Inside the space explorer, you can test how the rule affects other parameters.
For example, here, you see a positive effect on NOx and CO. Both are lower when you accept the rule.
- Risk: Absolute value of a negative rank. This is the worst range - there are mostly bad points in this range, within the rule context. The red area is the exact opposite of a green area of an high ranked variable.
- Redundancy: Max (rank in the job - rank in the rule, 0). If redundancy is high, the variable was interesting in the analysis, but its effect inside the rule is hidden by other variables. Adding such a variable to the rule, will not change much, but will increase the robustness. The higher the value, the more the ability to separate the bad ones in the job exploited by the rule. One or more dimensions of the rule may be redundant to this variable.
- Presence: percentage of points of the total population between the lowest and highest extreme point (of the variable under consideration) of the population that complies with the rule. This gives an indication of the extent of the range of the variable when the rule is followed, relative to the extent of the current range of that variable.