fp_growth library

A comprehensive, memory-efficient implementation of the FP-Growth algorithm for frequent itemset mining and association rule generation.

Classes

AssociationRule<T>: Represents an association rule: antecedent => consequent.
FPGrowth<T>: Implements the FP-Growth algorithm for mining frequent itemsets.
RuleGenerator<T>: Generates association rules from a set of frequent itemsets.

buildConditionalTransactions(Map<List<int>, int> conditionalPatternBases, Map<int, int> conditionalFrequentItems) → Map<List<int>, int>: Builds weighted conditional transactions from conditional pattern bases.
calculateAbsoluteMinSupport(double minSupport, int transactionCount) → int: Calculates the absolute minimum support count from a relative or absolute value.
filterFrequentItems(Map<int, int> frequency, int absoluteMinSupport) → Map<int, int>: Filters a frequency map to include only items that meet the absoluteMinSupport.
generateSubsets(List<FPNode> nodes) → List<List<FPNode>>: Generates all non-empty subsets for a given list of FPNodes.
mineForItem<T>(FPTree tree, int item, List<int> prefix, Map<int, int> frequency, int absoluteMinSupport, ItemMapper<T> mapper, Logger logger) → Map<List<int>, int>: Recursively mines frequent itemsets for a specific item and its conditional tree.
mineLogic<T>(FPTree tree, List<int> prefix, Map<int, int> frequency, int absoluteMinSupport, ItemMapper<T> mapper, Logger logger) → Map<List<int>, int>: The core recursive mining logic of the FP-Growth algorithm.
transactionsFromCsv(Stream<List<int>> csvStream, {String fieldDelimiter = ',', String eol = '\n'}) → Stream<List<String>>: Parses a CSV byte stream efficiently.