Mining for Mutually Exclusive Items in Transaction Databases

Association rule mining is a popular task that involves the discovery of co-occurences of items in transaction databases. Several extensions of the traditional association rule mining model have been proposed so far; however, the problem of mining for mutually exclusive items has not been directly tackled yet. Such information could be useful in various cases (e.g., when the expression of a gene excludes the expression of another), or it can be used as a serious hint in order to reveal inherent taxonomical information. In this article, we address the problem of mining pairs of items, such that the presence of one excludes the other. First, we provide a concise review of the literature, then we define this problem, we propose a probability-based evaluation metric, and finally a mining algorithm that we test on transaction data. IntroductIon Association rules are expressions that describe a subset of a transaction database. When mining for such patterns, it is quite often that we come up with a large number of rules that appear to be too specific and not very interesting. A rule that relates two specific products in a market basket database is not very likely to be really strong compared to a rule that relates two groups or two families of products. Hierarchical relationships among items in a database can be used in order to aggregate the weak, lower-level rules into strong, higher-level rules, producing hierarchical, multiple level, or generalized association rules. However, such information is not always explicitly provided, although it might exist. Mining for taxonomies is a really challenging task that, to the best of our knowledge, has not been approached yet. Taxonomies are conceptual