本文選題：融資融券　切入點(diǎn)：時間序列特征提取　出處：《山東大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：時間序列的聚類分析的研究對象常見于面板數(shù)據(jù),對時間序列進(jìn)行聚類分析則不同,是基于無監(jiān)督性學(xué)習(xí)的分類算法機(jī)制,在研究不同序列之間的相關(guān)性問題上能夠挖掘更深層次和更高維度的信息,這種思路在各類基于分形學(xué)與模式識別的量化交易中均有所體現(xiàn)。時間序列相似性的傳統(tǒng)判定方法多通過計算測試序列與目標(biāo)序列相對應(yīng)時刻的點(diǎn)與點(diǎn)之間的歐氏距離來進(jìn)行判定,進(jìn)而完成聚類的目的。但是傳統(tǒng)方法對于不同采樣頻率帶來的缺失點(diǎn)以及異常的離群點(diǎn)的處理上,有較大的缺陷,算法執(zhí)行起來會造成較大的偏差,聚類效果難以讓人滿意。本文給出了一種新型的判斷時間序列相似性的方法——特征因子判別法,通過對時間序列的統(tǒng)計特征、線性及非線性特征多個方面進(jìn)行特征提取,將一維的與時刻一一對應(yīng)的時間序列數(shù)據(jù)映射到高維空間,并據(jù)此構(gòu)建與之相對應(yīng)的特征向量,再通過主成分分析的方法,借以消除特征因子之間可能存在的多重共線性問題,最后使用K-means無監(jiān)督性學(xué)習(xí)算法對數(shù)據(jù)樣本進(jìn)行分類,達(dá)到數(shù)據(jù)層面的類別劃分的效果,在此基礎(chǔ)上,移動時間窗口重復(fù)以上操作,得到新的簇,與上次聚類得到的簇之間兩兩求交,交集占比最大的即我們所希望找到的相似性證券。本文以融資融券推出以來的交易數(shù)據(jù)與融資融券標(biāo)的股票股價走勢關(guān)系為例,通過提取其特征因子組成特征向量,劃分樣本內(nèi)外集合,將樣本內(nèi)的特征矩陣作為K-means算法的輸入樣本,進(jìn)行無監(jiān)督性學(xué)習(xí)分類,達(dá)到聚類分析的效果,從理論上看,聚類分析得到的相似性組合應(yīng)當(dāng),與市場指數(shù)走勢保持著較強(qiáng)正相關(guān)性的同時,還具備較強(qiáng)的趨勢性走勢,進(jìn)而可以嘗試通過量化擇時來對趨勢性對象進(jìn)行研究。與傳統(tǒng)方法相比,減小了不同長度時間序列數(shù)據(jù)缺失點(diǎn)與異常離群點(diǎn)的影響,提高了聚類分析的準(zhǔn)確性和時間序列識別的相似度問題。在此基礎(chǔ)上,探究其在實(shí)際市場中的應(yīng)用,將聚類分析得到的相似性證券構(gòu)建投資組合,探究其在樣本外與市場指數(shù)走勢之間的關(guān)系并探究兩者的價差,從頻域角度對傳統(tǒng)的EMA指標(biāo)進(jìn)行解讀,借助信號處理的觀點(diǎn),對其進(jìn)行修正,構(gòu)建二階低通濾波器盡可能地過濾高頻噪音信號保留低頻信號作為低延遲均線,并以此對兩者的價差進(jìn)行交易性擇時,根據(jù)其斜率的正負(fù)情況給出看多或者看空價差組合的觀點(diǎn)。觀察價差組合在樣本外的表現(xiàn),證實(shí)我們通過低延遲均線系統(tǒng)對聚類分析得到投資組合與市場指數(shù)滬深300構(gòu)造的價差進(jìn)行量化擇時的可行性和合理性。
[Abstract]:The research object of clustering analysis of time series is usually panel data, but the clustering analysis of time series is different, which is based on the unsupervised learning mechanism of classification algorithm. In studying the correlation between different sequences, we can mine deeper and higher-dimensional information. This idea is embodied in all kinds of quantitative transactions based on fractal and pattern recognition. The traditional method of judging the similarity of time series by calculating the points and points between the points corresponding to the test sequence and the target sequence. To determine the distance, But the traditional methods have great defects in the processing of missing points and abnormal outliers brought about by different sampling frequencies. The clustering effect is not satisfactory. This paper presents a new method to judge the similarity of time series, which is the feature factor discrimination method. The feature extraction is carried out from the statistical features, linear and nonlinear features of the time series. The one-to-one time series data of one dimension are mapped to the high-dimensional space, and the corresponding eigenvector is constructed accordingly, and then by principal component analysis (PCA), the problem of multiple collinearity between the feature factors can be eliminated. Finally, K-means unsupervised learning algorithm is used to classify the data samples to achieve the effect of data level classification. On this basis, the moving time window repeats the above operations to get a new cluster. In this paper, the relationship between the trading data and the stock price trend of the underlying stock is taken as an example. By extracting the feature factors to form the feature vector, dividing the set inside and outside the sample, taking the feature matrix in the sample as the input sample of K-means algorithm, the unsupervised learning classification is carried out to achieve the effect of clustering analysis. The similarity combination obtained by cluster analysis should keep a strong positive correlation with the market index trend, but also have a strong trend trend. Furthermore, we can try to study the trend objects by quantitative timing. Compared with the traditional method, the influence of missing points and abnormal outliers in time series of different lengths is reduced. The accuracy of clustering analysis and the similarity of time series recognition are improved. On this basis, the application of clustering analysis in real market is explored, and the portfolio of similar securities obtained by clustering analysis is constructed. This paper probes into the relationship between the outside sample and the trend of the market index, and probes into the price difference between the two, and interprets the traditional EMA index from the angle of frequency domain, and modifies it with the help of the viewpoint of signal processing. The second order low-pass filter is constructed to filter the high-frequency noise signal as much as possible to retain the low-frequency signal as the low-delay mean line, and to make a transactional timing for the price difference between the two. According to the positive and negative slope, the view of the combination of long or short spread is given. It is proved that the feasibility and rationality of the quantitative timing of the spread between the investment portfolio and the Hu-Shen 300 structure of the market index is obtained by using the low delay mean line system.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：F832.51;F224

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 萬濤;鄭婷婷;張琛;章意成;;不同股市的多重分形特性分析——基于統(tǒng)計物理和MF-DFA方法[J];計算機(jī)技術(shù)與發(fā)展;2010年12期

2 劉維奇;牛奉高;;滬深兩市多重分形特征的成因及其變化[J];經(jīng)濟(jì)管理;2009年12期

3 鐘維年;高清維;陳燕玲;;基于小波和多重分形的金融時間序列聚類[J];系統(tǒng)工程;2009年03期

4 苑瑩;莊新田;;金融時間序列的標(biāo)度特性實(shí)證研究[J];管理工程學(xué)報;2008年02期

5 苑瑩;莊新田;;國際匯率的多重分形消除趨勢波動分析[J];管理科學(xué);2007年04期

6 黃超;吳清烈;武忠;朱揚(yáng)勇;;基于方差波動多重分形特征的金融時間序列聚類[J];系統(tǒng)工程;2006年06期

7 廖士光;楊朝軍;;賣空交易機(jī)制對股價的影響——來自臺灣股市的經(jīng)驗(yàn)證據(jù)[J];金融研究;2005年10期

8 胡雪明,宋學(xué)鋒;深滬股票市場的多重分形分析[J];數(shù)量經(jīng)濟(jì)技術(shù)經(jīng)濟(jì)研究;2003年08期

，

本文編號：1584631