IEEE Systems, Man and Cybernetics Magazine - April 2023 - 47

terms of different input-output relations. Through the
release of synthetic datasets, we aim encouraging research
groups within the broader ML and DL community to investigate
and quantify the uncertainty of their methods using
these synthetic datasets and thereafter apply their models
to various real datasets.
We have sample python scripts showing how to use
data in Kaggle. Also, we save the data in both " pkl " and
" csv " formats. The user of dataset can load " csv " files in
other programming languages while training models in
another language.
Code Availability
The following Kaggle notebook generates datasets along
with detailed relationships between inputs and the output:
https://www.kaggle.com/dipuk0506/toy-dataset
-for-regression-and-uq. The following Kaggle notebook
presents example shallow NN training on datasets:
https://www.kaggle.com/dipuk0506/shallow-nn-on-toydatasets.
Version-N of the notebook applies a shallow NN
to Data-N.
The following Kaggle notebook presents example RVFL
network training on datasets: https://www.kaggle.com/
dipuk0506/rvfl-on-synthetic-dataset. Version-N of the notebook
applies an RVFL network to Data-N. We also upload
datasets and example scripts at the following GitHub
repository: https://github.com/dipuk0506/UQ-Data.
Conclusion
This article has proposed ten uncertainty and regression-related
datasets with details of their construction
and example performance evaluations. Presented
datasets and methodologies may help future researchers
develop novel ML and DL algorithms. Regression-type
models are useful for predicting electricity demand, generations,
renewables, economics, traffic, cloud quantities,
and so on. This article has also presented examples
of model training with publicly available codes. We also
provided scripts for data generation, plotting, and the
training of models on the datasets to assist future users
of the dataset.
Acknowledgments
This work was supported by the Australian Research
Council through Discovery Projects funding scheme
(DP190102181).
About the Authors
H M Dipu Kabir (hussain.kabir@deakin.edu.au) earned
his Ph.D. degree in uncertainty quantification from Deakin
University. He is with Deakin University, Geelong, Vic 3216,
Australia. His research interests include uncertainty quantification
and artificial intelligence.
Moloud Abdar (mabdar@deakin.edu.au) earned his
Ph.D. degree in machine learning from Deakin University.
He is with Deakin University, Geelong, Vic 3216, Australia.
His research interests include data mining, machine learning,
deep learning, computer vision, and sentiment analysis.
Abbas Khosravi (abbas.khosravi@deakin.edu.au)
earned his Ph.D. degree in uncertainty quantification from
Deakin University. He is with Deakin University, Geelong, Vic
3216, Australia. His research interests include uncertainty
quantification, deep learning, probabilistic forecasting, artificial
intelligence (AI) for medicine, and AI for healthcare.
Darius Nahavandi (darius.nahavandi@deakin.edu.au)
earned his Ph.D. degree in human factors modeling and
simulation from Deakin University. He is with Deakin University,
Geelong, Vic 3216, Australia, and also with Harvard
University, Allston, MA 02134 USA. His research interests
include human factors. He is a Member of IEEE.
Subrota Kumar Mondal (skmondal@must.edu.mo)
earned his Ph.D. degree in cloud reliability from The Hong
Kong University of Science and Technology. He is with
Macau University of Science and Technology, Taipa, Macau.
His research interests include cloud/serverless computing,
cybersecurity, deep learning, NLP, and smart city. He is a
Member of IEEE.
Sadia Khanam (epnsugan@ntu.edu.sg) earned his BDS
degree in dental surgery from Dhaka Dental College. He is
with Dhaka Dental College, Dhaka, Bangladesh. His research
interests include artificial intelligence for healthcare.
Shady Mohamed (shady.mohamed@deakin.edu.au)
earned his Ph.D. degree in data fusion from Deakin University.
He is with Deakin University, Geelong, Vic 3216,
Australia. His research interests include motion simulators,
signal processing, and control theory. He is a Member
of IEEE.
Dipti Srinivasan (dipti@nus.edu.sg) earned his Ph.D.
degree in artificial intelligence techniques in power from the
National University of Singapore. He is with the National University
of Singapore, 11920 Singapore. His research interests
include smart grid, computational intelligence, microgrids,
renewable energy forecasting, and energy systems.
Saeid Nahavandi (saeid.nahavandi@deakin.edu.au)
earned his Ph.D. degree in control systems from Durham
University. He is with Deakin University, Geelong, Vic
3216, Australia, and also with Harvard University, Allston,
MA 02134 USA. His research interests include intelligent
systems, intelligent control, haptics, system of systems,
and modeling and simulation. He is a Fellow of IEEE.
Ponnuthurai Nagaratnam Suganthan (epnsugan@
ntu.edu.sg) earned his Ph.D. degree in object recognition
from Nanyang Technological University. He is with Nanyang
Technological University, Singapore 639798, and also with
Qatar University, Doha 2713, Qatar. His research interests
include machine learning (ML), swarm-evolutionary computation
(SEC), neural networks (ANN), and applications of
ML SEC ANN. He is a Fellow of IEEE.
References
[1] A. Coates, A. Ng, and H. Lee, " An analysis of single-layer networks in unsupervised
feature learning, " in Proc. 14th Int. Conf. Artif. Intell. Statist., 2011, pp. 215-223.
April 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE 47
https://www.kaggle.com/dipuk0506/shallow-nn-on-toy-datasets https://www.kaggle.com/dipuk0506/shallow-nn-on-toy-datasets https://www.kaggle.com/dipuk0506/rvfl-on-synthetic-dataset https://www.kaggle.com/dipuk0506/rvfl-on-synthetic-dataset https://www.github.com/dipuk0506/UQ-Data
IEEE Systems, Man and Cybernetics Magazine - April 2023

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - April 2023
