Q4. Explain the reasons to take the log differences rather than the differenced original series in modelling the stochastic term in the series. Q5. ARIMA models include a parameter, d, that controls the number of times a time series is differenced before being modeled by an ARMA process. Why is differencing a time series sometimes necessary? Q6. Define weak stationarity. Why is stationarity a desirable property for a time series process