Highsupport

Highsupport is a term used in data mining and association rule mining to describe itemsets that occur frequently within a transactional database. The concept centers on a user-defined minimum support threshold; itemsets whose observed frequency in the data meets or exceeds this threshold are considered frequent. In practice, analysts refer to “high-support” itemsets to distinguish them from rare or infrequent patterns.

Formally, given a database D of transactions and an itemset X, the support of X is Supp(X) = |{t in D : X ⊆ t}| / |D|, the proportion of transactions containing X. An itemset is frequent if Supp(X) ≥ min_sup, where min_sup is the minimum support threshold chosen for the analysis. Some discussions use absolute counts instead of proportions, with min_sup defined as the required number of transactions.

High-support itemsets play a central role in frequent itemset mining algorithms, such as Apriori and FP-Growth. These methods rely on the monotonic property that all subsets of a frequent itemset are themselves frequent, enabling pruning of unlikely candidates and reducing computational effort. High-support thresholds shape the search space: higher thresholds yield fewer, more general patterns; lower thresholds reveal more, potentially noisy patterns.

Limitations and trade-offs should be noted. Setting a high minimum support may overlook rare but interesting associations, while a very low threshold can generate a large number of candidates and overfit to noise. Domain knowledge often guides the choice of min_sup, sometimes combined with additional measures like confidence or lift to evaluate the strength of discovered rules. Applications include market basket analysis, web usage mining, and bioinformatics.