DEVELOPMENT OF STATISTICAL ANALYSIS OF REMOTE OBSERVATIONS (OUTLIERS) FROM THE END OF THE WWI INTIL THE MID-80S OF THE XX CENTURY

Lyubomir Ivanov

Abstract: In the paper we consider the historical development of the methods for statistical analysis of the outlying observation (aka outliers), covering the period from the end of the First World War till the middle of the 80s years of the XX century. During that timeframe in the analysis are formed two main directions - robust evaluation methods and outlier detection. The first one encompasses the development of methods for accurate estimation even in the presence of the extreme values in the data, while the second one overawes the creation of statistical tests that can identify both the number and location of the outliers. The distinction between the two tracks led to their gradual separation into completely independent fields of the statistical analysis. In the same time with the development of the methods for robust estimation and detection of outliers, the researchers discovered many new problem areas - outlier's dimensions, masking, time series outliers; as well as new possibilities to enrich the analysis by application of graphical methods and automation of analytic procedures.

Keywords: Outliers, Extreme values, Robust evaluation methods, Outlier detection

Date published: 2020-09-18