Methodological developments and applications for the measurement of non-response in surveys and for the correction of their effects, especially the biases that can be introduced in the inferential process.
Discovery of individual behaviours from aggregated information.
Application and development of models for electoral prediction.
Longitudinal and survival data joint models with time or survival targets are studied, with special emphasis on dynamic predictive targets.
Our group addresses the problem of model selection from a Bayesian target point of view. We particularly work on the study of criteria that allow us to establish optimal prior distributions in order to carry out an effective selection and modeling.
Methodological and software developments to find solutions to the problem of partial non-response in surveys and gaps in databases.
Methodological developments and applications for active listening on the Internet, favouring the identification of feelings and authors to convert the huge flows of data and opinions into information.
Development of statistical methodologies for the rapid and reliable detection of influenza epidemics. We approach the inference and prediction of these models from the Bayesian paradigm, which allows the implementation of complex models with spatial, temporal and hierarchical structures.
Development of models to predict the spatial and spatial-temporal distribution of species. The incorporation of uncertainty in covariates, the problems generated by missing values, the effect of preferential sampling and the handling of large volumes of data are addressed.
Analysis and preparation of opinion surveys and polls.