本文選題：WACD模型　切入點(diǎn)：超高頻數(shù)據(jù)　出處：《上海師范大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：伴隨著大數(shù)據(jù)時(shí)代的到來(lái)，計(jì)算機(jī)存儲(chǔ)海量數(shù)據(jù)的技術(shù)及相關(guān)的數(shù)據(jù)統(tǒng)計(jì)和數(shù)據(jù)挖掘發(fā)現(xiàn)理論不斷發(fā)展，這類技術(shù)方法也將必然會(huì)用到股票市場(chǎng)的數(shù)據(jù)中。傳統(tǒng)分析股票市場(chǎng)數(shù)據(jù)都是以單位時(shí)間為軸，，運(yùn)用GARCH、SV等模型對(duì)這類數(shù)據(jù)進(jìn)行建模，旨在發(fā)現(xiàn)時(shí)間序列上的交易量、價(jià)格等波動(dòng)性和集聚性等特征。這類模型的一個(gè)重要特點(diǎn)是等間隔取樣，當(dāng)間隔不等時(shí)，運(yùn)用該類模型分析容易造成結(jié)論的失真。 在股市中，由各類逐筆交易構(gòu)成的數(shù)據(jù)我們稱之為超高頻數(shù)據(jù)。該類數(shù)據(jù)的一個(gè)重要特點(diǎn)是交易時(shí)隨機(jī)到達(dá)的，即到達(dá)的時(shí)間間隔不等，這類數(shù)據(jù)一般有時(shí)間間隔、交易及價(jià)格等幾個(gè)變量。任何擯棄時(shí)間間隔的分析方法都有一定得不合理性，此前傳統(tǒng)的模型顯然不太適用這類超高頻數(shù)據(jù)，因此需要考慮到時(shí)間變量對(duì)此類超高頻數(shù)據(jù)建模。另外，證券市場(chǎng)的一個(gè)主要功能就是在交易成本盡可能低的情況下，使投資者能夠迅速、有效地執(zhí)行交易。換句話說(shuō),也就是市場(chǎng)必須提供足夠的流動(dòng)性。 本文正是在這樣的背景下，從時(shí)間的角度根據(jù)交易量的變化提出了使用交易量期間來(lái)刻畫(huà)日內(nèi)流動(dòng)性的觀點(diǎn)，利用這一指標(biāo)來(lái)衡量執(zhí)行完給定交易量所耗費(fèi)的時(shí)間。不僅通過(guò)整體交易量期間來(lái)度量整體的流動(dòng)性，而且同時(shí)也利用主動(dòng)性賣出或買入交易量期間分別來(lái)度量單邊的市場(chǎng)流動(dòng)性。由于交易量期間是隨著時(shí)間而動(dòng)態(tài)變化的，并且具有一定得集聚性，因此其也是可以被預(yù)測(cè)的。 本文首先回顧了國(guó)內(nèi)外相關(guān)領(lǐng)域的研宄現(xiàn)狀和流動(dòng)性的概念及度量方法,隨后就使用了較大篇幅來(lái)介紹針對(duì)期間的計(jì)量模型分析框架，涵蓋了線形和非線性ACD模型（以WACD(1，1)模型為例），期間數(shù)據(jù)的處理和分析過(guò)程。本文選取了上海、深圳（主板/中小板/創(chuàng)業(yè)板）的12只股票為樣本，從整體和主動(dòng)性買單以及主動(dòng)賣單構(gòu)成的交易量持續(xù)期三個(gè)維度進(jìn)行對(duì)比分析，刻畫(huà)其流動(dòng)性特點(diǎn)。在交易量持續(xù)期基礎(chǔ)上，同時(shí)引進(jìn)價(jià)格變化，針對(duì)不等時(shí)間間隔的數(shù)據(jù)進(jìn)行調(diào)整后，運(yùn)用UHF-GARCH模型分析持續(xù)期條件下的價(jià)格分布特點(diǎn)^發(fā)現(xiàn)WACD(1，1)和UHF-GARCH模型能夠較好的分析我國(guó)證券市場(chǎng)的流動(dòng)性。
[Abstract]:With the advent of big data's era, the technology of computer storage of massive data and the related data statistics and data mining discovery theory are constantly developing. The traditional analysis of stock market data is based on the unit time axis and models such as GARCHN SV are used to model the data in order to find the trading volume in time series. One of the important characteristics of this kind of model is the equi-interval sampling. When the interval is not equal, it is easy to use this model to analyze the distortion of the conclusion. In the stock market, data consisting of all kinds of individual trades is called UHF data. One of the important features of these data is the random arrival at the time of trading, that is, the time interval of arrival varies, and such data usually have time intervals. Some variables, such as trading and price. Any analytical method that abandons the time interval has some irrationality. The traditional model is obviously not suitable for this kind of UHF data. One of the main functions of the securities market is to enable investors to execute transactions quickly and efficiently at the lowest possible transaction cost. In other words, That is, the market must provide sufficient liquidity. In this context, this paper puts forward the viewpoint of using trading volume period to depict intraday liquidity from the perspective of time, according to the change of trading volume. This indicator is used to measure the time taken to execute a given transaction volume. At the same time, the active selling or buying trading volume is used to measure the unilateral market liquidity, which can be predicted because the trading volume period changes dynamically with time and has a certain degree of agglomeration. This paper first reviews the current research situation and the concept and measurement methods of liquidity in related fields at home and abroad, and then introduces the econometric model analysis framework for the period. This paper covers the linear and nonlinear ACD model (taking WACD-1) model as an example, and the process of data processing and analysis. 12 stocks in Shanghai and Shenzhen (main board / small and medium-sized board / gem) are selected as samples. This paper makes a comparative analysis from the three dimensions of transaction duration, which consists of total and active payment and active selling order, to characterize its liquidity characteristics. On the basis of the duration of trading volume, the price changes are introduced at the same time. After adjusting the data of different time intervals, it is found that the UHF-GARCH model and the UHF-GARCH model can better analyze the liquidity of China's securities market by using the UHF-GARCH model to analyze the price distribution characteristics under the condition of duration.
【學(xué)位授予單位】：上海師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：F224;F832.51

