Did You Know?


LOESS is an acronym ("LOcal regrESSion") for locally estimated scatterplot smoother. It is a nonparametric method for smoothing a series of data. Therefore there are no assumptions made about the underlying structure of the data. LOESS uses local regression to fit a smooth curve through a scatterplot of data. The LOESS curve is typically smoother than a binomial filter or running average. LOESS is also effective when there are outliers in the data. The LOESS methodology includes techniques for constructing confidence intervals around the curve. For more information see Cleveland, W.S., 1979: Robust locally-weighted regression and smoothing scatterplots. Journal of the American Statistical Association, 74, 829-836. DOI: 10.1080/01621459.1979.10481038