Better than DFA? A Bayesian Method for Estimating the Hurst Exponent in Behavioral Sciences