Machine Learning
274 views
Dropout
Quick Definition
Regularization randomly deactivating neurons during training
Full Definition
A regularization technique randomly deactivating neurons during training to prevent co-adaptation and overfitting.
Examples
neural network regularization, ensemble effect, overfitting prevention
Related Terms
regularization
overfitting
neural-network
More Machine Learning Terms
DBSCAN
Density-based clustering finding clusters of varying shapes
Active Learning
Algorithm selecting most informative data points for labeling
Dimensionality Reduction
Reducing feature count while preserving important information
Hidden Markov Model
Statistical model with hidden states and observable outputs
CatBoost
Gradient boosting library with native categorical feature support
Bayesian Optimization
Optimizing expensive functions using a probabilistic model