دانلود مقاله ISI انگلیسی شماره 100933
کد مقاله سال انتشار مقاله انگلیسی ترجمه فارسی تعداد کلمات
100933 2018 12 صفحه PDF سفارش دهید 5656 کلمه
خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.
عنوان انگلیسی
Multidimensional scaling analysis of financial time series based on modified cross-sample entropy methods
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Physica A: Statistical Mechanics and its Applications, Volume 500, 15 June 2018, Pages 210-221

پیش نمایش مقاله
پیش نمایش مقاله

چکیده انگلیسی

Stocks, as the concrete manifestation of financial time series with plenty of potential information, are often used in the study of financial time series. In this paper, we utilize the stock data to recognize their patterns through out the dissimilarity matrix based on modified cross-sample entropy, then three-dimensional perceptual maps of the results are provided through multidimensional scaling method. Two modified multidimensional scaling methods are proposed in this paper, that is, multidimensional scaling based on Kronecker-delta cross-sample entropy (MDS-KCSE) and multidimensional scaling based on permutation cross-sample entropy (MDS-PCSE). These two methods use Kronecker-delta based cross-sample entropy and permutation based cross-sample entropy to replace the distance or dissimilarity measurement in classical multidimensional scaling (MDS). Multidimensional scaling based on Chebyshev distance (MDSC) is employed to provide a reference for comparisons. Our analysis reveals a clear clustering both in synthetic data and 18 indices from diverse stock markets. It implies that time series generated by the same model are easier to have similar irregularity than others, and the difference in the stock index, which is caused by the country or region and the different financial policies, can reflect the irregularity in the data. In the synthetic data experiments, not only the time series generated by different models can be distinguished, the one generated under different parameters of the same model can also be detected. In the financial data experiment, the stock indices are clearly divided into five groups. Through analysis, we find that they correspond to five regions, respectively, that is, Europe, North America, South America, Asian-Pacific (with the exception of mainland China), mainland China and Russia. The results also demonstrate that MDS-KCSE and MDS-PCSE provide more effective divisions in experiments than MDSC.

خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.