%0 Journal Article %A Millogo, Die Joseph Hassan %A Chan, Kuei-Yuan %D 2019 %I Begell House %K data analysis, uncertainty identification, curve fitting, principal component analysis, extrapolation %N 6 %P 569-587 %R 10.1615/Int.J.UncertaintyQuantification.2019028125 %T MULTIVARIATE ANALYSIS OF EXTRAPOLATING TIME-INVARIANT DATA WITH UNCERTAINTY %U https://www.dl.begellhouse.com/journals/52034eb04b657aea,65c85ecf41e4a59b,02b761143194570f.html %V 9 %X Data analysis deciphers phenomena and system behaviors within a large number of experimental realizations. Transforming these massive quantities of raw data into knowledge about the data is made possible thanks to continuously improved computing techniques. In science and engineering, a particular interest lies within surrogate models for system behaviour prediction and data extrapolation. These models could, however, be under- or over- fitted when confronted to a complex dataset or one embedded with uncertainty. In this paper, we suggest a treatment approach of experimental data under uncertainty prior to its surrogate model creation. We specially focus on extrapolation an attempt to estimate the true underlying phenomena. We quantify the uncertainty quantity through eigenvalues, copy the behavior of the data through its covariance matrix, and reproduce an almost identical dataset whose particularity is a perfectly correlated inputs and output. This new dataset is then used as the basis for the creation of a surrogate model. Our approach shows consistency and a clear opportunity to obtain better predictions under uncertainty as it focuses on the overall dataset's behavior and stays faithful to each data. %8 2019-12-17