The Bollinger Band Mode Based On Hilbert-Huang Transformation

The traditional time analysis method is likely to make serious distortion when it is used on some non-stationary time series. However, the Hilbert-Huang Transformation, which was proposed by Huang.N.E in1998, has a good advantage of dealing with non-stationary time series. The Hilbert-Huang Transformation consists of two parts:Empirical Mode Decomposition and Hilbert spectrum analysis. This method is totally adaptive and doesn’t subject to Heisenberg uncertainty principle.In this paper,the author utilizes the HHT theory to process the Shanghai and Shenzhen300stock index futures(from April,2010to Dec,2013),and the steps include applying EMD to the data,calculating Hilbert spectrum, observing the cycle and data denoising.This paper also make a quantitative investment mode-Bollinger Band for the data. This mode can alleviate the signal lag phenomenon of analysis. What’s more, in order to avoid "furute information ",the author keep the Bolliner Band up-to-date based on the past data.The author use the15minutes data to test the performance of the Bollinger mode and calculate its net profit,the maximum drawdown,sharp rate and so on. In addition,this paper also analyses the performance of the mode.