Modelling of Engineering Systems with Small Data: A comparative study

Morteza Mohammadzaheri; Mojtaba Ghodsi; Hamidreza Ziaiefar; Issam Bahadur; Musaab Zarog; Mohammadreza Emadi; Payam Soltani; Amirhosein Amouzadeh

doi:10.4018/978-1-6684-7684-0.ch006

Modelling of Engineering Systems with Small Data: A comparative study

Morteza Mohammadzaheri^*, Mojtaba Ghodsi, Hamidreza Ziaiefar, Issam Bahadur, Musaab Zarog, Mohammadreza Emadi, Payam Soltani, Amirhosein Amouzadeh

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

2 Citations (SciVal)

Abstract

This chapter equitably compares five different Artificial Intelligence (AI) techniques for data-driven modelling. All these techniques were used to solve two real-world engineering data-driven modelling problems with small number of experimental data samples, one with sparse and one with dense data. The models of both problems are shown to be highly nonlinear. In the problem with available dense data, Multi-Layer Perceptron (MLP) evidently outperforms other AI models and challenges the claims in the literature about superiority of Fully Connected Cascade (FCC). However, the results of the problem with sparse data shows superiority of FCC, closely followed by MLP and neuro-fuzzy network.

Original language	English
Title of host publication	Perspectives and Considerations on the Evolution of Smart Systems
Editors	Maki Habib
Publisher	IGI Global
Pages	120-136
Number of pages	17
ISBN (Electronic)	9781668476864
ISBN (Print)	9781668476840
DOIs	https://doi.org/10.4018/978-1-6684-7684-0.ch006
Publication status	Published (VoR) - 1 Jul 2023

Keywords

Modelling
Artificial Intelligence
Small Data
Sparse Data
Dense Data
Piezoelectric Actuator
Electrical Submersible Pump

Access to Document

10.4018/978-1-6684-7684-0.ch006

https://www.igi-global.com/book/perspectives-considerations-evolution-smart-systems/309144

Cite this

@inbook{78d3f77e5ee2458f9dae8749777cda40,

title = "Modelling of Engineering Systems with Small Data: A comparative study",

abstract = "This chapter equitably compares five different Artificial Intelligence (AI) techniques for data-driven modelling. All these techniques were used to solve two real-world engineering data-driven modelling problems with small number of experimental data samples, one with sparse and one with dense data. The models of both problems are shown to be highly nonlinear. In the problem with available dense data, Multi-Layer Perceptron (MLP) evidently outperforms other AI models and challenges the claims in the literature about superiority of Fully Connected Cascade (FCC). However, the results of the problem with sparse data shows superiority of FCC, closely followed by MLP and neuro-fuzzy network.",

keywords = "Modelling, Artificial Intelligence, Small Data, Sparse Data, Dense Data, Piezoelectric Actuator, Electrical Submersible Pump",

author = "Morteza Mohammadzaheri and Mojtaba Ghodsi and Hamidreza Ziaiefar and Issam Bahadur and Musaab Zarog and Mohammadreza Emadi and Payam Soltani and Amirhosein Amouzadeh",

year = "2023",

month = jul,

day = "1",

doi = "10.4018/978-1-6684-7684-0.ch006",

language = "English",

isbn = "9781668476840",

pages = "120--136",

editor = "Maki Habib",

booktitle = "Perspectives and Considerations on the Evolution of Smart Systems",

publisher = "IGI Global",

}

TY - CHAP

T1 - Modelling of Engineering Systems with Small Data

T2 - A comparative study

AU - Mohammadzaheri, Morteza

AU - Ghodsi, Mojtaba

AU - Ziaiefar, Hamidreza

AU - Bahadur, Issam

AU - Zarog, Musaab

AU - Emadi, Mohammadreza

AU - Soltani, Payam

AU - Amouzadeh, Amirhosein

PY - 2023/7/1

Y1 - 2023/7/1

N2 - This chapter equitably compares five different Artificial Intelligence (AI) techniques for data-driven modelling. All these techniques were used to solve two real-world engineering data-driven modelling problems with small number of experimental data samples, one with sparse and one with dense data. The models of both problems are shown to be highly nonlinear. In the problem with available dense data, Multi-Layer Perceptron (MLP) evidently outperforms other AI models and challenges the claims in the literature about superiority of Fully Connected Cascade (FCC). However, the results of the problem with sparse data shows superiority of FCC, closely followed by MLP and neuro-fuzzy network.

AB - This chapter equitably compares five different Artificial Intelligence (AI) techniques for data-driven modelling. All these techniques were used to solve two real-world engineering data-driven modelling problems with small number of experimental data samples, one with sparse and one with dense data. The models of both problems are shown to be highly nonlinear. In the problem with available dense data, Multi-Layer Perceptron (MLP) evidently outperforms other AI models and challenges the claims in the literature about superiority of Fully Connected Cascade (FCC). However, the results of the problem with sparse data shows superiority of FCC, closely followed by MLP and neuro-fuzzy network.

KW - Modelling

KW - Artificial Intelligence

KW - Small Data

KW - Sparse Data

KW - Dense Data

KW - Piezoelectric Actuator

KW - Electrical Submersible Pump

UR - http://www.open-access.bcu.ac.uk/14624/

U2 - 10.4018/978-1-6684-7684-0.ch006

DO - 10.4018/978-1-6684-7684-0.ch006

M3 - Chapter

SN - 9781668476840

SP - 120

EP - 136

BT - Perspectives and Considerations on the Evolution of Smart Systems

A2 - Habib, Maki

PB - IGI Global

ER -

Modelling of Engineering Systems with Small Data: A comparative study

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this