An association rule is a pattern that states when an event occurs, another event occurs with certain probability. The selection of association rules is based on the minimum confidence level, which is an indicator, which states that association rules below minimum threshold should be discarded. The association rule is given as X Y where X and Y are non-empty subset and X ∩ Y = Ø. The support for a rule is given by support(X ∪ Y) and the confidence is given as support(X ∪ Y) / support(X) [3]. In the association rules grid displayed in the application, if one of the strong rule is given as "Drug --> Liquor, Weapon" and confidence is 0.6; This means that if a crime of type Drug is recoded, then there is a 60% probability that the crime of type both Liquor and Weapon are also recorded.
The grid shown in Figure 20 displays the association rules based on the frequent item sets which satisfy the user supplied minimum confidence. The rules generated are based on the on-campus crime records for Alabama State University for a five year period from 2009 to 2013.