Scores for hazelnut-spread-contaminant-detection are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for polish_companies_bankruptcy are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for anneal are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for APSFailure are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for airfoil_self_noise are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for churn are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for physiochemical_protein are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for NATICUSdroid are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Marketing_Campaign are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for SDSS17 are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for maternal_health_risk are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for taiwanese_bankruptcy_prediction are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for customer_satisfaction_in_airline are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for houses are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for website_phishing are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for qsar-biodeg are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Bioresponse are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for coil2000_insurance_policies are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for concrete_compressive_strength are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for splice are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for seismic-bumps are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for jm1 are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for superconductivity are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for QSAR-TID-11 are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for online_shoppers_intention are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for diamonds are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for miami_housing are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for MIC are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Bank_Customer_Churn are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Is-this-a-good-customer are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for credit-g are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Another-Dataset-on-used-Fiat-500 are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for credit_card_clients_default are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for healthcare_insurance_expenses are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for students_dropout_and_academic_success are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Diabetes130US are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for wine_quality are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for heloc are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for QSAR_fish_toxicity are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, linear regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for E-CommereShippingData are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for HR_Analytics_Job_Change_of_Data_Scientists are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for GiveMeSomeCredit are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for Fitness_Club are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for diabetes are averaged across 30 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
Scores for bank-marketing are averaged across 9 evaluation splits. Causilo uses its default configuration; CatBoost, LightGBM, XGBoost, Random forest, Extra trees, logistic regression, and k-nearest neighbors are tuned and ensembled. Scores retain their original ROC AUC, log loss, or RMSE values. Dots are positioned on a linear scale within each dataset, from the worst result on the left to the best on the right; higher ROC AUC and lower log loss or RMSE are better. Positions show relative performance within a dataset, not score ratios.
