unisa ITA  unisa ENG


STATLAB

Scientific research by area

The activities of the STATLAB range in different areas, thanks to the research carried out by its members. For a detailed list of scientific publications, please refer to the personal web pages of the staff of STATLAB. An overview of the different research areas is given below.

  • Time Series Analysis: Stochastic Processes: inference and dynamic properties. Nonlinear time series models. Evaluation and combination of forecasts. 
  • Social Statistics: Questionnaire Design and Surveys Sampling. Data and Web Mining.. Multidimensional Data Analysis. Conjoint analysis. Social Network Analysis, Structural Equation Modeling 
  • Statistical Methods for Finance and Risk Management: Modelling and forecasting financial volatility. Tail risk forecasting. Backtesting. Multivariate models for dynamic portfolio optimization. Quantitative Finance. 
  • Methods and models for heterogeneous data: Mixture models. Model-based clustering. Algorithms for classification and clustering of large scale dataset. Rrobustness and noise-resistant data analysis methods.
  • Methods and models for time-to-event data: Models for truncated and censored data. Bivariate survival analysis under truncation and censoring. Joint models for longitudinal and time-to-event data.
  • Nonparametric Estimation and Inference: Free distribution inference based on resampling methods and empirical likelihood techniques. Kernel regression and local polynomial regression. Neural networks. Nonparametric methods for dependent data.
  • Analysis of high-dimensional models: Methods for choosing the relevant variables. Asymptotic analysis of models and estimators when the number of parameters diverges to infinity.
  • Analysis of Big Data: Methods and Algorithms for large-scale experiments. Models for heterogeneous data. Feature selection and efficient data reduction methods. Robustness and noise-free methods for massive datasets. Statistical inference for big data.