Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal JOIV : International Journal on Informatics Visualization
Aisy, Rahida Rihhadatul
Unknown Affiliation
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Multivariate Time Series Forecasting using Hybrid Vector Autoregressive and Neural Network for Coupled Roll-Sway-Yaw Motions Prediction Suhermi, Novri; Suhartono, -; Rahayu, Santi Puteri; Ali, Baharuddin; Dahlila, Dea; Aisy, Rahida Rihhadatul
JOIV : International Journal on Informatics Visualization Vol 9, No 4 (2025)
Publisher : Society of Visual Informatics

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62527/joiv.9.4.3077

Abstract

There are six types of motion referred to as the six degrees of freedom, which define the motion of a ship. For a ship to remain stable, it must be in a symmetrical position. Therefore, a ship's stability can be determined based on its motion. Ship motions can be analyzed either in an uncoupled system or a coupled system. One of the coupled motion systems that is often studied is the roll-sway-yaw motion. In this study, we apply the Hybrid Vector Autoregressive–Neural Network (VAR-NN) model to build a multivariate time series model for predicting the roll-sway-yaw motions of a prototype ship. The Hybrid VAR-NN is a data analysis technique that integrates the linear capabilities of the VAR model with the nonlinear capabilities of the NN model to capture both linear and nonlinear trends simultaneously. The dataset for this study was generated from waves in a prototype ship experiment and divided into in-sample and out-of-sample data. The model was trained using the in-sample data, and predictions were made on the out-of-sample data using the trained model. The forecast results of the VAR-NN model were compared with those from the pure VAR and pure NN models. Model selection was based on out-of-sample performance criteria, with the Root Mean Square Error (RMSE) employed as the prediction performance metric. According to the experimental results, the Hybrid VAR-NN model outperformed the other models, demonstrating its ability to improve the prediction performance of the pure models through its hybrid approach.
Co-Authors - Suhartono Baharuddin Ali, Baharuddin Dahlila, Dea Santi Puteri Rahayu Suhermi, Novri