(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Analisis taburan hujan di sitiawan menggunakan rangkaian neural / Yazid Abu

Analisis taburan hujan di sitiawan menggunakan rangkaian neural_Yazid Abu Kasim _E3_2007_875002310
Projek ini membincangkan tentang kecerdikan buatan untuk menganalisis taburan hujan di Sitiawan. Teknik kecerdikan buatan adalah meliputi rangkaian neural buatan, sistem mahir, sistem fuzzi dan sebagainya. Metodologi yang digunakan bagi kecerdikan buatan adalah Perceptron Berbilang Lapisan. Data yang digunakan merangkumi data selama 30 tahun. Bilangan data yang digunakan secara keseluruhan adalah sebanyak 360 data dan dibahagikan kepada dua bahagian iaitu set data latihan dan set data pengujian. Jumlah sampel latihan adalah 200 manakala jumlah sampel ujian adalah 160. Penganalisisan taburan hujan menggunakan kedua-dua kaedah rangkaian neural dapat menentukan samada hujan atau sebaliknya melalui pembelajaran yang dilakukan. Perisian Matlab 7 digunakan untuk mengimplementasikan rangkaian neural buatan ini. Jenis algoritma pembelajaran yang dipilih untuk melatih rangkaian tersebut adalah algoritma `Bayesian regularization`. Fasa latihan dilakukan dahulu sebelum fasa pengujian. Fasa pengujian dilakukan untuk melihat sejauh mana kebolehan sistem rangkaian neural buatan dalam menganalisis taburan hujan. Keputusan yang diperolehi menunjukkan bahawa analisis tersebut memberikan nilai peratus kejituan yang tinggi iaitu 100%. Ini menandakan bahawa rangkaian neural buatan mempunyai keupayaan yang tinggi untuk menganalisis taburan hujan di Sitiawan. _________________________________________________________________________________________ This paper is about artificial intelligence (AI) applications to analyze rainfall in Sitiawan. Beside that, statistical analysis also have been use to analyze rainfall. The AI techniques include artificial neural networks, expert system, fuzzy system and multivariate regression while statistical analysis can be done using SPSS, Stata or SAS. For this project, multilayer perceptron has been chosen among other methodology in neural network. The data consists of 360 data which include 30 years of data from January 1951 to December 1980. All 360 data were divided onto 2 sets of data: a set of 200 samples for training phase and the remaining 160 samples are used to test for the validity and applicability of the artificial intelligence neural network approach. Neural network can detect rainfall or not through the learning process. Matlab 7 is used to design multilayer perceptron. A learning algorithm used in this project to train the multilayer perceptron is Bayesian regularization algorithm. Training process will be done first before testing process. Testing process is used to determine how accurate neural network system in analyze rainfall. From training and testing, the result of back propagation algorithm gives a high percentage of accuracy. The percentage is 100%. The results proved that the multilayer perceptron network has high capability to analyze rainfall in Sitiawan.
Contributor(s):
Yazid Abu Kasim - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875002310
Language:
English
Subject Keywords:
artificial intelligence (AI); rainfall; multilayer perceptron network
First presented to the public:
3/1/2007
Original Publication Date:
9/18/2018
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 50
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2018-09-18 11:07:37.185
Date Last Updated
2019-01-07 11:24:32.9118
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
Analisis taburan hujan di sitiawan menggunakan rangkaian neural / Yazid Abu1 2018-09-18 11:07:37.185