抽象的な

A Survey on Evolutionary Co-Clustering Formulations for Mining Time-Varying Data Using Sparsity Learning

R.Amsaveni, R. Suresh Kumar MCA, MPhil

The data matrix is considered as static in Traditional clustering and feature selection methods. However, the data matrices evolve smoothly over time in many applications. A simple approach to learn from these time-evolving data matrices is to analyze them separately. Such strategy ignores the time-dependent nature of the underlying data. Two formulations are proposed for evolutionary co-clustering and feature selection based on the fused Lasso regularization. The evolutionary co-clustering formulation is able to identify smoothly varying data embedded into the matrices along with the temporal dimension. Formulation allows for imposing smoothness constraints over temporal dimension of the data matrices. The evolutionary feature selection formulation can uncover shared features in clustering from time-evolving data matrices.