Feature Construction Methods : A Survey

62 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

AutoML: A Survey of the State-of-the-Art (2019, methodology): “In terms of searching, the algorithms like decision tree-based methods [79], [80] and genetic algorithm [81] require a pre-defined operation space, while annotation-based approaches eliminate that requirement, because itcan use domain knowledge in the form of annotation along with the training examp”
Benchmark and Survey of Automated Machine Learning Frameworks (2019, methodology): “New features are generated using a predefined set of operators transforming the original features (Sondhi, 2009): Unary Unary operators transform a single feature, for example by discretizing or normalizing numerical features, applying rule-based expansions of dates or using unary mathematical…”
SISSO: A compressed-sensing method for identifying the best low-dimensional descriptor in an immensity of offered candidates (2017, background): “All quantities that are hypothesized to be relevant for describing the target property (the so called primary features [7, 31]) are used as starting point for the construction of the space [37, 38].”
AutoLearn — Automated Feature Generation and Selection (2017, background): “The goal is to learn a predictor h : X → Y from T , with low error Errh.”