1. Musk (Version 1): The goal is to learn to predict whether new molecules will be musks or non-musks
2. Robot Execution Failures: This dataset contains force and torque measurements on a robot after failure detection. Each failure is characterized by 15 force/torque samples collected at regular time intervals
3. Ionosphere: Classification of radar returns from the ionosphere
4. Low Resolution Spectrometer: From IRAS data -- NASA Ames Research Center
5. Water Treatment Plant: Multiple classes predict plant state
6. Wine: Using chemical analysis determine the origin of wines
7. Function Finding: Cases collected mostly from investigations in physical science; intention is to evaluate function-finding algorithms
8. Glass Identification: From USA Forensic Science Service; 6 types of glass; defined in terms of their oxide content (i.e. Na, Fe, K, etc)
9. Intelligent Media Accelerometer and Gyroscope (IM-AccGyro) Dataset: The IM-AccGyro dataset is devised to benchmark techniques dealing with human activity recognition based on inertial sensors.
10. Connectionist Bench (Sonar, Mines vs. Rocks): The task is to train a network to discriminate between sonar signals bounced off a metal cylinder and those bounced off a roughly cylindrical rock.
11. Forest Fires: This is a difficult regression task, where the aim is to predict the burned area of forest fires, in the northeast region of Portugal, by using meteorological and other data (see details at: http://www.dsi.uminho.pt/~pcortez/forestfires).
12. Yacht Hydrodynamics: Delft data set, used to predict the hydodynamic performance of sailing yachts from dimensions and velocity.
13. QSAR fish toxicity: Data set containing values for 6 attributes (molecular descriptors) of 908 chemicals used to predict quantitative acute aquatic toxicity towards the fish Pimephales promelas (fathead minnow).
14. QSAR aquatic toxicity: Data set containing values for 8 attributes (molecular descriptors) of 546 chemicals used to predict quantitative acute aquatic toxicity towards Daphnia Magna..
15. Climate Model Simulation Crashes: Given Latin hypercube samples of 18 climate model input parameter values, predict climate model simulation crashes and determine the parameter value combinations that cause the failures.