Projection pursuit based on Gaussian mixtures and evolutionary algorithms
This is an incremental improvement for researchers in data visualization and dimensionality reduction, offering a semi-parametric method to enhance projection pursuit.
The authors tackled the problem of detecting informative structures in multivariate data by proposing a projection pursuit algorithm that uses Gaussian mixture models to estimate negentropy and genetic algorithms to optimize projections, showing it is flexible and effective on artificial and real datasets.
We propose a projection pursuit (PP) algorithm based on Gaussian mixture models (GMMs). The negentropy obtained from a multivariate density estimated by GMMs is adopted as the PP index to be maximised. For a fixed dimension of the projection subspace, the GMM-based density estimation is projected onto that subspace, where an approximation of the negentropy for Gaussian mixtures is computed. Then, Genetic Algorithms (GAs) are used to find the optimal, orthogonal projection basis by maximising the former approximation. We show that this semi-parametric approach to PP is flexible and allows highly informative structures to be detected, by projecting multivariate datasets onto a subspace, where the data can be feasibly visualised. The performance of the proposed approach is shown on both artificial and real datasets.