Outhers detection

Pena D, Prieto FJ. Multivariate outher detection and robust covariance estimation. Technometrics 2001 41 286—300. 

It is usually advisable to plot the observed pairs of y versus r, to support the linearity assumption and to detect potential outhers. Suspected outliers can be omitted from the least-squares Tit and then subsequently tested on the basis of the least-squares fit. 

Some outliers may also be identified by statistical tests (see Chapter 14), but no single method is capable of detecting outhers in every situation that may occur. The number of techniques suggested or recommended is, for that reason, very large. " The two main problems encountered are as follows ... 

We now present and discuss some of the results of the benchmark. The benchmark exercise features aU of the techniques applied to each of the examples, but in the scope of this document we only can present a representative subset. Most of the graphics are shown in form of box plots derived from the 25 available runs per sample size. The box plots show the lower quartile, the median, and the upper quartile values. The whiskers in the plots are lines showing the data range. Outhers are detected using a multiple (here 3) of the inter-quartile range. 

Unusual weather in any month is known to affeet road safety statisties of that month [BER 13]. In order to explain the outliers with low values, averaged monthly temperature data were collected and compared with the seasonal averages for the month Thus, six periods of cold spells were detected, and the three most severe were retained for explaining the low outhers of January 2006, January 2010 and December 2010. 

---