パーソナルコンピュータによる高効率ピーク分離法 : II.波形分割とデータ削減

概要

論文の詳細を見る
To speed up the curve-fitting calculation for peak-separation, efficient data division and data reduction techniques were developed on the DLS (damped least-squares) algrithm. According to those techniques, the observed curve is automatically divided to several limits of isolated overlapping peaks. After that the curve-fitting based on the DLS algorithm is carried out for each limits separately with reducing the data by equal-interval samplings. The reduction interval is changed after three iterations to reduce the calculation time, and these intervals were determined from the chi square test of fitting-errors. A visible emission and an IR absorption spectra were separated to several Gaussian and Lorentzian peaks by a personal computer in BASIC language. The calculation time was reduced less than one-third of a conventional DLS method.
社団法人日本分光学会の論文