Imbalanced Regression Dataset Repository

A curated collection of datasets for extreme value-aware regression tasks

This page provides access to 62 datasets with metadata on features, target imbalance, extreme values, and missing data characteristics. Ideal for benchmarking regression models under imbalanced conditions.

This repository has been constructed and used in the following work:
The data is available in two formats: CSV and ARFF.

Download CSVs Download ARFFs View on GitHub

Dataset	Description	Features	Nominal Features	Numeric Features	Instances	Missing Values	Type of Extreme	Relevance Threshold	# Rare	% Rare	Target Variable	Source
a1	The data points are taken on an annual basis from various streams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	Both	0.8	0	0.00%	a1	[1]
a2	The data points are taken on an annual basis from various streams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	Both	0.8	0	0.00%	a2	[1]
a3	The data points are taken on an annual basis from various streams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	Both	0.8	0	0.00%	a3	[1]
a4	The data points are taken on an annual basis from various streams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	Both	0.8	0	0.00%	a4	[1]
a6	The data points are taken on an annual basis from variousstreams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	Both	0.8	0	0.00%	a6	[1]
a7	The data points are taken on an annual basis from various streams / rivers in Europe, compiling features aimed at predicting the concentrations of seven algae species.	11	3	8	198	No	High	0.8	14	7.07%	a7	[1]
abalone	Predict the age of abalone from physical measurements.	8	1	7	4177	No	Both	0.8	1033	24.73%	Rings	[2]
acceleration	Dataset with acceleration statistics.	14	3	11	1732	No	Both	0.8	158	9.12%	acceleration	[3]
ailerons	The attributes describe the status of the aeroplane, while the goal is to predict the control action on the ailerons of the aircraft.	40	0	40	13515	No	Both	0.8	1622	11.80%	Goal	[4]
airfoil	NASA data set, obtained from a series of aerodynamic and acoustic tests of two and three-dimensional airfoil blade sections conducted in an anechoic wind tunnel.	5	0	5	1503	No	High	0.8	80	5.32%	scaled-sound-pressure	[5]
anacalt	The data contains information about the decisions taken by a supreme court.	7	0	7	4052	No	Both	0.8	0	0.00%	Log_exposure	[4]
appliances_energy	Experimental data used to create regression models of appliances energy use in a low energy building.	27	0	27	19735	No	Both	0.8	8212	41.61%	Appliances	[6]
autoprices	Dataset with feature leading to the prediction of its price.	16	1	15	159	No	Both	0.8	0	0.00%	class	[7]
availablePower	Dataset with power statistics.	15	7	8	1802	No	Both	0.8	305	16.93%	available.power	[8]
bank8FM	Part of a family of datasets synthetically generated from a simulation of how bank-customers choose their banks.	8	0	8	4499	No	Both	0.8	0	0.00%	rej	[9]
baseball	This dataset contains the 1992 salaries of the set of Major League Baseball players who played at least one game in both the 1991 and 1992 seasons, excluding pitchers.	16	0	16	337	No	Both	0.8	0	0.00%	Salary	[10]
californiaHousing	This data set contains information about all the block groups in California from the 1990 Census.	8	0	8	20640	No	Low	0.8	1802	8.73%	MedianHouseValue	[11]
cocomo	Software Engineering Repository data set made publicly available in order to encourage repeatable, verifiable, refutable, and/or improvable predictive models of software engineering.	16	1	1	60	No	Low	0.8	14	23.33%	ACT_EFFORT	[12]
