Clustering categorical data using a swarm-based method

The K-Modes algorithm is one of the most popular clustering algorithms in dealing with categorical data. But the random selection of starting centers in this algorithm may lead to different clustering results and falling into local optima. In this paper we proposed a swarm-based K-Modes algorithm. The experimental results over two well known Soybean and Congressional voting categorical data sets show that our method can find the optimal global solutions and can make up the K-Modes shortcoming.