(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

A study on detection of honey purity using computational intelligence method

A study on detection of honey purity using computational intelligence method / Teh Zhi Wern
Contoh madu dengan kepekatan madu Tualang tulen yang berbeza dianalisis dengan menggunakan hidung elektronik (E-Nose) dan Fourier Transform infra-merah Spektroskopi (FTIR). Data yang diekstrakkan daripada E-Nose dan FTIR akan dinilai dengan menggunakan Artificial Neural Network (ANN) untuk memberi keputusan klasifikasi dan penganggaran. ANN dapat mengkelaskan madu sebagai tulen atau dicemari. Selain itu, ANN juga berupaya untuk menganggarkan kepekatan madu tulen dalam sampel. Bahagian pertama dalam kerja ini adalah untuk mengelaskan madu sebagai tulen atau dicemari manakala bahagian kedua kerja ini adalah untuk menganggarkan kepekatan madu tulen. Sekiranya madu diklasifikasikan sebagai tulen, sistem ini akan tamat. Namun, sekiranya madu diklasifikasikan sebagai dicemari, sistem ini akan terus ke bahagian kedua untuk menganggarkan kepekatan madunya. ANN yang digunakan dalam kerja ini merupakan satu MLP. Bilangan neuron yang tersembunyi dan fungsi pengaktifan MLP tersebut adalah berbeza-beza untuk menentukan struktur MLP yang paling mudah. Keputusan yang baik diperolehi dengan menggunakan E-Nose dan FTIR dalam pengelasan madu sampel dengan menggunakan model rangkaian neural yang berdasarkan satu lapisan perceptron. Hasil kajian menunjukkan bahawa pada dasarnya tiada perbezaan dalam prestasi dengan menggunakan E-Nose ataupun FTIR data. Walau bagaimanapun, bagi MLP sebagai estimator, data E-Nose adalah lebih baik daripada data FTIR. Keputusan ini menunjukkan bahawa hidung elektronik boleh dijadikan alat yang berguna untuk pengelasan dan anggaran madu. _______________________________________________________________________________________________________ Samples of honey with different concentration of pure Tualang honey were analyzed using electronic nose (E-Nose) and Fourier Transform Infrared Spectroscopy (FTIR). The data extracted from E-Nose and FTIR were evaluated by Artificial Neural Network (ANN) to give classification and estimation results. The ANN is able to classify the honey as pure or adulterated. Besides that, the ANN is also capable of estimating the concentration of pure honey in a sample. The first part of this work was about classifying the honey as pure or adulterated while the second part of this work was about estimating the concentration of pure honey. If the honey is classified as pure, then the system ends. However, if the honey is classified as adulterated, the system proceeds to the second part i.e. to estimate the concentration of pure honey in the sample. The ANN used in this work was a MLP. The number of hidden neurons and activation functions of the MLP were varied to determine the simplest MLP structure. Good results were obtained in the classification of honey samples by using a neural network model based on a multilayer perceptron for both E-Nose and FTIR data. The results showed that there was essentially no difference between the network performances on the two data sets. However, for MLP as estimator, the E-Nose data performed better than FTIR data. This result suggests that the electronic nose could be a useful tool for the classification and estimation of honey.
Contributor(s):
Teh Zhi Wern - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875005067
Language:
English
Subject Keywords:
honey; (E-Nose); (FTIR)
First presented to the public:
6/1/2013
Original Publication Date:
11/8/2019
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 101
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2019-11-29 16:42:57.215
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
A study on detection of honey purity using computational intelligence method1 2019-11-29 16:42:57.215