Particle Swarm Optimization based K-Prototype Clustering Algorithm

摘要: Clustering in data mining is a discovery process that groups set of so as to maximize the intra- cluster similarity and minimize inter-cluster similarity. The K-Means algorithm best suited for clustering large numeric sets when at possess only values. K-Modes extends domain categorical. But some applications, objects are described by both categorical features. K-Prototype one most important algorithms this type data. This produces locally optimal solution dependent on initial prototypes order object Particle Swarm Optimization simple optimization techniques, which can be effectively implemented enhance results. discrete or binary mechanisms useful handle mixed set. leads better cost evaluation description space subsequently enhanced processing Optimization. paper proposes new variant reach global problem. proposed evaluated standard benchmark dataset taken from UCI machine learning repository. comparative analysis proved based provides performance than traditional K-modes K- Prototype algorithms.