Sparse Semi-supervised Heterogeneous Interbattery Bayesian Analysis

• DOI: 10.48550/arxiv.2001.08975
• Publication Date: 2020-01-01
• Authors: Carlos Sevilla-Sa...; Vanessa Gómez-Ver...; Pablo M. Olmos
• Keywords: Interpretability; Representation; Feature (linguistics)

The Bayesian approach to feature extraction, known as factor analysis (FA), has been widely studied in machine learning obtain a latent representation of the data. An adequate selection probabilities and priors these bayesian models allows model better adapt data nature (i.e. heterogeneity, sparsity), obtaining more representative space. objective this article is propose general FA framework capable modelling any problem. To do so, we start from Inter-Battery Factor Analysis (BIBFA) model, enhancing it with new functionalities be able work heterogeneous data, include selection, handle missing values well semi-supervised problems. performance proposed Sparse Semi-supervised Heterogeneous Interbattery (SSHIBA) tested on 4 different scenarios evaluate each one its novelties, showing not only great versatility an interpretability gain, but also outperforming most state-of-the-art algorithms.