本文選題：參數(shù)估計(jì) + 非參數(shù)核估計(jì)�。� 參考：《安慶師范大學(xué)》2016年碩士論文

【摘要】：水文時(shí)間序列是一個(gè)錯(cuò)綜復(fù)雜的不確定性研究系統(tǒng),它的主要研究內(nèi)容有水文模擬、水文頻率計(jì)算和水文預(yù)報(bào)等三大模塊知識(shí)。本文在學(xué)習(xí)、總結(jié)和引用國內(nèi)外學(xué)者對水文學(xué)統(tǒng)計(jì)理論研究的基礎(chǔ)上,分析、對比、總結(jié)并在水文頻率計(jì)算方面應(yīng)用了參數(shù)估計(jì)和非參數(shù)估計(jì)兩大類模型,而且應(yīng)用時(shí)間序列相關(guān)模型以及水文模擬相關(guān)知識(shí)進(jìn)行了水文預(yù)報(bào)。本文的研究內(nèi)容和結(jié)論如下:參數(shù)估計(jì)的幾大類方法對比分析及應(yīng)用。水文水資源系統(tǒng)的參數(shù)估計(jì)方法有矩法、權(quán)函數(shù)法、概率權(quán)重法、線性矩法和適線法等;經(jīng)過無偏性檢驗(yàn)可知權(quán)函數(shù)法的無偏性最好,然后依次是線性矩法、概率權(quán)重法和常規(guī)矩法;且它們的穩(wěn)健性能的排名分別為線性矩法、權(quán)函數(shù)法、概率權(quán)重法和常規(guī)矩法;另外,適線法適合在水文工程上的機(jī)器模擬;因此,我們用無偏性與穩(wěn)健性都較好的線性矩法來估計(jì)皮爾遜Ⅲ型分布,進(jìn)而求出其頻率對應(yīng)表。非參數(shù)核密度估計(jì)及應(yīng)用。水文頻率計(jì)算中非參數(shù)核估計(jì)無需假定總體分布類型,只需要利用最小二乘交叉驗(yàn)證法(簡稱LSCV法)求取恰當(dāng)?shù)拇皩捯约扒‘?dāng)?shù)暮撕瘮?shù),便可求出水文頻率密度函數(shù)。進(jìn)而,由非參數(shù)核密度估計(jì)對長江安慶站52年的最大洪峰流量進(jìn)行估計(jì),得到其洪水頻率函數(shù),并與參數(shù)估計(jì)的結(jié)果進(jìn)行對比,可知在總體分布類型不確定時(shí)用非參數(shù)法更適合,而確定時(shí)二者偏差不大,但參數(shù)法更具有針對性。時(shí)間序列分析法建模。針對長江安慶水文站1959年至2010年這52年間的歷年月流量數(shù)據(jù)進(jìn)行了季節(jié)調(diào)整以及趨勢分解;并對其總流量數(shù)據(jù)進(jìn)行了描述性時(shí)序分析、平穩(wěn)性檢驗(yàn)、模型判定,最終建立水文預(yù)報(bào)的ARMA模型;最后利用水文模擬中的蒙特卡洛法對該水文預(yù)報(bào)模型進(jìn)行穩(wěn)健性檢驗(yàn),可知模型可靠,可以用來對安慶的水文環(huán)境進(jìn)行預(yù)報(bào)。
[Abstract]:Hydrological time series is a complicated uncertainty research system. Its main research contents include three modules: hydrological simulation hydrological frequency calculation and hydrological forecasting. On the basis of studying, summarizing and quoting domestic and foreign scholars' research on hydrological statistics theory, this paper analyzes, compares, sums up and applies two kinds of models, parameter estimation and non-parameter estimation, in hydrologic frequency calculation. Moreover, hydrological prediction is carried out by using time series correlation model and hydrologic simulation knowledge. The contents and conclusions of this paper are as follows: comparison and application of several methods for parameter estimation. The parameter estimation methods of hydrology and water resources system include moment method, weight function method, probability weight method, linear moment method and line fitting method, etc. The unbiased test shows that the weight function method is the best, followed by the linear moment method. The probability weight method and the regular rule method are ranked as linear moment method, weight function method, probability weight method and regular rule method respectively. In addition, the line fitting method is suitable for the machine simulation of hydrological engineering. We estimate Pearson type 鈪，

本文編號(hào)：2054825