One-pass algorithm: Difference between revisions

Revision as of 06:44, 16 April 2019

Revision as of 19:13, 19 November 2019

Line 4:

Basically one-pass algorithm operates as follows:

(1)# theThe object descriptions are processed serially;

(2)# theThe first object becomes the cluster representative of the first cluster;

(3)# eachEach subsequent object is matched against all cluster representatives existing at its processing time;

# A given object is assigned to one cluster (or more if overlap is allowed) according to some condition on the matching function;

its processing time;

(4)# awhen givenan object is assigned to onea cluster (orthe morerepresentative iffor overlapthat iscluster allowed)is accordingrecomputed;

(6)# if an object fails a certain test it becomes the cluster representative of a new cluster

to some condition on the matching function;

cluster (7)# nothing happened

(5) when an object is assigned to a cluster the representative for that cluster is

recomputed;

(6) if an object fails a certain test it becomes the cluster representative of a new

cluster (7) nothing happened

==Example problems solvable by one-pass algorithms==

Given any list as an input: