
182 Knowledge Discovery from Data Streams
Figure 11.8: The main steps in SAX. a) Piecewise Aggregate Approximation;
b) Symbolic Discretization; c) The output SAX string.
schema is: If a point is less than the smallest breakpoint, then it is denoted
as a. Otherwise and if the point is greater than the smallest breakpoint and
less than the next larger one, then it is denoted as b, etc.
11.4.1.3 Distance Measure
The output of the second step in SAX is a string. How can we work with
the string? How can we define a sound metric to work with strings?
The following distance measure is applied when comparing two different
SAX strings:
MIN DIST (
ˆ
Q,
ˆ
C) =
v
u
u
t
w
n
w