A5082624229- Statistical Methods and Inference
- Advanced Statistical Methods and Models
- Direction-of-Arrival Estimation Techniques
- Point processes and geometric inequalities
- Bayesian Modeling and Causal Inference
- Gene Regulatory Network Analysis
- Morphological variations and asymmetry
- Stochastic processes and statistical mechanics
- Sparse and Compressive Sensing Techniques
- Bayesian Methods and Mixture Models
- Blind Source Separation Techniques
- Markov Chains and Monte Carlo Methods
- RNA and protein synthesis mechanisms
- Financial Risk and Volatility Modeling
- interferon and immune responses
- RNA modifications and cancer
- Optical Imaging and Spectroscopy Techniques
- Statistical Methods and Bayesian Inference
- Systemic Lupus Erythematosus Research
- Fault Detection and Control Systems
- Control Systems and Identification
- Probabilistic and Robust Engineering Design
- Protein Structure and Dynamics
- Stochastic processes and financial applications
- Neural dynamics and brain function
University of Copenhagen
2014-2025
Copenhagen Business School
2018
John Wiley & Sons (United States)
2018
Hudson Institute
2018
Rigshospitalet
2014
Copenhagen University Hospital
2014
Due to its low computational cost, Lasso is an attractive regularization method for high-dimensional statistical settings. In this paper, we consider multivariate counting processes depending on unknown function parameter be estimated by linear combinations of a fixed dictionary. To select coefficients, propose adaptive $\ell_{1}$-penalization methodology, where data-driven weights the penalty are derived from new Bernstein type inequalities martingales. Oracle established under assumptions...
Transcripts from spacer sequences within chromosomal repeat clusters [CRISPRs (clusters of regularly interspaced palindromic repeats)] archaea have been implicated in inhibiting or regulating the propagation archaeal viruses and plasmids. For crenarchaeal thermoacidophiles, spacers show a high level matches (∼30%) with viral plasmid genomes. Moreover, their distribution along virus/plasmid genomes, as well DNA strand specificity, appear to be random. This is consistent hypothesis that are...
Abstract Background A central question in molecular biology is how transcriptional regulatory elements (TREs) act combination. Recent high-throughput data provide us with the location of multiple regions for regulators, and thus possibility analyzing multivariate distribution occurrences these TREs along genome. Results We present a model TRE known as Hawkes process. illustrate use this by two different publically available sets. are able to model, detail, occurrence one affected others, we...
Common Chromosomal Fragile Sites (CFSs) are specific genomic regions prone to form breaks on metaphase chromosomes in response replication stress. Moreover, CFSs mutational hotspots cancer genomes, showing that the mechanisms operate at highly active cells. Orthologs of human found a number other mammals, but extent CFS conservation beyond mammalian lineage is unclear. Characterization from distantly related organisms can provide new insight into biology underlying CFSs. Here, we have mapped...
ABSTRACT In dose‐response modeling, several models can often yield satisfactory fits to the observed data. The current practice in risk assessment is use model averaging, which a way combine multiple weighted average. A key parameter benchmark dose, dose resulting predefined abnormal change response. Current when applying frequentist averaging weights based on Akaike Information Criterion (AIC). This paper introduces stacking as an alternative for modeling and generalizes Diversity Index...
We give a causal interpretation of stochastic differential equations (SDEs) by defining the postintervention SDE resulting from an intervention in SDE. show that under Lipschitz conditions, solution to is equal uniform limit probability structural equation models based on Euler scheme original SDE, thus relating our definition mainstream concepts. prove when driving noise Lévy process, distribution identifiable generator
Symmetric independence relations are often studied using graphical representations. Ancestral graphs or acyclic directed mixed with $m$-separation provide classes of symmetric models that closed under marginalization. Asymmetric appear naturally for multivariate stochastic processes, instance, in terms local independence. However, no class representing such asymmetric relations, which is also marginalization, has been developed. We develop the theory $\mu $-separation and show this provides...
We give sufficient criteria for the Dol\'eans-Dade exponential of a stochastic integral with respect to counting process local martingale be true martingale. The are adapted particularly case processes and sufficiently weak useful verifiable, as we illustrate by several examples. In particular, allow construction example nonexplosive Hawkes well intensities depending on diffusion processes.
