Exportar Publicação

Manuela Oliveira, AMC, JW, Vera Miguéis , Ferreira, N. B. & Miguel Gaspar (2017). Forecasting of bivalve landings with multiple regression and data mining: The case of the Portuguese artisanal dredge fleet. 21st Conference of the International Federation of Operational Research Societies.

	. M. Oliveira et al.,  "Forecasting of bivalve landings with multiple regression and data mining: The case of the Portuguese artisanal dredge fleet", in 21st Conf. of the Int. Federation of Operational Research Societies, Quebec, 2017

@misc{oliveira2017_1733301063071,
	author = "Manuela Oliveira and AMC and JW and Vera Miguéis  and Ferreira, N. B. and Miguel Gaspar",
	title = "Forecasting of bivalve landings with multiple regression and data mining: The case of the Portuguese artisanal dredge fleet",
	year = "2017",
	howpublished = "Digital",
	url = "http://ifors2017.ca/"
}

TY  - CPAPER
TI  - Forecasting of bivalve landings with multiple regression and data mining: The case of the Portuguese artisanal dredge fleet
T2  - 21st Conference of the International Federation of Operational Research Societies
AU  - Manuela Oliveira
AU  - AMC
AU  - JW
AU  - Vera Miguéis 
AU  - Ferreira, N. B.
AU  - Miguel Gaspar
PY  - 2017
CY  - Quebec
UR  - http://ifors2017.ca/
AB  - The bivalve dredge fishery is one of the most important artisanal fisheries in mainland Portugal involving a large number of fishers and vessels and the value of catches represents a large proportion of all revenue from traditional fisheries of coastal communities.The sustainability of this fishery has been at risk in the last few years, in part due to the occasional compulsory closures of the fishery activity as a result of phytotoxin episodes.In the absence of an accurate system to predict these phenomena, aggravated by their increased frequency, the goal of this analysis is to develop a decision support tool that can help administrative fishery authorities to forecast bivalve landings accounting for several contextual conditions.With data of 6 years relating to indicators of vessels characterization, fishing effort, weather conditions, phytotoxin episodes, stock-biomass indicators per species and tourism levels, it was explored the relationship between these factors and the monthly quantities landed using multiple linear regression models. The results showed that the impact of the contextual factors varies between regions, and also depends of the vessels target species. The accuracy of monthly bivalve landings forecasts was then improved using a Data Mining technique (Random Forests). This model has proved to be a robust decision support technique in this context, as the forecasts obtained showed accuracy levels ranging from 74% in the Southwest coast to 99% in the South.
ER  -