[Back]


Scientific Reports:

P. Filzmoser, J. Gussenbauer, M. Templ:
"Detecting outliers in household consumption survey data";
Report No. Deliverable 4. Contract with world bank (1157976), 2015; 85 pages.



English abstract:
The World Bank has standardized a large collection of about 200 survey datasets from 90 developing countries. These datasets are standard in the sense that variable names and the structure of the datasets have been harmonized, in such a way that a (Stata, R or other) program developed for one particular dataset will run on all. These datasets are used to conduct a
large-scale assessment of the extent of outliers issue, to test detection and imputation methods and identify the best-performing options.

Keywords:
outlier detection, multivariate statistics, robust statistics

Created from the Publication Database of the Vienna University of Technology.