Affiliations 

  • 1 Universiti Putra Malaysia
  • 2 Universiti Teknologi Malaysia
ASM Science Journal, 2014;8(2):150-158.
MyJurnal

Abstract

We had reviewed the current practice in stocks market analysis where stock is represented by its closing price, and then found that this approach may be misleading. In actuality, in the daily activity of stocks market, stock is represented by four prices, namely opening, highest, lowest, and closing prices. Thus, stock is a multivariate time series of those four prices and not a univariate time series of closing price only. In this paper all four prices will be considered. Then, the notion of multivariate time series similarity among stocks will be developed as a generalisation of univariate time series similarity. The results are used to construct stocks network in multivariate setting. To filter the economic information contained in that network, the standard tools in econophysics is used. Furthermore, the topological properties of stocks are analysed by using the most common centrality measures. As an example, Bursa Malaysia data are investigated and we show that the proposed approach can better figure out the real situation compared to the current one.