Peak-hopping stepwise wavelength selection algorithm

Wavelength selection is an important preprocessing step for improving and simplifying calibrations in both quantitative and qualitative problems. An improved variable selection algorithm has been developed to improve upon existing methods in terms of speed and prediction error. The new technique uses a novel peak-hopping strategy to move quickly between important spectral regions. Results when applied to both Raman and near-infrared data show that the algorithm is very fast, decreases prediction errors, and chooses a small subset of the full spectral range available. A comparison with other techniques is given and the respective advantages and disadvantages are discussed.