concrete	Concrete Compressive Strength data set	8	0	8	1030	No	Low	0.8	0	0.00%	ConcreteCompressiveStrength	[13]
cpuActiv	Computer activity data set	21	0	21	8192	No	Low	0.8	371	4.53%	Usr	[9]
cpuSm	The Computer Activity databases are a collection of computer systems activity measures. The data was collected from a Sun Sparcstation 20/712 with 128 Mbytes of memory running in a multi-user university department. The final dataset is taken from both occasions with equal numbers of observations coming from each collection epoch.	12	0	12	8192	No	Low	0.8	371	4.53%	usr	[9]
debutanizer	This dataset aims to predict the butane concentration on a Debutanizer column.	7	0	7	2394	No	High	0.8	212	8.86%	y	[14]
deltaAirlerons	This data set is also obtained from the task of controlling the ailerons of a F16 aircraft.	5	0	5	7129	No	Both	0.8	1206	16.92%	Sa	[4]
deltaElevators	This data set is also obtained from the task of controlling the elevators of a F16 aircraft.	60	0	6	9517	No	High	0.8	4785	50.28%	Se	[4]
diabetes	This data set concerns the study of the factors affecting patterns of insulin-dependent diabetes mellitus in children.	2	0	2	43	No	High	0.8	6	13.95%	C_peptide	[4]
ele-1	Electrical Length data set	2	0	2	495	No	High	0.8	21	4.24%	Length	[4]
ele-2	Electrical-Maintenance data set	4	0	4	1056	No	Both	0.8	0	0.00%	Y	[4]
elevators	The attributes describe the status of the aeroplane, while the goal is to predict the control action on the ailerons of the aircraft.	18	0	18	16599	No	Both	0.8	4390	26.45%	Goal	[15]
forestFires	Forest Fires data set	12	0	12	517	No	High	0.8	15	2.90%	Area	[16]
friedman	Friedman Benchmark Function data set	5	0	5	1200	No	Both	0.8	0	0.00%	Output	[4]
fuelConsumption	The data contains information about car’s emissions and fuel consumption.	37	12	25	1764	No	Both	0.8	167	9.47%	fuel.counsumption.country	[17]
geographical_origin_music	Instances in this dataset contain audio features extracted from 1059 wave files. The task associated with the data is to predict the geographical origin of music.	117	0	117	1059	No	Both	0.8	104	9.82%	V100	[18]
heat	Dataset with heating statistics.	11	3	8	7400	No	Both	0.8	833	11.26%	heat	[8]
house16H	This database was designed on the basis of data provided by US Census Bureau.	16	0	16	22784	No	Both	0.8	6098	26.76%	Price	[9]
housing	The Ames Housing Dataset is a well-known dataset in the field of machine learning and data analysis. It contains various features and attributes of residential homes in Ames, Iowa, USA.	79	43	36	1460	Yes (7829)	Both	0.8	179	12.26%	SalePrice	[19]
housingBoston	This dataset contains information collected by the U.S Census Service concerning housing in the area of Boston Mass.	13	0	13	506	No	Both	0.8	105	20.75%	HousValue	[20]
kdd_coil_1	This data set is from the 1999 Computational Intelligence and Learning (COIL) competition. The data contains measurements of river chemical concentrations and algae densities.	11	3	8	316	Yes (56)	Both	0.8	0	0.00%	algae_1	[21]
kinematics8nm	This is data set is concerned with the forward kinematics of an 8 link robot arm.	8	0	8	8192	No	Both	0.8	0	0.00%	y	[9]
laser	Laser generated data set	4	0	4	993	No	Both	0.8	0	0.00%	Output	[4]
lungcancer-shedden	Prediction in Lung Adenocarcinoma	23	3	20	442	No	High	0.8	12	2.71%	OS_years	[22]
machineCPU	Machine CPU Performance data set	6	0	6	209	No	Both	0.8	46	22.01%	PRP	[4]
maxTorque	Dataset with torque statistics.	32	13	19	1802	No	Both	0.8	235	13.04%	maximal.torque	[23]
