Application of PCA-RBF Method to QSRR Studies on Hydrocarbon in FCC Gasoline with PCA-RBF
A novel Artificial Neutral Network (ANN) algorithm based on Principle component analysis(PCA) is proposed and applied to predicted the retention index of a series of hydrocarbons in FCC gasoline. The PCA technology is utilized to preprocess the mass spectrogram of FCC gasoline for parameter selection and to reduce the input of prediction model, which thus improve the input factors and eliminates the correlation among the inputs. Then new sample data are input into radial basis function-artificial neutral network(RBF-ANN) to construct the prediction model.156 compounds were divided into two subsets. RSM of training set and testing set is 0.9958 and 0.9991, respectively. The satisfactory results were obtained.
Weiguo Pan, Jianxing Ren and Yongguang Li
X. T. Zhang et al., "Application of PCA-RBF Method to QSRR Studies on Hydrocarbon in FCC Gasoline with PCA-RBF", Advanced Materials Research, Vols. 347-353, pp. 1769-1773, 2012