The College of Computer Science and Information Technology at University of Basrah held a seminar on the representation of individuals in society.
The seminar in which graduate student Adhraa Qais Obaid lectured dealt with improving the k-means algorithm using swarm intelligence algorithms with individuals of variable length.
It showed that despite the simplicity and speed of the algorithm and its implementation, it has some limitations that lead to different clustering results depending on the values chosen. To address these limitations, this study proposed an approach that takes advantage of swarm intelligence algorithms.
The study contributed to modifying the SMO, WOA, and GWO algorithms to deal with individuals with variable height, in addition to improved population initialization using the GPS large point set method instead of random initialization.
It was concluded that the proposed techniques outperform the k-means algorithm and other techniques in terms of accuracy and clustering ability. It is worth noting that it is part of a research work submitted to obtain a master’s degree in computer science.
