
196 Big Data: Techniques and Technologies in Geoinformatics
Given a set of sequences, FSM algorithms aim at nding unique sets of ordered
itemsets whose frequencies are above a predened threshold termed as support value,
which can be either an absolute value or a percentage. The task becomes simpler
when only one item is allowed in the itemsets that are used as the building blocks for
frequent sequences. While allowing multiple items in itemsets provides higher ex-
ibility, using a single item per itemset may actually be more suitable for trajectory
data where road segments are typically unique in simple frequent sequences. The
Sequoia packag ...