查詢結果分析
相關文獻
- Comparisons of Time Series Models for the Forecasting of Albacore Commercial Harvest in the Indian Ocean
- Reconsideration of Assessing South Pacific Albacore Stocks(Thynnus alalunga)
- Simulation of Age-Structured Leslie Matrix Model to Evaluate the Reproductive Values of the North Atlantic Albacore Stock
- Recruit Estimation and Its Relationship with Spawner for the North Atlantic Albacore Stock Using Commercial Catch Data
- 時間序列分析在品質管制上的應用
- Estimation of SOPDT Transfer Function Models Using an Auto-Tuning Test
- Seasonal Changes of the Distribution of South Pacific AlBacore Based on Taiwan's Tuna Longline Fisheries, 1971-1985
- 北太平洋長鰭鮪資源的研究與開發(2):表層漁業漁場
- Examination on Recent Increase in Juvenile Catch of South Atlantic Albacore Thunnus alalunga: Observations and Implications
- Fluctuation of the South Pacific Albacore Stocks(Thynnus alalunga) Relative to the Sea Surface Temperature
頁籤選單縮合
題 名 | Comparisons of Time Series Models for the Forecasting of Albacore Commercial Harvest in the Indian Ocean=印度洋長鰭鮪商業性漁獲預測之時序列模式的比較 |
---|---|
作 者 | 汪淑慧; 許建宗; 劉錫江; | 書刊名 | 臺灣水產學會刊 |
卷 期 | 27:2 2000.06[民89.06] |
頁 次 | 頁63-75 |
分類號 | 439.2 |
關鍵詞 | 時間序列分析; 自我回歸整合移動平均模式; 轉換函數模式; 長鰭鮪; Time series analysis; ARIMA; Transfer function; Albacore; |
語 文 | 英文(English) |
中文摘要 | 本研究比較不同時間序列模式用來預測印度洋長鰭鮪之漁獲量,以月別漁獲量為因 變數及經一般線性模式法校正後的月別單位努力漁獲量為自變數,建立三類時間序列模式, 分吸為:自我回歸整合移動平均模式、回繳模式涵自我回繳整合移動平均誤差項及轉換函數 模式。月別漁獲量序列分吸以1968-1996年為內部模擬區間及1997年為外部預測區間。六種 統計決策值分別用來比較三類模式的預測能力。 結果顯示,三類時間序列模式皆能模擬漁獲量序列的年內季節變動性和年間的震盪,且 時延遲並不存在於月別漁獲量和其相對的月別標準化單位努力漁獲量之間,個別的統計誤差 決策值比較顯示,在內部模擬區間,雙變量模式略便於單變量模式;在外部預測區間,單變 量模式則較優於雙變量模式,但全部統計百分誤差決策值比較之下,雙變數之回歸模式涵自 我回歸整合移動平均誤差項則在預測印度洋長鰭鮪漁獲量上似較優於單變數之自我回歸整合 移動平均模式,但具有較保守的預測值。 |
英文摘要 | Time series models were used and compared to forecast commercial catch of albacore in the Indian Ocean. Three time series models, i.e., auto-regressive integrated moving average (ARIMA), regression model with the ARIMA error (RAE) and transfer function noise (TN) models were built. Monthly catch was used as dependent variable and adjusted catch per unit effort (ACPUE) standardized by the general linear model was used as the independent variable in the RAE and TN models. Catch data 1969-1996 were used as the inner simulated period, and 1997 data were used in the outer forecasting period. Six statistical criteria, ME, MPE, MAE, MAPE, RMSE and RMSPE were used to evaluated the performance of the built models. The results indicate that three selected time series models can closely trace the pattern of the yearly periodicity and the fluctuation of the catch series. Time delay did not exist between the catch and the ACPUE series. The bivariate model (TN, RAE) used in the inner simulation and the ARIMA model used in outer forecasting seems sound by the individual comparison of the statistical criteria in the present study. For overall comparison the bivariate RAE model seems best among the three built time series models using in forecasting the albacore catch in the Indian Ocean, but a conservative prediction may be observed during forecasting using RAE. |
本系統中英文摘要資訊取自各篇刊載內容。