Particle Swarm Optimization based K-Prototype ClusteringAlgorithm
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2015, Vol 17, Issue 2
Abstract
Abstract: Clustering in data mining is a discovery process that groups a set of data so as to maximize the intraclustersimilarity and to minimize the inter-cluster similarity. The K-Means algorithm is best suited forclustering large numeric data sets when at possess only numeric values. The K-Modes extends to the K-Meanswhen the domain is categorical. But in some applications, data objects are described by both numeric andcategorical features. The K-Prototype algorithm is one of the most important algorithms for clustering this typeof data. This algorithm produces locally optimal solution that dependent on the initial prototypes and order ofobject in the data. Particle Swarm Optimization is one of the simple optimization techniques, which can beeffectively implemented to enhance the clustering results. But discrete or binary Particle Swarm Optimizationmechanisms are useful for handle mixed data set. This leads to a better cost evaluation in the description spaceand subsequently enhanced processing of mixed data by the Particle Swarm Optimization. This paper proposesa new variant of binary Particle Swarm Optimization and K-Prototype algorithms to reach global optimalsolution for clustering optimization problem. The proposed algorithm is implemented and evaluated on standard benchmark dataset taken from UCI machine learning repository. The comparative analysis proved that ParticleSwarm based on K-Prototype algorithm provides better performance than the traditional K-modes and KPrototypealgorithms.
Authors and Affiliations
K. Arun Prabha , N. Karthi Keyani Visalakshi
Keywords
Related Articles
- EP ID EP158384
- DOI -
- Views 103
- Downloads 0
A Survey on Research work in Educational Data Mining
Abstract: Educational Data Mining is an emerging discipline that focuses on applying Data Mining tools andtechniques to educationally related data. The discipline focuses on analyzing educational data to developmod...
Positional Prediction of Individuals Present In a Crime Scene
Abstract: The scientific study/interpretation of bloodstain patterns at a crime scene, provide invaluable evidence for sequencing, reconstruction of events that might have occurred at the crime scene. the authors have tr...
Holistic and Procedural Features for Authenticating Users
Security systems help to protect machines or sensitive data from unauthorized users. The need for better and cheap security systems is growing with the growth in technologies and hacking skills of the people. Var...
Multi-Resolution Pruning Based Co-Location Identification In Spatial Data
A co-location spatial pattern is a pattern of multiple groups which co-relates spatial features or events that are frequently located in same zone. Co-location pattern mining emphasizes overall analysis bymanipulating th...
Particle Swarm Optimization based K-Prototype ClusteringAlgorithm
Abstract: Clustering in data mining is a discovery process that groups a set of data so as to maximize the intraclustersimilarity and to minimize the inter-cluster similarity. The K-Means algorithm is best suited f...