Robust spectral analysis by moving block bootstrap designs Original Research Article

A moving block bootstrap (MBB) technique for statistical analysis of data with correlated residuals is introduced. The MBB does not require independently and identically distributed (i.i.d.) residuals as is a fundamental requirement for classical least-square regression approaches. Therefore, the MBB is robust against residuals correlation. An MBB design evaluates probability distribution functions and therefore opens access to non-parametric statistical analysis of the data. It does not require sum-of-squared residuals as fitting criteria and, hence, allows data analysis that is also robust against extraneous data. The MBB will be discussed for UV–VIS spectroscopic data, where the correlation in the residuals is demonstrated by autoregression analysis. The performance of the new technique is compared to linear least-squares regression.