Nonlinearity Characteristic for Clean Energy Stock Market: An Integrated Exploration Approach

This paper proposes an integrated nonlinearity exploration approach to discover the nonlinearity characteristic in clean energy stock market, integrating a set of data characteristics analysis technologies. In the proposed approach, the stock data are first analysed in terms of Hurst exponent, a basic nonlinear exponent, from an overall perspective. Second, a series of nonlinear testing methods are employed to analyse sample data from different perspectives:residuals of linear fit-based method is used by testing whether linear method is sufficient to model the data (linearity) or not(nonlinearity); parametric method, by direct comparing whether linear method (linearity) or nonlinear method(nonlinearity) is better for modelling the data; and an indirect method, i.e., surrogate data based method, is employed by testing the main statistic properties of original data and its linear surrogate data similar (linearity) or different(nonlinearity). For illustration purpose, ten time series data of daily return in clean energy stock market are selected as sample data, and the empirical results reveal that all of the mare nonlinear. The results also imply the proposed integrated method can explore the nonlinearity characteristic from different perspectives, avoiding partial analysis results via single method.