Which of the following are true of the "linear regression" forecasting method? 1. the data pattern can allow for stationarity. 2. a trend in the data can also be spotted, but that trend must be non-linear. 3. the amount of historical data required is typically in the range of 10-20 observations, though more is generally preferred. 4. a trend in the data can also be spotted, but that trend must be linear. 5. linear regression is the most popular of all forecasting methods. 6. seasonality in the data cannot be accomodated. 7. the forecast horizon can range from short to medium.