An Improved Pheromone-Based Kohonen Self- Organising Map in Clustering and Visualising Balanced and Imbalanced Datasets

The data distribution issue remains an unsolved clustering problem in mining, especially dealing with imbalanced datasets. Kohonen Self-Organising Map (KSOM) is one of the well-known algorithms that can solve various problems without a pre- defined number clusters. However, similar to other algorithms, this algorithm requires sufficient for its unsupervised learning process. inadequate amount class label dataset significantly affects process, leading inefficient and unreliable results. Numerous research have been conducted by hybridising optimising KSOM optimisation techniques. Unfortunately, are still unsolved, separation boundary overlapping Therefore, proposed improved pheromonebased PKSOM known as iPKSOM mentioned problem. Six different datasets, i.e. Iris, Seed, Glass, Titanic, WDBC, Tropical Wood datasets were chosen investigate effectiveness algorithm. All observed compared original This modification impacted process improving refining scatteredness reducing be implemented complex such high dimensional streaming data.