9 An Example Model - Utility Learning Agents

9.2 Formal Structure

• PS is ,

• L is the language of functions expressible using: +, -, *, /, max, min, log, exp, average, cutbetween, a random assortment of constants between 0 and 100, and the amount spent on products 1 and 2.

• is defined by the evaluation of the expressions in L. So if then

• The distance function, d, is the RMSE of the past predictions of a model and the actual utility resulting

• Complexity, C, is estimated by the maximum depth of an expression in L.

• The volume is estimated by the number of different products mentioned (0, 1 or 2) in an expression of L.

The action is determined by a limited binary search for the spending pattern that the model predicts will return the best utility. The cost of action inference is thus represented by the number of binary search refinements.

Generated with CERN WebMaker