What are the criteria of frequent pattern mining?

There are several criteria of frequent pattern mining which are as follows −

Based on the completeness of patterns to be mined − It can mine the whole collection of frequent itemsets, the closed frequent itemsets, and the maximal frequent itemsets, provided a minimum support threshold.

It can also extract constrained frequent itemsets (It can satisfy a collection of user-defined constraints), approximate frequent itemsets (It can change only approximate support counts for the mined frequent itemsets), near-match frequent itemsets (It can count the support count of the relatively matching itemsets), top-k frequent itemsets (i.e., the k most frequent itemsets for a user-specified value, k), etc.

There are several applications can have multiple requirements regarding the integrity of the patterns to be mined, which can lead to different computations and optimization approaches.

Based on the levels of abstraction included in the rule set − There are several methods for association rule mining can discover rules at multiple levels of abstraction. For instance, consider that a group of association rules mined involves the following rules where X is a variable defining a customer −

buys(X, “computer”) ⇒ buys(X, “HP printer”)

buys(X, “laptop computer”) ⇒ buys(X, “HP printer”)

Based on the number of data dimensions included in the rule − If the items or attributes in an association rule reference only one dimension, thus it is a single-dimensional association rule.

Based on the types of values handled in the rule − If a rule contains associations among the existence and absence of items, it is a Boolean association rule. If a rule defines associations among quantitative items or attributes, thus it is a quantitative association rule. In these rules, quantitative values for items or attributes are separated into intervals.

Based on the kinds of rules to be mined − Frequent pattern analysis can create several kinds of rules and different interesting relationships. Association rules are the famous kind of rules generated from frequent patterns.

Based on the kinds of patterns to be mined − Several kinds of frequent patterns can be mined from multiple kinds of data sets. The main target is on frequent itemset mining, that is, the mining of frequent itemsets (sets of items) from transactional or relational data sets.

Sequential pattern mining searches for frequent subsequences in a sequence data set, where a sequence data an ordering of events. For instance, with sequential pattern mining, it can study the series in which items are generally purchased. For example, users can tend to first purchase a PC, followed by a digital camera, and then a memory card.