No packages match

LogisticEnsembles - Automatically Runs 18 Logistic Models-14 Individual Logistic Models and 4 Ensembles of Models

Automatically returns results from 18 logistic models including 14 individual logistic models and 4 logistic ensembles of models. The package also returns 25 plots, 5 tables, and a summary report. The package automatically builds all 18 models, reports all results, and provides graphics to show how the models performed. This can be used for a wide range of data, such as sports or medical data. The package includes medical data (the Pima Indians data set), and information about the performance of Lebron James. The package can be used to analyze many other examples, such as stock market data. The package automatically returns many values for each model, such as True Positive Rate, True Negative Rate, False Positive Rate, False Negative Rate, Positive Predictive Value, Negative Predictive Value, F1 Score, Area Under the Curve. The package also returns 36 Receiver Operating Characteristic (ROC) curves for each of the 18 models.

Last updated

6.12 score 3 stars 11 scripts 378 downloads

NumericEnsembles - Multi-Model Stacking Optimization and Continuous Regression Ensembles

Automated parallelized regression ensemble pipeline executing concurrent evaluations across 17 base machine learning architectures, 6 specialized stacking meta-learners, and 136 pairwise hybrid combinations. The package implements dynamic performance-weighted stacking blends alongside internal data dictionary profiling metrics. It provides iterative multicollinearity pruning via Variance Inflation Factors (VIF) and automated executive Quarto report generation.

Last updated

quarto

4.68 score 2 stars 21 scripts 550 downloads

ForecastingEnsembles - Time Series Forecasting Using 23 Individual Models

Runs multiple individual time series models, and combines them into an ensembles of time series models. This is mainly used to predict the results of the monthly labor market report from the United States Bureau of Labor Statistics for virtually any part of the economy reported by the Bureau of Labor Statistics, but it can be easily modified to work with other types of time series data. For example, the package was used to predict the winning men's and women's time for the 2024 London Marathon.

Last updated

3.70 score 329 downloads

ClassificationEnsembles - Automated Multi-Class Stacking Ensembles and Predictive Diagnostics

Automated multi-class classification ensemble pipeline executing concurrent parameter optimizations across logistic, tree, kernel, and neural configurations. Implements dynamic out-of-sample blending, automated feature reduction via iterative Variance Inflation Factors (VIF), and native S3 diagnostic dashboards.

Last updated

3.60 score 2 stars 5 scripts 283 downloads