meta	Meta-Data was used in order to give advice about which classification method is appropriate for a particular dataset.	21	2	19	528	Yes (504)	Both	0.8	165	31.25%	class	[24]
mortgage	Mortgage data set	15	0	15	1049	No	Low	0.8	133	12.68%	30Y-CMortgageRate	[25]
pdgfr	This is one of 41 drug design datasets.	320	0	320	79	No	High	0.8	15	18.99%	oz322	[26]
pollen	This dataset is synthetic. It was generated by David Coleman at RCA Laboratories in Princeton, N.J.	5	0	5	3848	No	Both	0.8	242	6.29%	DENSITY	[27]
puma32NH	This is a family of datasets synthetically generated from a realistic simulation of the dynamics of a Unimation Puma 560 robot arm.	32	0	32	8192	Yes (33)	Both	0.8	0	0.00%	thetadd6	[9]
puma8NH	This is a family of datasets synthetically generated from a realistic simulation of the dynamics of a Unimation Puma 560 robot arm.	8	0	8	8192	Yes (9)	Both	0.8	0	0.00%	thetadd3	[9]
quake	Quake data set	3	0	3	2178	No	Both	0.8	0	0.00%	Richter	[4]
sensory	Data for the sensory evaluation experiment in Brien, C.J. and Payne, R.W. (1996) Tiers, structure formulae and the analysis of complicated experiments.	11	0	11	576	No	Both	0.8	69	11.98%	Score	[28]
servo	This is an interesting collection of data provided by Karl Ulrich. It covers an extremely non-linear phenomenon - predicting the rise time of a servomechanism in terms of two (continuous) gain settings and two (discrete) choices of mechanical linkages.	4	4	0	167	No	Both	0.8	59	35.33%	class	[15]
space_ga	The dataset contains 3,107 observations on U.S. county votes cast in the 1980 presidential election.	6	0	6	3107	No	Both	0.8	182	5.86%	ln_votes_pop	[29]
stock	Stock Prices data set	9	0	9	950	No	Both	0.8	0	0.00%	Company10	[30]
strikes	The data consist of annual observations on the level of strike volume (days lost due to industrial disputes per 1000 wage salary earners), and their covariates in 18 OECD countries from 1951-1985.	6	0	6	625	No	High	0.8	15	2.40%	strike_volume	[31]
sulfer_1	The sulfur recovery unit (SRU) removes environmental pollutants from acid gas streams before they are released into the atmosphere. Furthermore, elemental sulfur is recovered as a valuable by-product.	5	0	5	10081	No	Both	0.8	1117	11.08%	y1	[32]
sulfer_2	The sulfur recovery unit (SRU) removes environmental pollutants from acid gas streams before they are released into the atmosphere. Furthermore, 0.8elemental sulfur is recovered as a valuable by-product.	5	0	5	10081	No	Both	0.8	1444	14.32%	y2	[32]
supercondutivity	Two files contain data on 21263 superconductors and their relevant features.	81	0	81	21263	No	Both	0.8	0	0.00%	critical_temp	[33]
treasury	This file contains the Economic data information of USA from 01/04/1980 to 02/04/2000 on a weekly basis.	15	0	15	1049	No	Low	0.8	137	13.06%	1MonthCDRate	[4]
triazines	A triazine dataset. The goal is to predict the inhibition of dihydrofolate reductase by triazines.	60	0	60	186	No	Both	0.8	23	12.37%	activity	[34]
wankara	This file contains the weather information of Ankara from 01/01/1994 to 28/05/1998.	9	0	9	1609	No	Both	0.8	0	0.00%	Mean_temperature	[4]
wine-quality	The two datasets are combined and related to red and white variants of the Portuguese "Vinho Verde" wine.	12	1	11	6497	No	High	0.8	4113	63.31%	quality	[35]
yachtHydrodynamics	Delft data set, used to predict the hydodynamic performance of sailing yachts from dimensions and velocity.	6	0	6	308	No	Both	0.8	0	0.00%	Residuary_Resistance	[36]

