Glosarium AI
Kamus lengkap Kecerdasan Buatan
AutoML
End-to-end automation process of the machine learning lifecycle, including data preparation, feature engineering, model selection and hyperparameter optimization without significant human intervention.
Automated Feature Engineering
Set of algorithmic techniques that automatically generate new features from existing data, transform variables and select the most relevant characteristics to improve model performance.
Automated Feature Selection
Algorithmic process that automatically identifies and selects the most relevant subset of features by eliminating redundancies, correlations and uninformative characteristics to optimize model performance.
Auto-Sklearn
Open-source AutoML framework based on scikit-learn that automates algorithm selection, hyperparameter optimization and pipeline construction using Bayesian Optimization and meta-learning.
TPOT
Python AutoML tool that uses genetic algorithms to automatically optimize machine learning pipelines by exploring the space of preprocessing, feature selection and classification/regression algorithm combinations.
H2O AutoML
Machine learning automation platform that automatically trains and validates a wide range of models, including GLM, GBM, Random Forest and Deep Learning, with hyperparameter optimization and automatic stacking.
Bayesian Optimization
Sequential global optimization method that uses a probabilistic model to model the objective function and an acquisition criterion to decide which points to evaluate next, particularly effective for hyperparameter optimization.
Automated Model Selection
Process that automatically evaluates multiple machine learning algorithms on a given dataset to select the optimal model based on predefined performance metrics and computational constraints.
Feature Importance Automation
Set of algorithms that automatically calculate the relative importance of each feature in the model using techniques like SHAP, LIME, or permutation-based methods to interpret predictions.
Ensemble Learning Automation
Automatic process that strategically combines multiple base models to create a more robust predictive model, including automation of bagging, boosting, and stacking with optimization of combination weights.
Cross-Validation Automation
System that automates the configuration and execution of optimal cross-validation strategies based on data characteristics and model type to reliably evaluate generalized performance.
Pipeline Automation
Automated orchestration of all machine learning workflow steps, from data ingestion to model deployment, including preprocessing, training, and validation with dependency management.
AutoKeras
AutoML library for deep learning that automates neural network architecture search, hyperparameter optimization, and selection of best configurations for classification and regression tasks.
Hyperband
Hyperparameter optimization algorithm based on adaptive resource allocation that quickly eliminates poor configurations and allocates more resources to promising configurations for efficient exploration.
Population Based Training (PBT)
Optimization method that combines reinforcement learning with evolutionary algorithms, where a population of models trains in parallel with simultaneous exploration and exploitation of hyperparameters.
Automated Data Preprocessing
System that automatically detects and applies necessary transformations to raw data, including type detection, missing value handling, normalization, encoding, and anomaly detection.
Neural Architecture Transfer
AutoML technique that automatically transfers and adapts optimized neural network architectures from source tasks to new target tasks with similar characteristics to accelerate the NAS process.