Abstract
We present a new technical approach based on the autocorrelation function, widely used in physics, to determine and to analyze the business cycle turning points of an economic activity. This method is adapted to stochastic processes and does not require a smoothing technique. The application of this method to the industrial production seasonally adjusted of Tunisia, for the period 1994:4-2006:8 gives similar results to these obtained by two-state Markov switching model.