Forecasting pelagic fish in Malaysia using ets state space approach
Modelling and forecasting fish catch has been undertaken for a long time over the world. However, From time to time, researchers are always looking for a new model that can predict more accurately the number of fish catch. The objective of this study is to propose the Error Trend and Seasonal...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/1471/1/24p%20HADIZA%20YAKUBU%20BAKO.pdf http://eprints.uthm.edu.my/1471/2/HADIZA%20YAKUBU%20BAKO%20COPYRIGHT%20DECLARATION.pdf http://eprints.uthm.edu.my/1471/3/HADIZA%20YAKUBU%20BAKO%20WATERMARK.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Modelling and forecasting fish catch has been undertaken for a long time over the
world. However, From time to time, researchers are always looking for a new model
that can predict more accurately the number of fish catch. The objective of this study
is to propose the Error Trend and Seasonal (ETS) state space approach.In this study,
two techniques of time series analysis were used to forecast fish catch of three
commercial fish species found in the Malaysian waters. One of such techniques is the
Box-Jenkins method which concerns the building of linear and stochastic dynamic
models with minimum data requirements. The second technique is the Error Trend
and Seasonal (ETS) state space exponential method which requires no assumptions
about the correlations between successive values of the time series. The two class
models were used to model and forecast two years monthly catches of the three fish
species based on the collected data for the period 2007 – 2011. The
SARIMA(1,1,1)(0,0,1)[12], SARIMA(1,1,4)(0,0,1)[12], SARIMA(2,1,1)(0,0,1)[12]
and ETS (M, A, M), ETS (M, N, M), ETS (M, A, M) for Dussumiera acuta (tamban
buloh), Rastrelliger kanagurta (kembong) and Thunnus tonggol (Tongkol hitam)
were proposed respectively. The diagnostic checking for all the fitted models
confirmed the adequacy of the models. Results based on the root mean square error
(RMSE) and mean absolute error (MAE) demonstrated that the ETS models per�formed better for Thunnus tonggol and Rastrelliger kanagurta, while SARIMA
model performed better for Dussumiera acuta. This shows that ETS model which has
so far not been used in fisheries in Malaysia is our main contribution in this research.
Nevertheless, both models have proven successful in describing and forecasting the
monthly fishery dynamics. These forecasts proves helpful in formulating the needed
strategies for sustainable management and conservation of the stocks, and can also
help the decision makers to establish priorities in terms of fisheries management. |
|---|