Affiliations: Center for Soft Computing Research, Indian Statistical
Institute, Kolkata 700108, India. [email protected] | Machine Intelligence Unit and Center for Soft
Computing Research, Indian Statistical Institute, Kolkata 700108, India.
[email protected] | Department of Computer Science Engineering and
Jadavpur University, Kolkata 700032, India
Note: [] Address for correspondence: Center for Soft Computing Research
Indian Statistical Institute, Kolkata 700108, India
Abstract: The study of ant colonies behavior and their self-organizing
capabilities is of interest to machine learning community, because it provides
models of distributed adaptive organization which are useful to solve difficult
optimization and classification problems among others. Social insects like
ants, bees deposit pheromone (a type of chemical) in order to communicate
between the members of their community. Pheromone, that causes clumping
behavior in a species and brings individuals into a closer proximity, is called
aggregation pheromone. This article presents a new algorithm (called, APC) for
pattern classification based on this property of aggregation pheromone found in
natural behavior of real ants. Here each data pattern is considered as an ant,
and the training patterns (ants) form several groups or colonies depending on
the number of classes present in the data set. A new test pattern (ant) will
move along the direction where average aggregation pheromone density (at the
location of the new ant) formed due to each colony of ants is higher and hence
eventually it will join that colony. Thus each individual test pattern (ant)
will finally join a particular colony. The proposed algorithm is evaluated with
a number of benchmark data sets as well as various kinds of artificially
generated data sets using three evaluationmeasures. Results are compared with
four other well known conventional classification techniques. Experimental
results show the potentiality of the proposed algorithm in terms of all the
evaluation measures compared to other algorithms.
Keywords: Swarm intelligence, Ant colony optimization, Aggregation pheromone, Pattern classification
