METHOD FOR DETECTING INFORMATION IN SIGNALS - APPLICATION TO 2-DIMENSIONAL TIME-DOMAIN NMR DATA

Time domain (TD) NMR is used in industry for quality control. Like nea r-infrared (NIR) spectrometry, it has many advantages over wet chemist ry including speed, ease of use and versatility. Unlike NIR, TD-NMR ca n generate a wide range of responses depending on the particular pulse sequences used, The resulting relaxation curves may vary as a functio n of the physico-chemical properties or even the biological and geogra phical origin of the product. The curves are usually decomposed into s ums of exponentials and the relaxation parameters are then used in reg ression models to predict water content, iodine number, etc, The diver sity of possible signals is both an advantage and disadvantage for TD- NMR as it broadens the range of potential applications of the techniqu e but also complicates the development and optimisation of new analyti cal procedures. It is shown that univariate statistical techniques, su ch as analysis of variance or chi-squared, may be used to determine wh ether a signal contains any information relevant to a particular appli cation. These techniques are applied to 2D TD-NMR signals acquired for a series of traditional and 'light' spreads. Once it has been demonst rated that the signals contain relevant information, partial least-squ ares (PLS) regression is applied directly to the signals to create a p redictive model, The Durbin-Watson function is shown to be a means cha racterising the signal-to-noise ratio of the vectors calculated by PLS to select the components to be used in PLS regression.