查詢結果分析
來源資料
相關文獻
- 水文頻率分析--核方法
- Model Indexing and Smoothing Parameter Selection in Nonparametric Function Estimation
- 臺灣北部區域雨量強度公式之研究(2):無母數區域雨量頻率分析
- 非常態週期性水文序列之預處理及其在預測上之應用
- 蓮華池地區降雨-延時-頻率關係及不同降雨延時之設計雨型歷線
- Nonparametric Tests for the Multivariate Multi-Sample Location Problem
- 應用集水區降雨特性進行洪水頻率分析
- 大陸泵浦產業生產曲線估計
- 類神經網路應用於臺灣股市預測:統合基本面與技術面資訊
- 混合機率分布於水文頻率分析適用性之研究
頁籤選單縮合
| 題 名 | 水文頻率分析--核方法=Hydrological Frequency Analysis--Kernel Method |
|---|---|
| 作 者 | 虞國興; 何志軒; | 書刊名 | 農業工程學報 |
| 卷 期 | 42:1 1996.03[民85.03] |
| 頁 次 | 頁13-23 |
| 分類號 | 443 |
| 關鍵詞 | 頻率分析; 無母數; 核方法; Frequency analysis; Non-parameter; Kernel method; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 本研究主要探討Parzen (1962)所提核方法之適用性,並與傳統頻率分析中常用之分布(NOR、LN2、LN3、EV1、PT3及LPT3)做一比較。研究中亦探討五種不同核函數之優劣,並應用核方法分析臺灣地區年最大一日暴雨量。 研究結果顯示,核方法所得結果與以傳統方法且正確選定資料所屬之機率分布所得之精確度為差,此乃核方法無須推估資料所屬機率分布所須付之代價,但因兩者所得之精確度差異不大,顯示核方法之適用性;五種不同核函數之整體表現亦無顯著差異,應用時可任意選用;應用於臺灣地區年一日最大暴雨量之結果與傳統方法中之最佳分布LPT3等差不多。然而由於核方法,不須涉及資料所屬機率分布之判定、資料轉換及偏態係數修正等,應用上較傳統方法為簡便,更具彈性,故核方法值得加以推廣應用於水文頻率分析上。 |
| 英文摘要 | The major objective of this present study is to investigate the aptness of Kernel method proposed by Parzen (1962). The results of Kernel method and those of obtained by traditional frequency analysis by using Normal, 2-parameter Log Normal, 3-parameter Log Normal, Extreme Value Type I Pearson Type III and Log Pearson Type III distributions are compared. The performance of five different Kernel functions are also investigated in this present study. Besides, the annual maximum 1-day rainfall data in Taiwan area are employed to study the aptness of Kernel method for real data. The results indicate that the accuracy of Kernel method is worse than those of obtained by the traditional frequency analysis when the probability density function of data obeyed is accurately detected. This is the price Kernel method must pay because it does not need to know the probability density function. However, the differences between these two methods are not significant. This result strongly suggests that Kernel method be quite apt for frequency analysis. The performance of five different Kernel functions used in the present study are the same. Thus, the choice of Kernel function can be arbitrary. Meanwhile, for the annual maximum 1-day rainfall data, the results of Kernel method are not significantly different with those of obtained by the most appropriate distribution--LPT3. Because Kernel method does not need to identify the probability density function, data transformation or modify the estimated skewness coefficient, it is much easier and flexiable than traditional frequency analysis in practice. Therefore, Kernel method is worth of applying in the hydrological frequency analysis. |
本系統中英文摘要資訊取自各篇刊載內容。