# New clusterization of global seaport countries based on their DEA and FDEA network efficiency scores

## TL;DR

This paper clusters 133 countries by seaport network efficiency using DEA and FDEA scores, revealing four distinct connectivity groups.

## Contribution

The study introduces hkmeans clustering for seaport network efficiency, creating new connectivity clusters (LC, MC, HC, VHC).

## Key findings

- hkmeans clustering is shown to be consistent and practical for grouping seaport efficiency data.
- 24, 47, 40, and 22 countries were classified into low, medium, high, and very high connectivity clusters, respectively.
- The method works effectively with both fuzzy and non-fuzzy datasets.

## Abstract

Global seaport network efficiency can be measured using the Liner Shipping Connectivity Index (LSCI) with Gross Domestic Product. This paper utilizes k-means and hierarchical strategies by leveraging the results obtained from Data Envelopment Analysis (DEA) and Fuzzy Data Envelopment Analysis (FDEA) to cluster 133 countries based on their seaport network efficiency scores. Previous studies have explored hkmeans clustering for traffic, maritime transportation management, swarm optimization, vessel trajectory prediction, vessels behaviours, vehicular ad hoc network etc. However, there remains a notable absence of clustering research specifically addressing the efficiency of global seaport networks. This research proposed hkmeans as the best strategy for the seaport network efficiency clustering where our four newly founded clusters; low connectivity (LC), medium connectivity (MC), high connectivity (HC) and very high connectivity (VHC) are new applications in the field. Using the hkmeans algorithm, 24 countries have been clustered under LC, 47 countries under MC, 40 countries under HC and 22 countries under VHC. With and without a fuzzy dataset distribution, this demonstrates that the hkmeans clustering is consistent and practical to form grouping of general data types. The findings of this research can be useful for researchers, authorities, practitioners and investors in guiding their future analysis, decision and policy makings involving data grouping and prediction especially in the maritime economy and transportation industry.