【參考文獻(xiàn)】

相關(guān)期刊論文 前9條

1 蔡艷萍;謝家泉;;中國(guó)股市收益率波動(dòng)實(shí)證研究──基于自回歸條件持續(xù)性模型[J];財(cái)經(jīng)理論與實(shí)踐;2006年01期

2 曾勇,王志剛,李平;基于高頻數(shù)據(jù)的金融市場(chǎng)微觀結(jié)構(gòu)實(shí)證研究綜述[J];系統(tǒng)工程;2005年03期

3 陳怡玲,宋逢明;中國(guó)股市價(jià)格變動(dòng)與交易量關(guān)系的實(shí)證研究[J];管理科學(xué)學(xué)報(bào);2000年02期

4 劉向麗;程剛;成思危;汪壽陽(yáng);洪永淼;;中國(guó)期貨市場(chǎng)價(jià)格久期波動(dòng)聚類特征研究[J];管理科學(xué)學(xué)報(bào);2010年05期

5 馬丹;尹優(yōu)平;;交易間隔、波動(dòng)性和微觀市場(chǎng)結(jié)構(gòu)——對(duì)中國(guó)證券市場(chǎng)交易間隔信息傳導(dǎo)的實(shí)證分析[J];金融研究;2007年07期

6 劉偉;陳敏;吳武清;;高頻數(shù)據(jù)交易量久期與價(jià)格變化的動(dòng)態(tài)行為研究[J];數(shù)理統(tǒng)計(jì)與管理;2010年03期

7 補(bǔ)馮林,張衛(wèi)國(guó),何偉;基于超高頻數(shù)據(jù)的股票流動(dòng)性度量研究[J];統(tǒng)計(jì)與決策;2005年04期

8 徐國(guó)祥;金登貴;;基于金融高頻數(shù)據(jù)的ACD模型非參數(shù)設(shè)定檢驗(yàn)[J];統(tǒng)計(jì)研究;2007年04期

9 戴麗娜;;非參數(shù)可加ACD模型及其應(yīng)用研究[J];統(tǒng)計(jì)與信息論壇;2008年02期

本文編號(hào)：1641188