Decomposition of overlapping peaks in gas chromatograph
This is an article about Decomposition of overlapping peaks in gas chromatograph. If you are interested, please contact us!
The high dimensionality and high degree of nonlinearity of the gas chromatograph Shimadzu vials overlap peak decomposition problem make the calculation take a long time and is generally only suitable for post-processing. The high-dimensional problem not only makes the algorithm very slow, but also has a great influence on the convergence of the algorithm.
This is because there are often local minimum points in the objective function, and the parameter optimization process often falls into the local minimum point and cannot be self-extracted, and cannot converge to the global minimum point. Shimadzu vialsThe more the number of peaks in the gas chromatograph, the more model parameters, the more local small points in the objective function. The gradient descent algorithm searches for the direction of the improved solution each time. This "greedy" algorithm often leads to finding only one local optimal solution, rather than the global optimal solution. Shimadzu vials An artificial neural network with optimization functions can solve the problem of local minimum points.
This is the end of the introduction of Decomposition of overlapping peaks in gas chromatograph. I hope it can help you.