18.5 A good "straw man" learning algorithm is as follows: create a table out of all the training examples. Determine which output occurs most often among the training examples; call it d.

Then when given an input that is not in the table, just return

d. For inputs that are in the table, return the output associated with it (or the most frequent output, if there is more than one). If the input does not appear in the table, then return

d. Implement this algorithm and see how well it does in a sample domain. This should give you an idea of the baseline for the domain—the minimal performance that any algorithm should be able to obtain (although many published algorithms have managed to do worse).