Learning concise representations for regression by evolving networks of trees

William La Cava, Tilak Raj Singh, James Taggart +2

We propose and study a method for learning interpretable representations for the task of regression. Features are represented as networks of multi-type expression trees comprised o…