Learning from Data by Interval Linear Programming
The linear programming based method are popular methods for learning from empirical data (observations, samples, examples, records). In this paper, an interval linear programming based method for regression problems is proposed. The explicit representation of the general optimal solution of regression problem is obtained in terms of a generalized inverse of the constraint matrix. This explicit solution has obvious theoretical (and possibly computational) advantages over the well-known iterative methods of linear programming.
L. Wei "Learning from Data by Interval Linear Programming", Key Engineering Materials, Vols. 439-440, pp. 710-714, 2010