Cristian Preda

Learn More
This paper proposes the first model-based clustering algorithm for multivariate functional data. After introducing multivariate functional principal components analysis (MFPCA), a parametric mixture model, based on the assumption of normality of the principal components, is defined and estimated by an EM-like algorithm. The main advantage of the proposed(More)
The main contributions to functional data clustering are reviewed. Most approaches used for clustering functional data are based on the following three methodologies: dimension reduction before clustering, nonparametric methods using specific distances or dissimilarities between curves and model-based clustering methods. These latter assume a probabilistic(More)
In Parkinson's disease, precise analysis of gait disorders remains essential for the diagnostic or the evaluation of treatments. During a gait analysis session, a series of successive dynamic gait trials are recorded and data involves a set of continuous curves for each patient. An important aspect of such data is the infinite dimension of the space data(More)
Every year adverse drug events (ADEs) are known to be responsible for 98,000 deaths in the USA. Classical methods rely on report statements, expert knowledge, and staff operated record review. One of our objectives, in the PSIP project framework, is to use data mining (e.g., decision trees) to electronically identify situations leading to risk of ADEs.(More)
Missing data is a common feature of large data sets in general and medical data sets in particular. Depending on the goal of statistical analysis, various techniques can be used to tackle this problem. Imputation methods consist in substituting the missing values with plausible or predicted values so that the completed data can then be analysed with any(More)
Model-based clustering for functional data is considered. An alternative to model-based clustering using the functional principal components is proposed by approximating the density of functional random variables. An EM-like algorithm is used for parameter estimation and the maximum a posteriori rule provides the clusters. Real data applications illustrate(More)
Geographical variations in Crohn’s Disease incidence have been reported worldwide suggesting putative variations in the distribution of environmental risk factors. A spatial heterogeneity in standardized incidence ratios of Crohn’s Disease was previously detected in northern France. The goals of this study were to highlight significant atypical clusters in(More)
Plasmodium falciparum malaria in India is characterized by high rates of severe disease, with multiple organ dysfunction (MOD)—mainly associated with acute renal failure (ARF)—and increased mortality. The objective of this study is to identify cytokine signatures differentiating severe malaria patients with MOD, cerebral malaria (CM), and cerebral malaria(More)