A Similarity Measurement with Entropy-Based Weighting for Clustering Mixed Numerical and Categorical Datasets

Many mixed datasets with both numerical and categorical attributes have been collected in various fields, including medicine, biology, etc. Designing appropriate similarity measurements plays an important role clustering these datasets. traditional treat equally when measuring the similarity. However, different may contribute differently as amount of information they contained could vary a lot. In this paper, we propose measurement entropy-based weighting for The data are first transformed into by automatic categorization technique. Then, strategy is applied to denote importances attributes. We incorporate proposed iterative algorithm, extensive experiments show that algorithm outperforms OCIL K-Prototype methods 2.13% 4.28% improvements, respectively, terms accuracy on six from UCI.