Measure of Surprise for Outlier Detection-ECO6416

Measure of Surprise for Outlier Detection-ECO6416 - Measure...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Measure of Surprise for Outlier Detection Robust statistical techniques are needed to cope with any undetected outliers; otherwise they are more likely to invalidate the conditions underlying statistical techniques , and they may seriously distort estimates and produce misleading conclusions in test of hypotheses. A common approach consists of assuming that contaminating models, different from the one generating the rest of the data, generate the (possible) outliers. Because of a potentially large variance, outliers could be the outcome of sampling errors or clerical errors such as recording data. Therefore, you must be very careful and cautious. Before declaring an observation "an outlier," find out why and how such observation occurred. It could even be an error at the data entering stage while using any computer package. In practice, any observation with a standardized value greater than 2.5 in absolute value is a candidate for being an outlier. In such a case, one must first investigate the source of the datum. candidate for being an outlier....
View Full Document

This note was uploaded on 10/04/2011 for the course ECO 6416 taught by Professor Staff during the Spring '08 term at University of Central Florida.

Ask a homework question - tutors are online