Statistical inference has been very successful with outlier detection in univariate time series data problem. In this paper, we use the method of modified Z-score in detecting outlier observation, and to ascertain the effect of the outlier on the long term behaviour of the series, the outlier observation was deleted. Using the basic idea of underlying statistical inference for time series data gives computational efficiency of achieving better solution in the absence of outlier.

statistical inference, outlier detection, modified Z-score, univariate time series.