Baseline drift in chromatography data often results in degraded quantitative analysis results. Baseline correction is required in order to achieve higher accuracy. It is difficult, however, to correct the baseline when its signal-to-noise ratio (SNR) ...
Baseline drift in chromatography data often results in degraded quantitative analysis results. Baseline correction is required in order to achieve higher accuracy. It is difficult, however, to correct the baseline when its signal-to-noise ratio (SNR) is low or when the baseline drift occurs. This study proposes an algorithm to identify the correct baseline using the moving standard deviation and penalized least squares(PLS). The proposed method can be applied to noisy chromatographic data or even to any non-stationary data. This method calculates the moving standard deviation of data and identifies the points above the user-defined threshold. The initial weight vector of PLS for the corresponding points is initialized with the value of zero. The accurate baseline can be determined through the PLS method in an iterative manner. The performance of this method can be evaluated by comparing its root-mean-square error (RMSE) with other previous methods.