Compressing Decision Tables

Using common sense, people can often compress a larger decision table into a smaller one that produces identical results — an 18-rule table might collapse to just 6 rules once the redundant conditions are spotted. But once a table has more attributes (columns), doing this compression by hand becomes difficult or impossible, even allowing for some margin of error — you generally need a tool.

This challenge asks you to compress the following multi-hit decision table (18 rules, including the default). Each row applies to a “Type,” tests either an upper or lower bound on an “Adjustment” value and a “Loss” value, and classifies the case as TOP or BOTTOM:

Try to compress this table into fewer rules while preserving exactly the same classification behavior for every possible input.

Send your solutions to DecisionManagementCommunity@gmail.com, or open a pull request to add yours here.