Abstract Ridge regression is a biased estimate regression method which analysis collinearity data, it is essentially a modified least squares estimation method, by giving up the unbiasedness of the least squares, with the cost of losing information sections and accuracy to get a more practical and reliable regression coefficients。 It’s a better way to fit pathological data the the least squares。
The resonance line in AAS are not strictly geometric, it’s more like a narrow range of frequency and wavelength, which means it has width。 The central wavelength and half width are the characterization of the contour of spectrum。 The specific absorption spectrum varies from different elements, it’s rare that two elements’ absorption spectrum can overlap, when they do, the absorption will overlay and make the results high。 This study is specially designed to choose two elements whose absorption spectrum are extremely close: Cobalt and Indium, the peak of their absorption spectrum is 252。136nm and 252。137nm。 the purpose of this study is to research the influence between two elements in AAS with effects from the external conditions, and by using ridge regression to correct the results, which turned out to be between 95% to 105%。
On the other hand, for researching the influence of pH in AAS, this study designed several aqueous that have same concentration of measured elements but different pH。 Which turned out to be not quite relevant between pH and AAS。
Key words: ridge regression; atomic absorption spectrometry; AAS instrument; Cobalt; Indium。