Evidence of persistent behaviour at Etna: some stochastic thoughts for forecasting