static equivalent resultant force acting in hip joint. Computational model of system and self solution, including depiction results, was done by ANSYS Workbench 11.0 for four model variants.

Dans l'étude du vieillissement cérébral, le suivi des personnes âgées est soumis à une forte sélection avec un risque de décès associé à de faibles performances cognitives. La modélisation de l'histoire naturelle du vieillissement cognitif est complexe du fait de données longitudinales et données de survie incomplètes. Par ailleurs, un déclin accru des performances cognitives est souvent observé avant le diagnostic de démence sénile, mais le début de cette accélération n'est pas facile à identifier. Les profils d'évolution peuvent être variés et associés à des risques différents de survenue d'un événement; cette hétérogénéité des déclins cognitifs de la population des personnes âgées doit être prise en compte. Ce travail a pour objectif d'étudier des modèles conjoints pour données longitudinales et données de survie incomplètes afin de décrire l'évolution cognitive chez les personnes âgées. L'utilisation d'approches à variables latentes a permis de tenir compte de ces phénomènes sous-jacents au vieillissement cognitif que sont l'hétérogénéité et l'accélération du déclin. Au cours d'un premier travail, nous comparons deux approches pour tenir compte des données manquantes dans l'étude d'un processus longitudinal. Dans un second travail, nous proposons un modèle conjoint à état latent pour modéliser simultanément l'évolution cognitive et son accélération pré-démentielle, le risque de démence et le risque de décès. / In cognitive ageing study, older people are highly selected by a risk of death associated with poor cognitive performances. Modeling the natural history of cognitive decline is difficult in presence of incomplete longitudinal and survival data. Moreover, the non observed cognitive decline acceleration beginning before the dementia diagnosis is difficult to evaluate. Cognitive decline is highly heterogeneous, e.g. there are various patterns associated with different risks of survival event. The objective is to study joint models for incomplete longitudinal and survival data to describe the cognitive evolution in older people. Latent variable approaches were used to take into account the non-observed mechanisms, e.g. heterogeneity and decline acceleration. First, we compared two approaches to consider missing data in longitudinal data analysis. Second, we propose a joint model with a latent state to model cognitive evolution and its pre-dementia acceleration, dementia risk and death risk.