Publication: Beer classification by electronic nose
1
Issued Date
2008-12-01
Resource Type
Other identifier(s)
2-s2.0-56749151792
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
Proceedings of the 2008 International Conference on Wavelet Analysis and Pattern Recognition, ICWAPR. Vol.1, (2008), 333-338
Suggested Citation
Chomtip Pornpanomchai, Natt Suthamsmai Beer classification by electronic nose. Proceedings of the 2008 International Conference on Wavelet Analysis and Pattern Recognition, ICWAPR. Vol.1, (2008), 333-338. doi:10.1109/ICWAPR.2008.4635799 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/19109
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Beer classification by electronic nose
Author(s)
Other Contributor(s)
Abstract
An electronic nose is a smart instrument that is designed to detect and discriminate among complex odors by using arrays of sensors. The arrays of sensors are treated with a variety of odor-sensitive biological or chemical materials. An electronic nose is a project that uses two researches areas which are hardware for developing sensors and software using theorem from neuron network technology. The operation begins when sensors hit the smell of beer. The result is converted from analog to digital and represented in a graph form. An artificial intelligence is a tool of a thinking system which can create knowledge as if a human does. This project concerns training and testing beer by using 10 types of beer which are Asahi, Chang, Cheer, Samiguel, Singha, Kloster, Heineken, Leo, Tiger and Tai. We separate the experiment into two parts. The first part is immediate checking, which is performed immediately after the beer can is opened. The second part is to check the beer after the can is opened for 24 hours. This project consists of two data classifications which are Rule base and Neural Network. Rule base is used to classify unknown data. Neural network is used to check types of beer. Our structure in a neural network consists of 25 input nodes, 28 hidden nodes, and 10 output nodes. The percentage of correctness is equal to 87.5%. ©2008 IEEE.