The traditional spectrum analysis it from Engineering, how every the financial time series are very complicated and have nonlinear term, so other methods are tested, such as fractal Brownian Motion and Hurst statistic for long memory.
However according to my knowledge, the industry does not like too complex ones, and NP is good for VaR, High freq is for fund strategy.
If there is no meaning for arbitrage, no one cares about Hurst nor Fractal BM. Nor Levy, there is a post on Copula, but I think its application is done in market and its econometrics estimation method is too difficult to set up and hardly can find a good theoretical frame work such as ARMA or Kalman filter.
However according to my knowledge, the industry does not like too complex ones, and NP is good for VaR, High freq is for fund strategy.
If there is no meaning for arbitrage, no one cares about Hurst nor Fractal BM. Nor Levy, there is a post on Copula, but I think its application is done in market and its econometrics estimation method is too difficult to set up and hardly can find a good theoretical frame work such as ARMA or Kalman filter.
