Machine Learning / High Dimensional Statistics / Multivariate Analysis

Álvaro Veiga and collaborators

Exploiting low-rank structure in semi-definite programming to obtain an approximation of a data matrix by the product of a low rank matrix and a sparse matrix.

• Exploiting low-rank structure in semi-definite programming to obtain an approximation of a data matrix by the product of a low rank matrix and a sparse matrix. (with Mario Souto)
• Boost: boosting smooth transition regression trees. This is a new non-parametric regression model that allow for the calculation of partial effects, i.e., the calculation of derivatives in respect to the input variables. (with Yuri Resende, Gabriel Vasconcelos e Marcelo Medeiros)
• Copulas in Time Series Models with applications to scenario generation for energy sources (hydro, wind, etc). (with Guilherme Pereira)