Does a correlation coefficient close to 1 or -1 indicate the strength of the effect? How can one use the confidence interval for the linear regression slope to test a null hypothesis of zero linear correlation? Under what conditions can one successfully use linear regression to forecast future data? Why does the exponentially smoothed time series data appear to be more amenable to forecasting as opposed to raw time series data
