Interblend fusing of genetic algorithm-based attribute selection for clustering heterogeneous data set

Different clustering strategies to partition heterogeneous data set with numeric, binary, categorical and ordinal attributes are explored by the researchers. All the real-life applications data set is often heterogeneous in nature; if it is converted to homogeneous, then it leads to information loss. In this paper, we propose an interblend fusing of genetic algorithm-based attribute selection and increase the clustering accuracy in credit risk assessment. The proposed technique classifies the similar objects together without changing the characteristics of heterogeneous data sets. This algorithm also identifies the importance of attributes in clustering large number of objects with good many attributes. The fusing technique yields contextual distance measure for clustering the objects. The result presented in this paper provides clear interpretation of applying our methodology to the data sets. The performance of this algorithm is of the higher standard when compared to the related literature.