Abstract
This research is addresses to determine the dominant species that located in the overlapped clusters produced by the Kohonen Self-organizing Map (KSOM). Before, KSOM algorithm able to cluster the tropical wood species data set effectively and accurately according to the wood features, which is wood pores sizes. Unfortunately, there are seven overlapped clusters in the clustering result and this is due to similarity features among the wood species. This problem has caused difficulty in determining the separation boundary amongst clusters, where the most dominant species for every overlapped cluster is difficult to identify. As for a solution, the Histogram Intersection (HI) method is proposed in this research to solve this problem. From the experiments, the HI has proved that it can determine the dominant species of the overlapped clusters effectively resulting in an improvement of 1.12% of the classification accuracy, compared with the clustering result without HI technique. It has shown the implementation of this method has successfully solved the problem occurred.
Original language | English |
---|---|
Title of host publication | IECON 2015 - 41st Annual Conference of the IEEE Industrial Electronics Society |
Publisher | Institute of Electrical and Electronics Engineers Inc. |
Pages | 3075-3080 |
Number of pages | 6 |
ISBN (Electronic) | 9781479917624 |
DOIs | |
Publication status | Published - 2016 Jan 25 |
Event | 41st Annual Conference of the IEEE Industrial Electronics Society, IECON 2015 - Yokohama, Japan Duration: 2015 Nov 9 → 2015 Nov 12 |
Other
Other | 41st Annual Conference of the IEEE Industrial Electronics Society, IECON 2015 |
---|---|
Country | Japan |
City | Yokohama |
Period | 15/11/9 → 15/11/12 |
Keywords
- Clustering
- pheromone
- Pheromone-Based KSOM (PKSOM)
- Tropical Wood Species
ASJC Scopus subject areas
- Electrical and Electronic Engineering
- Industrial and Manufacturing Engineering