(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

A comparison between levenberg-marquardt (lm) intelligent system and bayesian regularization (br) intelligent system for flow regime classification / Mohamad Iqbal Sa'ad

A comparison between levenberg-marquardt (lm) intelligent system and bayesian regularization (br) intelligent system for flow regime classification_Mohamad Iqbal Sa'ad_E3_2006
Projek ini dilaksanakan untuk membuat perbandingan di antara Sistem Kecerdikan Buatan Levenberg-Marquardt (LM) dan Sistem Kecerdikan Buatan Bayesian Regularization (BR) dari segi prestasi, tempoh pembelajaran serta output yang dihasilkan oleh kedua-dua sistem kecerdikan buatan ini melalui masalahan pengklasifikasian. Perbandingan ini dibuat adalah untuk membantu di dalam pemilihan algoritma pembelajaran untuk penyelesaian masalah. Masalah pengklasifikasian dihasilkan menggunakan proses Tomografi Kemuatan Elektrik (TKE) yang menyediakan data bagi mengenalpasti rejim-rejim aliran minyak di dalam saluran penghantaran. TKE mengenalpasti kemuatan bendalir yang berbeza dan seterusnya menghasilkan data untuk proses pengklasifikasian ini. Perseptron Berbilang Lapisan (MLP) merupakan bentuk Rangkaian Neural Buatan (RNB) yang biasa digunakan untuk masalah pengklasifikasian dibina dengan menggunakan perisian MATLAB 7®. Perbandingan yang dilaksanakan akan menunjukkan bahawa algoritma pembelajaran LM adalah algoritma yang lebih pantas berbanding algoritma pembelajaran BR manakala algoritma pembelajaran BR mampu membina sebuah sistem kecerdikan buatan yang lebih bagus dari segi prestasi keseluruhan sistem. ______________________________________________________________________________________ The purpose of this project is to study the performance, leaning time and, output of Levenberg-Marquardt (LM) intelligent system and Bayesian Regularization (BR) intelligent system through a classification problem. These studies will help in choosing the right training algorithm for classification problem involved. These intelligent systems have to classify flow regimes in a closed line with the data are provided by Electrical Capacitance Tomography (ECT). ECT measured the different capacitance value of fluid and produced the data for the classification problem. Multilayed Perceptron (MLP), a type of artificial neural network (ANN) which is widely used in a classification problem is developed using MATLAB 7®. The comparison made showed that LM learning algortihm is a faster training algorithm compared to BR training algorithm meanwhile BR learning algorithm capable of building a superior intelligent system in term of the overall system performance.
Contributor(s):
Mohamad Iqbal Sa'ad - Author
Primary Item Type:
Final Year Project
Language:
English
Subject Keywords:
performance; leaning time; output of Levenberg-Marquardt (LM)
First presented to the public:
5/1/2006
Original Publication Date:
1/22/2019
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 45
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2019-01-22 10:56:47.91
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
A comparison between levenberg-marquardt (lm) intelligent system and bayesian regularization (br) intelligent system for flow regime classification / Mohamad Iqbal Sa'ad1 2019-01-22 10:56:47.91
Engineering Library USM @ 2017