Glosarium AI

Feature Cross

Technique creating new features by combining two or more existing features to capture non-linear relationships between variables.

Polynomial Feature Generation

Feature Interaction Detection

Algorithm automatically identifying significant interactions between features to generate relevant combined variables.

Feature Importance Ranking

Algorithm automatically ranking features based on their predictive contribution using metrics like Gini importance or permutation importance.

Automated Feature Extraction

Technique automatically transforming high-dimensional data into a lower-dimensional space while preserving relevant information.

Automated Feature Transformation

Automatic application of mathematical transformations (log, sqrt, box-cox) to features to improve their distribution and normality.

Automated Feature Scaling

Automatic normalization or standardization of features to put them on a common scale, essential for many ML algorithms.

Automated Feature Encoding

Automatic conversion of categorical variables into appropriate numerical representations like one-hot encoding, target encoding or embeddings.

Automated Feature Discretization

Process that automatically converts continuous variables into discrete intervals using methods like equal-width or equal-frequency binning.

Automated Feature Aggregation

Automatic generation of aggregated features (mean, sum, max) from data groups to capture statistical information.

Automated Text Feature Engineering

Automatic extraction of features from textual data including TF-IDF, n-grams, semantic embeddings, and linguistic metrics.

Feature Space Exploration

Systematic and automatic exploration of the possible feature space to identify optimal transformations.

Automated Feature Pruning

Automatic pruning of redundant or uninformative features to reduce model complexity and avoid overfitting.

Feature Creation via Deep Learning

Using deep neural networks to automatically learn hierarchical and abstract feature representations.

Genetic Feature Engineering

Application of genetic algorithms to evolve and automatically optimize feature sets over multiple generations.

Meta-Feature Engineering

Automatic generation of meta-features describing the statistical and structural properties of the original data.