I have identified the outlier of the time series "duration of conversation in minutes" of the telephone traffic to an operator, through residuals analysis.I must provide an estimate of the abnormality and I thought to consider the value of the residual and multiply by the coefficient of variation of estimated values by the model. I also calculated the coefficient of variation of the average time of conversation that best interprets the trend of time series. How to can calculate a single measure of variability for these two indices? There is another multiplicative factor which allows to obtain a correct estimate abnormality? -- View this message in context: http://www.nabble.com/estimate-abnormality-tp23277168p23277168.html Sent from the R help mailing list archive at Nabble.com.