L. Amorosi, T. Padellini
Recent developments in the interplay between Operational Research and Statistics allowed to exploit advances in Mixed Integer Optimization (MIO) solvers to improve the quality of statistical analysis. In this work we tackle Canonical Correlation Analysis (CCA), a dimensionality reduction method that summarises multiple data sources jointly, retaining their dependency structure. We propose a new technique to encode Sparsity in CCA by means of a new mathematical programming formulation which allows to obtain an exact solution using readily available solvers (like Gurobi). We show a preliminary investigation of the performance of our proposal through a simulation study, which highlights the potential of our approach.
Canonical Correlation Analysis, Mixed Integer Optimization, Sparsity
Tipo di pubblicazione: