pattern discovery task the task of pattern discovery is to find all local patterns from a given pattern language that satisfy the required conditions. A prototypical instantiation of this task is the task of finding frequent itemsets (sets of items, such as {bread, butter}), which are often found together in a transaction (e.g., a market basket) (Aggrawal et al 1993). The condition that a pattern (itemset) has to satisfy in this case is to appear in (hold true for) a sufficiently high proportion (called support and denoted by s) of the transactions in the input dataset. With the increasing interest in mining complex data, mining frequent patterns is also considered for structured data. We can thus talk about mining frequent subsequences or mining frequent subgraphs in sequence or graph data. We can consider as frequency the multiple occurrences of a pattern in a single data structure (e.g., sequence or graph) or the single occurrences of a pattern in multiple data structures. data mining task