dm.cs.tu-dortmund.de/mlbits/frequent-pattern-apriori-hashtree/
The Apriori Hash Tree – Lecture Notes
refers to a child node at level \(d+1\)
root node is at level 1
hashing at level \(d\) is based on the \(d\) th smallest value in the \(k\) -itemset; at most \(k+1\) levels
Example
Hash Tree Index
Search [...] Frequent Itemsets using Hash Tree
Search all candidates that are contained in transaction \(T = (t_1 t_2\ldots t_m)\)
at the root node:
determine the hash value for each item in \(T\)
continue search in [...] contained in transaction \(T\)
Optimization: at every level \(d\) , only consider the hash codes of \((t_{i+1},\ldots ,t_{m-k+d})\) as possible next-smallest item, where \(t_i\) was the item last used for splitting …
