Artificial neural network in time series analysis: An ocean engineering experience

K. Ramanitharan, Chi Wai Li

Research output: Journal article publicationJournal articleAcademic researchpeer-review


The applicability of an Artificial Neural Network (ANN) to the ocean wave system as a forecasting tool is analyzed. ANNs without hidden layers and with linear sigmoid functions are used. Single step ahead forecasting of significant wave period is considered with minimum Root Mean Square (RMS) error as the basis of the stopping criterion. Different networks are tried with different pre-processing of real time data before inputting at input neurons. For comparison, appropriate Auto Regressive Integrated Moving Average (ARIMA) models for the same data set are developed. Results from these models are checked for their statistical correctness. The ANN approach is found more flexible and superior as compared to the ARIMA modeling approach in terms of accuracy and efficiency.
Original languageEnglish
Pages (from-to)781-786
Number of pages6
JournalIntelligent Engineering Systems Through Artificial Neural Networks
Publication statusPublished - 1 Dec 1996

ASJC Scopus subject areas

  • Software


Dive into the research topics of 'Artificial neural network in time series analysis: An ocean engineering experience'. Together they form a unique fingerprint.

Cite this