Abstract—In chromatogram analysis, overlapped chromatograms are difficult to analyze if they are not resolved. Experimental separation procedures are costlier and time consuming. Conventional multivariate curve resolution techniques have difficulties when the chromatograms are severely overlapped. Non-negative Matrix Factorization (NMF) is a tool generally used for image processing and data mining. The original NMF can also be applied for chemical analysis, after imposing some constraints. In this work, modified NMF with divergence objective algorithm (NMFdiv) has been proposed for the separation of severely overlapped chromatograms of acetone and acrolein mixture. The experimental chromatograms were obtained using Gas Chromatography –Flame Ionization Detector (GC-FID).The proposed NMFdiv algorithm effectively separates the overlapping chromatograms. Principal Component Analysis (PCA) was applied to determine the number of components in the mixtures taken. The correlation co-efficient is greater than 0.99 and resolved signal power is greater than 20dB always. The algorithm also effectively separates the simulated overlapped chromatograms even though the numbers of overlapped components are three or four. The simulated results are encouraging even for three and four components and the correlation coefficient is greater than 0.9 and resolved signal power is greater than 10 dB always.

Index Terms—Modified NMFdiv, Overlapping chromategrams,PCA, resolution.

S.Anbumalar is with Sri Manakula Vinayagar Engineering College, Puducherry, India . e-mail: saravanan.anbumalar@gmail.com
P.Rameshbabu and R.Anandanatarajan are with Pondicherry Engineering College, Puducherry, India

Cite: S.Anbumalar, P.Rameshbabu, and R.Anandanatarajan, "Overlapped Chromatograms Separation Using Nonnegative Matrix Factorization," International Journal of Computer and Electrical Engineering vol. 3, no. 5, pp. 654-658, 2011.