What are they, and why should we care

Outliers can be defined as any observation that does not fit a pattern. In analytical chemistry, an outlier could be a pure analyte concentration that is Very different from the rest of the analyte concentrations obtained from a set of calibration samples. In a typical quantitative calibration problem in PAC, one can encounter three different types of outliers  

It is very important to note that the term outlier does not imply incorrect. An outlier could be caused by an error or an incorrect action, but it could just as easily be caused by a real phenomenon that is relevant to the problem. 

Outliers demand special attention in chemometrics for several different reasons. In calibration data, their extremeness often gives them an unduly high influence in the calculation of the calibration model. Therefore, if they represent erroneous readings, then they will add disproportionately more error to the calibration model. Even if they represent informative data, it might be determined that this specific information does not need to be included in the model. 

Outliers are also very important when one is applying a model because they can be used to indicate whether the model is being applied to an inappropriate sample. Details regarding such on-line outlier detection are provided in a later section (Section 8.4.3). 

---