When working with [[Association Rules|association rules]], a high conviction [[Error Metrics|metric]] means that the consequent is highly depending on the antecedent. It can be derived from the [[Support of an Item Set|support]] and [[Confidence of an Association Rule|confidence]] metrics as: $ \text{conviction}(A \rightarrow C) = \frac{ 1 - \text{support}(C) }{ 1 - \text{confidence}(A \rightarrow C) } $ In the case of a perfect [[Confidence of an Association Rule|confidence]] score, the denominator becomes 0 and the conviction $\infty$. Similar to [[Lift of an Association Rule|lift]], if items are independent, the conviction is 1.