From: A new pipeline for structural characterization and classification of RNA-Seq microbiome data
Category | Description | Weaknesses | References |
---|---|---|---|
Filter | - Extract features from the data without any type of learning involved. | - Ignore interaction with the classifier. | |
Wrapper | - Use learning approaches to evaluate which features are useful. | - Risk of overfitting. - Classifier dependent selection. | |
Embedded | - Combine the traditional feature selection step with the classifier construction. | - Classifier dependent selection. |