
Decision Trees from Data Streams 125
Algorithm 23: The algorithm to compute P (x
j
|C
k
) for numeric at-
tribute x
j
and class k at a given leaf.
input : BTree: Binary Tree for attribute x
j
nrExs: Vector of the number of examples per Class
X
h
: the highest value of x
j
observed at the Leaf
X
l
: the lowest value of x
j
observed at the Leaf
N
j
: the number different values of x
j
observed at the Leaf
output: Counts The vector of size Nintervals with the percentage of
examples per interval
begin
if BT ree == NULL then return 0
/* number of intervals */
Nintervals ← min(10, |BT ree|)
/* interval range */
inc ←
X
h
−X
l
Nintervals
for i = 1 to Nintervals do
Counts[i] ← LessT han(x