Feature Selection for Latent Factor Models
This addresses the challenge of identifying relevant features for classification tasks, particularly in high-dimensional settings, though it appears incremental as it builds on traditional feature selection approaches.
The paper tackles the problem of feature selection in high-dimensional datasets by proposing a method that selects features for each class separately using low-rank generative models and a signal-to-noise ratio criterion, resulting in theoretical guarantees and improved performance over some existing methods on standard classification datasets.
Feature selection is crucial for pinpointing relevant features in high-dimensional datasets, mitigating the 'curse of dimensionality,' and enhancing machine learning performance. Traditional feature selection methods for classification use data from all classes to select features for each class. This paper explores feature selection methods that select features for each class separately, using class models based on low-rank generative methods and introducing a signal-to-noise ratio (SNR) feature selection criterion. This novel approach has theoretical true feature recovery guarantees under certain assumptions and is shown to outperform some existing feature selection methods on standard classification datasets.