Module automl 

Automated Machine Learning (AutoML) module.

Provides hyperparameter optimization with type-safe parameter definitions, multiple search strategies, and integration with aprender’s Estimator trait.

Quick Start

use aprender::automl::{SearchSpace, RandomSearch, Trial};
use aprender::automl::params::RandomForestParam as RF;

// Define type-safe search space (Poka-Yoke: compile-time typo prevention)
let space = SearchSpace::new()
    .add(RF::NEstimators, 10..500)
    .add(RF::MaxDepth, 2..20);

// Random search with 50 trials
let search = RandomSearch::new(50);

Design Principles

• Type Safety: Parameter keys are enums, not strings—typos caught at compile time
• Zero Unsafe: Pure Rust implementation leveraging trueno SIMD
• Extensible: Custom parameter enums for any model family

References

• Bergstra & Bengio (2012). Random Search for Hyper-Parameter Optimization. JMLR.
• Snoek et al. (2012). Practical Bayesian Optimization. NeurIPS.

Re-exports

pub use params::GradientBoostingParam;

pub use params::RandomForestParam;

Modules

params

Type-safe hyperparameter definitions.

Structs

ActiveLearningSearch

Active Learning search optimizer.

AutoTuner

AutoTuner for hyperparameter optimization.

DESearch

Differential Evolution search optimizer.

EarlyStopping

Early stopping callback.

GridSearch

Grid search optimizer.

LogScale

Marker for log-scale continuous ranges.

ProgressCallback

Progress logging callback.

RandomSearch

Random search optimizer.

SearchSpace

Type-safe search space for hyperparameters.

TPE

Tree-structured Parzen Estimator optimizer.

TPEConfig

TPE optimizer configuration.

TimeBudget

Time budget constraint.

Trial

A hyperparameter configuration to evaluate.

TrialResult

Result of evaluating a trial.

TuneResult

Result of hyperparameter optimization.

Enums

HyperParam

Hyperparameter value that can be sampled.

ParamValue

A concrete parameter value.

Traits

Callback

Callback trait for monitoring optimization progress.

SearchStrategy

Search strategy trait for hyperparameter optimization.