AI-ordlista
Den kompletta ordlistan över AI
Categorical Feature Binarization
Process of converting categorical features into binary features optimized by CatBoost to improve model performance.
Prediction Time Transformation
Transformation applied to categorical features at prediction time to maintain consistency with training in CatBoost.
Minimal Variance Sampling
Sampling strategy in CatBoost that minimizes the variance of categorical feature estimates to improve model stability.
CatBoostClassifier
Specific implementation of CatBoost for classification problems, optimized with adapted loss functions like Logloss or MultiClass.
CatBoostRegressor
Version of CatBoost specialized in regression tasks, using loss functions like RMSE or MAE for optimization.
Greedy Categorical Splitting
Node splitting algorithm for categorical features in CatBoost that combines categories in a greedy manner to maximize gain.
Random Strength
CatBoost parameter controlling the degree of randomness in split selection, similar to regularization through stochasticity in gradient boosting.