Algorithm 1: Dataset reduction using clustering.

Input:Dold, t, n

Output:Dnew

Function DatasetReduction(Dold, t, n): Dnew ← [ ]

for c in findClasses(Dold) do

Xc ← extractClassPoints(Dold,c)

db ← DBSCAN.fit(Xc)

for l in db.clusters() do

m ← size(l)/size(Xc)

if m > t do

Xp ← db.extractClusterPoints(p)

q ← createCentroid(Xp)

dist ← distances(q, Xp) Dnew ← addClosestNPoints(Dnew, dist, Xp, n)

end

end

end

end