查詢結果分析
來源資料
相關文獻
- Forecasting Trending Tiem Series: A Comparative Analysis of Nonlinear Models
- On the Model-Free Predictor & Its Applications in Economic Forecasting
- 應用狀態空間模型與基因類神經網路濾波技術於風險值預測之研究:以臺股指數與臺指期貨為例
- Neural Network Procedures for Taguchi's Dynamic Problems
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
- 專家系統振動訊號圖型判別之研究
- 反傳遞模糊類神經網路於流量推估之應用
- 類神經網路(Neural Networks)的種類及其在影像處理上的應用
- C++Fuzzy類神經網路物件導向發展系統之建立
- 臺灣汽保費率之估計--對數線性費率模式與類神經網路之比較
頁籤選單縮合
題名 | Forecasting Trending Tiem Series: A Comparative Analysis of Nonlinear Models=趨勢型時間數列預測: 非線性模式之比較研究 |
---|---|
作者姓名(中文) | 吳正義; 吳柏林; | 書刊名 | 正修學報 |
卷期 | 8 1995.07[民84.07] |
頁次 | 頁117-131 |
分類號 | 550.19 |
關鍵詞 | 時間數列趨勢; 狀態空間; 指數成長率; 自由模式; 神經網路; Trending time series; State space; Exponential growth rate; RGARMA; Model free; Neural networks; |
語文 | 英文(English) |
中文摘要 | 傳統上,時間數列分析大都在穩定隨機過程上考慮。但是在實證研究時,很多趨 勢型時間數列資料呈現穩定成長之非線型圖形,如指數成長。本研究即針對此趨勢型時間數 列提出三種模式建構與預測分析方法:( 1 )狀態空間方程式( 2 ) REGARIMAModel ( 3 )神經網路學習法;詳細說明其特性。最後我們以中央政府總支出與教科文總支出做 一實證研究,並比較此三種模式建構與預測分析方法預測效果之優劣。 |
英文摘要 | Conventionally the research on time series analysis has focused upon the modeling of dynamic data under the assumptions of linearity and stationarity. While in the socioeconomic system, many time series are steadily increasing with time and exhibit certain nonlinear shapes such as exponential curve. In this paper, we introduce three alternative techniques for modeling trending data. One is to use the state space representation. Another is to use the reltive growth rate of ARMA models. The other is to use the neurocomputing as the learning and forecasting tools by the neural networks. Finally, we present an illustrative example about the annual expenditures on government and expenditures on science-education-culture of Taiwan and compare the forecasting performance by abov methods. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。