Hyperparameter Search Spaces and Fixed Settings is a small dataset by Megumi Shiomi, last updated on 2026-05 04. It contains the hyperparameter search spaces and fixed settings used for tuning LightGBM, Random Forest, and XGBoost models. The tuning was performed using Bayesian optimization with Optuna, employing five-fold stratified cross-validation and 100 trials per algorithm.
Use Cases
- Replicate hyperparameter optimization experiments based on the described search spaces and settings.
- Benchmark model performance based on the described Bayesian optimization and cross-validation methodology.
- Initialize hyperparameter tuning workflows based on the documented fixed settings for the three algorithms.
Strengths
- Documented hyperparameter tuning methodology using Bayesian optimization with 100 trials per algorithm.
- Explicitly covers three major algorithms: LightGBM, Random Forest, and XGBoost.
- Uses a specific random seed (42) for reproducibility.
Limitations
- Row count is unknown, which may limit suitability assessment.
- Column-level documentation is absent; field semantics must be inferred after download.
- The dataset is very small (11.2 KB), indicating limited scope.
Provenance
- Source
- Megumi Shiomi via figshare
- Collection Method
- Generated from hyperparameter tuning experiments using Optuna.
- Time Range
- null
- Freshness
- Last updated 2026-05-04 17:40:07; freshness should be verified.
- Geography
- null