References

[1] Torgo, L. (2016). Data mining with R: Learning with case studies (2nd ed.). Chapman & Hall/CRC. http://ltorgo.github.io/DMwR2
[2] Nash, W., Sellers, T., Talbot, S., Cawthorn, A., & Ford, W. (1994). Abalone. UCI Machine Learning Repository. https://doi.org/10.24432/C55C7W
[3] Moniz, N., Ribeiro, R. P., & Margarido, M. (2023). accel: Acceleration dataset [Dataset in the IRon R package, version 0.1.4]. https://CRAN.R-project.org/package=IRon
[4] Alcalá-Fdez, J., Fernández, A., Luengo, J., Derrac, J., García, S., Sánchez, L., & Herrera, F. (2011). KEEL data-mining software tool: Data set repository, integration of algorithms and experimental analysis framework. Journal of Multiple-Valued Logic and Soft Computing, 17(2–3), 255–287.
[5] Brooks, T., Pope, D., & Marcolini, M. (1989). Airfoil self-noise [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5VW2C
[6] Candanedo, L. (2017). Appliances energy prediction [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5VC8G
[7] Schlimmer, J. (1985). Automobile [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5B01C
[8] Camacho, L., & Bação, F. (2024). WSMOTER: A novel approach for imbalanced regression. Applied Intelligence, 54, 1–11. https://doi.org/10.1007/s10489-024-05608-6
[9] Rasmussen, C. E., Neal, R. M., Hinton, G. E., et al. (1996). DELVE: Data for evaluating learning in valid experiments[Software and dataset repository]. University of Toronto. https://www.cs.toronto.edu/~delve/
[10] Journal of Statistics Education. (1992). Pay for play: Are baseball salaries based on performance? [Dataset]. Dataset available from the Journal of Statistics Education. https://jse.amstat.org/datasets/baseball.dat.txt
[11] Carnegie Mellon University. (2016). StatLib: A data and software archive [Online dataset repository]. https://lib.stat.cmu.edu
[12] OpenML contributors. (2025). OpenML dataset 1051 [Dataset]. OpenML: An open platform for machine learning. https://www.openml.org/d/1051
[13] Yeh, I.-C. (1998). Concrete compressive strength [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5PK67
[14] Fortuna, L., Graziani, S., Rizzo, A., & Xibilia, M. G. (2007). Soft sensors for monitoring and control of industrial processes. In Advances in Industrial Control. Springer London. https://doi.org/10.1007/978-1-84628-480-9
[15] Torgo, L. (2019). Regression data sets [Online dataset repository]. LIACC / University of Porto. https://www.dcc.fc.up.pt/~ltorgo/Regression/DataSets.html
[16] Cortez, P., & Morais, A. (2007). Forest fires [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5D88D
[17] University of Toronto, DELVE Project. (2003). DELVE: Data for evaluating learning in valid experiments [Online dataset repository]. https://www.cs.toronto.edu/~delve/
[18] Zhou, F., Claire, Q., & King, R. D. (2014). Geographical origin of music [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5VK5D
[19] Necrothapa, S. (2020). Ames housing dataset [Dataset]. Kaggle. https://www.kaggle.com/datasets/shashanknecrothapa/ames-housing-dataset
[20] Schirmer, C. (2020). Boston housing [Dataset]. Kaggle. https://www.kaggle.com/datasets/schirmerchad/bostonhoustingmlnd
[21] Elkan, C. (2001). Magical thinking in data mining: Lessons from CoIL Challenge 2000. In Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 426–431). ACM. https://doi.org/10.1145/502512.502576
[22] Director's Challenge Consortium for the Molecular Classification of Lung Adenocarcinoma, Shedden, K., Taylor, J. M., Enkemann, S. A., Tsao, M. S., Yeatman, T. J., Gerald, W. L., Eschrich, S., Jurisica, I., Giordano, T. J., Misek, D. E., Chang, A. C., Zhu, C. Q., Strumpf, D., Hanash, S., & Shepherd, F. A. (2008). Shedden_2008: Gene expression–based survival prediction in lung adenocarcinoma [Dataset]. Lung Cancer Explorer, UT Southwestern. https://lce.biohpc.swmed.edu/lungcancer/datasetsearch.php?datasetid=1
[23] Branco, P., Torgo, L., & Ribeiro, R. P. (2025). Imbalanced regression data sets [Online dataset repository]. University of Porto. https://paobranco.github.io/Imbalanced-Regression-DataSets/
[24] Meta-data. (1994). Meta-data [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5X31P
[25] Board of Governors of the Federal Reserve System. (2025). H.15 selected interest rates [Statistical release and dataset repository]. https://www.federalreserve.gov/releases/h15/
[26] OpenML contributors. (2025). OpenML dataset 409 [Dataset]. OpenML platform. https://www.openml.org/d/409
[27] Coleman, D. (1986). pollen: Geometric features of pollen grains [Dataset]. StatLib Archive, Carnegie Mellon University. https://lib.stat.cmu.edu/data-expo/pollen.data
[28] Brien, C. J., & Payne, R. W. (1999). Tiers, structure formulae and the analysis of complicated experiments. Journal of the Royal Statistical Society: Series D (The Statistician), 48(1), 41–52.
[29] OpenML contributors. (n.d.). space_ga [Dataset]. OpenML. https://www.openml.org/d/507
[30] Carnegie Mellon University, Department of Statistics. (2016). StatLib: A data and software archive [Online dataset repository]. https://lib.stat.cmu.edu/datasets/
[31] Tibshirani, R. J. (2015). strike: Annual strikes data for OECD countries (1951–1985) [Dataset]. Dataset used in course “Statistical Computing,” Carnegie Mellon University. http://www.stat.cmu.edu/~ryantibs/statcomp-F15/homework/strike.txt
[32] OpenML contributors. (2025). sulfur [Dataset]. OpenML. https://www.openml.org/d/23515
[33] Hamidieh, K. (2018). Superconductivity data [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C53P47
[34] King, R. D., Hurst, J. D., & Sternberg, M. J. E. (1994). A comparison of artificial intelligence methods for modelling QSARs. Applied Artificial Intelligence, 9, 213–234. / Hirst, J. D., King, R. D., & Sternberg, M. J. E. (1994). Quantitative structure–activity relationships by neural networks and inductive logic programming: II. The inhibition of dihydrofolate reductase by triazines. Journal of Computer-Aided Molecular Design, 8(4), 421–432. https://doi.org/10.1007/BF00125376
[35] Cortez, P., Cerdeira, A., Almeida, F., Matos, T., & Reis, J. (2009). Wine Quality [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C56S3T.
[36] Gerritsma, J., Onnink, R., & Versluis, A. (1981). Yacht Hydrodynamics [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5XG7R.