hrvatski jezikClear Cookie - decide language by browser settings

Optimization of self-organizing polynomial neural networks

Marić, Ivan (2013) Optimization of self-organizing polynomial neural networks. Expert systems with applications, 40 (11). pp. 4528-4538. ISSN 0957-4174

| Request a personal copy from author
[img]
Preview
PDF - Accepted Version
Download (544kB) | Preview

Abstract

The main disadvantage of self-organizing polynomial neural networks (SOPNN) automatically structured and trained by the group method of data handling (GMDH) algorithm is a partial optimization of model weights as the GMDH algorithm optimizes only the weights of the topmost (output) node. In order to estimate to what extent the approximation accuracy of the obtained model can be improved the particle swarm optimization (PSO) has been used for the optimization of weights of all node-polynomials. Since the PSO is generally computationally expensive and time consuming a more efficient Levenberg–Marquardt (LM) algorithm is adapted for the optimization of the SOPNN. After it has been optimized by the LM algorithm the SOPNN outperformed the corresponding models based on artificial neural networks (ANN) and support vector method (SVM). The research is based on the meta-modeling of the thermodynamic effects in fluid flow measurements with time-constraints. The outstanding characteristics of the optimized SOPNN models are also demonstrated in learning the recurrence relations of multiple superimposed oscillations (MSO).

Item Type: Article
Uncontrolled Keywords: polynomial neural networks; GMDH; Levenberg–Marquardt algorithm; Particle swarm optimization; Time series modeling
Subjects: TECHNICAL SCIENCES > Electrical Engineering > Power Engineering
TECHNICAL SCIENCES > Electrical Engineering > Electromechanical Engineering
TECHNICAL SCIENCES > Electrical Engineering > Electronics
TECHNICAL SCIENCES > Electrical Engineering > Automation and Robotics
TECHNICAL SCIENCES > Computing > Data Processing
TECHNICAL SCIENCES > Computing > Artificial Intelligence
TECHNICAL SCIENCES > Computing > Process Computing
TECHNICAL SCIENCES > Computing > Program Engineering
TECHNICAL SCIENCES > Mechanical Engineering > Process Energy Engineering
Divisions: Division of Electronics
Projects:
Project titleProject leaderProject codeProject type
Computational Intelligence Methods in Measurement Systems (Postupci računalne inteligencije u mjernim sustavima)-Ivan Marić098-0982560-2565MZOS
Depositing User: Ivan Marić
Date Deposited: 27 Jan 2016 12:33
Last Modified: 27 Jan 2016 12:33
URI: http://fulir.irb.hr/id/eprint/2422
DOI: 10.1016/j.eswa.2013.01.060

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year