You are here

Article

UDC: 519.246
DOI: 10.14438/gn.2015.08
typology: 1.04
submitted: 2015-03-01, accepted: 2014-04-06, published: 2014-04-10


Normality distribution testing for leveling data obtained from production measurements

Milan Trifković1*

1University in Novi Sad, Faculty of Civil Engineering in Subotica, Serbia

Abstract. Normal distribution of data is of crucial importance in data processing and hypothesis testing in geodesy. Models of geodetic measurements adjustment assume that data are normally distributed. However, results of measurements could be affected by different influences because geodetic data are obtained under external conditions and under different limitations such as the deadlines or other processes on the construction site. These possibilities implicate certain risks that deviations from normal distribution in geodetic data could appear. Those deviations from normal distribution could spread through the mathematical and stochastic models and violate conclusions based on geodetic data. To avoid mentioned risks it is of considerable importance to devote attention to testing normality distribution of geodetic data obtained from production measurements. In this paper one set of leveling data obtained from production measurements was considered from aspect of its normal distribution curve.

Keywords. Shapiro-Wilk test, Pearson Chi-square test, skewness, kurtosis

Download the full article

BibTeX reference of the article