Research

Research Interests

• Generalized regression for time series analysis with VGLMs/VGAMs
• Statistical modeling with VGLMs/VGAMs
• Hierarchical and dynamic regression modeling for longitudinal data
• Computational statistics (R, Matlab, Python)

Research Summary

Victor is a (computational) statistician with a special interest in developing new methods and software for vector generalized linear and additive models (VGLMs/VGAMs), and some extensions directed to some popular data types. At present some outcomes refer to time series (with interest in forecasting and model selection methods), as well as cross-sectional and longitudinal data.

Current Research Projects Focus

• Developing new methodology for forecasting and model selection for vector generalized linear time series models (VGLTSMs)
• developing new VGLM modeling choices, e.g., new link functions for the mean and the quantiles of several 1-parameter distributions. The latter arises as an alternative to quantile regression, exceedingly addressed in the literature with many variants, but no overriding framework.

VGLTSMs are a new sub-class of VGLMs that handle time series data, introduced in my PhD thesis (2018) along with its companion software in R, developed in collaboration with Thomas W. Yee (University of Auckland). Unlike other software to fit time series, particularly those relying on optim() and arima(), VGLTSMs are estimated by MLE using Fisher scoring, and are designed to fit popular TS models (linear and non-linear, e.g., IGARCHs) as special cases, and accommodates covariates in the analysis allowing to model empirical features of stochastic volatility.

For more info,, please check