In this note we show a new version of the trek rule for continuous Lyapunov equation. This linear matrix equation characterizes cross-sectional steady-state covariance Gaussian Markov process, and links graphical structure drift process to entries matrix. general, is power series expansion matrix, while special case where acyclic, it simplifies polynomial in off-diagonal Using can give relatively explicit formulas some cases Furthermore, use derive lower bound variances acyclic case.
A discrete-time approximation scheme called local linearization of the Langevin diffusion on Rk is considered, with emphasis ergodic properties considered as a Markov chain. We will derive criteria for to be geometrically ergodic, and illustrate use these by means examples. Furthermore, we discuss in relation other schemes such discretization proposals Metropolis-Hastings algorithm.
We study a class of graphs that represent local independence structures in stochastic processes allowing for correlated error processes. Several may encode the same independencies and we characterize such equivalence classes graphs. In worst case, number conditions our characterizations grows superpolynomially as function size node set graph. show deciding Markov is coNP-complete which suggests cannot be improved upon substantially. prove global property case multivariate Ornstein-Uhlenbeck...
We consider local alignments without gaps of two independent Markov chains from a finite alphabet, and we derive sufficient conditions for the number essentially different with score exceeding high threshold to be asymptotically Poisson distributed. From approximation Gumbel maximal alignment is obtained. The results extend those obtained by Dembo, Karlin Zeitouni [Ann. Probab. 22 (1994) 2022–2039] sequences i.i.d. variables.
Identification of the primary tumor site in patients with metastatic cancer is clinically important, but remains a challenge. Hence, efforts have been made towards establishing new diagnostic tools. Molecular profiling promising approach, tissue heterogeneity and inadequacy may negatively affect accuracy usability molecular classifiers. We developed validated microRNA‐based classifier, which predicts liver biopsies, containing limited number cells. Concurrently we explored influence...
The partial copula provides a method for describing the dependence between two random variables $X$ and $Y$ conditional on third vector $Z$ in terms of nonparametric residuals $U_1$ $U_2$. This paper develops test independence by combining with quantile regression based estimating residuals. We consider statistic generalized correlation $U_2$ derive its large sample properties under consistency assumptions procedure. demonstrate through simulation study that resulting is sound complicated...
Une représentation des degrés de liberté comparable au lemme Stein est donnée pour une classe d’estimateurs du paramètre la moyenne dans $\mathbb{R}^{n}$. Contrairement aux résultats précédents, notre valable famille discontinus. Cela montre que même si les discontinuités sont mesure Lebesgue zéro, elles ne peuvent pas être ignorées lors calcul liberté. Les estimateurs avec apparaissent naturellement modèles régression sélection variables par données utilisée. Deux tels exemples, meilleur...
Abstract Motivation: Contamination of a cancer tissue by the surrounding benign (non-cancerous) is concern for molecular diagnostics. This because an observed signature will be distorted tissue, possibly leading to incorrect diagnosis. One example identification primary tumor site metastases biopsies typically contain significant amount tissue. Results: A model contamination presented. works independently training predictor, and it can combined with any predictor model. The usability...
Abstract Objective: To investigate the relationship between physical fitness and work integration following stroke. Design: Single-group study, measurement of pre post training, employment status in a follow-up assessment 2 to 36 months after rehabilitation. Setting: Interdisciplinary outpatient rehabilitation program. Participants: 58 stroke survivors (62% male, mean age at program start 46.7 years, time since 1.1 years) who were consecutively referred Intervention: 1½ hours intensive...
We give a general result on the effective degrees of freedom for nonlinear least squares estimation, which relates to divergence estimator. show that in framework, estimator is well defined but potentially negatively biased estimate freedom, and we an exact representation bias. This implies if use as plug-in Stein's unbiased risk (SURE), generally underestimate true risk. Our applies, instance, model searching problems, yielding finite sample characterization how much search contributes...
Conditional local independence is an asymmetric relation among continuous time stochastic processes. It describes whether the evolution of one process directly influenced by another given histories additional processes, and it important for description learning causal relations We develop a model-free framework testing hypothesis that counting conditionally locally independent process. To this end, we introduce new functional parameter called Local Covariance Measure (LCM), which quantifies...
Learning large scale nonlinear ordinary differential equation (ODE) systems from data is known to be computationally and statistically challenging. We present a framework together with the adaptive integral matching (AIM) algorithm for learning polynomial or rational ODE sparse network structure. The allows time course sampled multiple environments representing e.g. different interventions perturbations of system. AIM combines an initial penalised step adapted least squares based on solving...