R - apriori() support vs coverage?

Question

I am using apriori to find assocaition rules and I am running into an issue:

| rule           | support           | confidence   | lift            | coverage       |
|---------------|--------------------|------------------|--------------|-------------------|
| {A} => {B} | 8.616999e-05 | 0.01502544 | 19.11896 | 0.005734940 |
| {A} => {C} | 8.944227e-05 | 0.01559602 | 49.05084 | 0.005734940 |

The manual states that Coverage: Provides the generic function and the needed S4 method to calculate the coverage (support of the left-hand-side) of rules.

For a small ruleset coverage is equal to support. Why does coverage differ from support for large rulesets?

Answer 1

From the man page for interestMeasure:

"coverage", cover, LHS-support Support of the left-hand-side of the rule, i.e., supp(X). A measure of to how often the rule can be applied. Range: [0, 1]

So, coverage is greater or (in some rare cases) equal to the support of the rule.

Sorry that this is somewhat confusing in the documentation, but, you know, we all love to maintain documentation. I will improve the documentation for coverage in the next release...