1. UKM Learning and Research Repository
2. Master Thesis / Tesis Master
3. Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Title:	Prediction of carbon monoxide concentrations based on a hybrid multilayer perceptron & imperialist competitive algorithm
Authors:	Somaiyeh Mahmoudzadeh (P58882)
Supervisor:	Zalinda Othman, Assoc. Prof. Dr.
Keywords:	Intelligent agents (Computer software) Neural networks (Computer science) Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia
Issue Date:	12-Oct-2012
Description:	Exposures to contaminated air in polluted environment may affect human health. The polluted environment may be contaminated by air pollutant concentrations such as carbon monoxide (CO), ozone and benzene. CO is a main air pollutant produced by incomplete combustion process and in urban areas, the main source of it is traffic. There are number of approaches in air pollution modeling such as traditional statistical approaches and artificial intelligence approaches. The literature demonstrated that AI techniques are successfully used for air pollution modeling mainly in prediction. The main objective of this work is to utilize artificial neural networks (ANNs) as a predictive tool of CO concentrations. In order to ensure the ANNs work effectively, an evolutionary algorithm is proposed to optimize the ANNs weights. The proposed algorithm is Imperialist Competitive Algorithm (ICA). ICA is a global heuristic search method that uses imperialism and imperialistic competition process as a source of inspiration. In this study, the results showed the hybrid ICA-MLP algorithm provides an acceptable performance particularly in the prediction accuracy. The obtained value for mean squared error (MSE) of ICA-MLP model for training and testing are 0.0090 and 0.0166 respectively. Meanwhile, the obtained value for MSE of MLP model for training and testing are 0.0166 and 0.1906. From the comparison results of the proposed approaches, it can be concluded that the hybrid ICA-MLP is more accurate than only MLPin predicting CO pollutant.,Certification of Master's / Doctoral Thesis" is not available
Pages:	48
Call Number:	QA76.76.I58M337 2012 3 tesis
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.