/AMPL-1

Constrained portfolio optimization and machine learning with linear regression classification

AMPL Projects

  1. Machine learning classification system for evaluating recessionary economic states with X matrix as FRED time series economic indicators and y vector as NBER recessionary periods. Method optimizes theta parameters by maximizing/(minimizing) average positive/(negative) Euclidean distances to a hyperplane.

  2. Optimization of rebalanced portfolio weights to the Russell Global Index under sector, capitalization, transaction cost, and tax lot constraints.