Services on Demand
Journal
Article
Indicators
Cited by SciELO
Access statistics
Related links
Similars in
SciELO
Share
Revista Técnica de la Facultad de Ingeniería Universidad del Zulia
Print version ISSN 0254-0770
Abstract
SOCORRO MAC-QUHAE, Hely Saúl and BERMUDEZ, Mauricio A.. Proposal of a methodology to calculate propagation of a variable statistical error (VEE): applications to well log data and thermocronologic dating. Rev. Téc. Ing. Univ. Zulia [online]. 2017, vol.40, n.1, pp.26-33. ISSN 0254-0770.
The purpose of this research on Basic Statistics is to propose an alternative method to calculate errors based on the data natural variability. Because of the world wide application of the Standard Deviation formula, proposed methodology is defined in a similar manner. This methodology consists in defining an Error Function corresponding to the regression function with maximum Pearson correlation factor of deviation magnitudes, which are defined in relation to a generic model, arithmetic average in this case. Accordingly, firstly deviations are calculated, secondly negative signs are discarded by means of absolute value function, thirdly, to propagate error over the statistical set of samples, regression function with maximum correlation factor is taken as “Error function” and, finally, errors are calculated for each sample. This way, statistical error varies as inverse function to representativeness, more representative samples/observations, which are closer to average have a relative smaller error, less representative samples/observations are associated to greater errors. Proposed methodology is applied to well log and thermocronologic data, the latter experimentally obtained in the laboratory.
Keywords : Error function; Pearson correlation factor; standard deviation; representativeness; regression.












