Journal of Computational Science

Abstract: This paper uses panel data from 30 provinces and autonomous regions from 2004-2010 to theoretically analyze and empirically test the relationship between the real estate industry and the independent innovation activities of enterprises in China. This paper employs the amount of patient authorization and sales of new products as dependent variables, house price and investment in real estate as independent variables, R&D spending, FDI, average consumption, fixed investment as control variables. The research suggests that (1) there exist a negative relationship between house price and the independent innovation output of enterprises (the amount of patient authorization) in China. And at the same time for the upcoming increase of house price has a remarkable influence on the sales of new products. (2) The relationship between the real estate investment, independent innovation output of enterprises, and sales of innovative products is positive. However, this positive relationship is incredible. Therefore, controls of the excessive increase of house price and regulation of the real estate investment play a decisive role in the independent innovation activities of enterprises in China.

Abstract: Dyes have been applied and are playing increasingly important role in many industries including textile, printing, medical, and energy industry etc. The wide applications of them imply that specific dyes possessing given properties should be effectively designed. In present study, a QSPR modeling work performed on the maximum absorption of 1, 4-naphthoquinone derivatives was investigated using the classical 2D-QSPR method by linear multiple linear regression (MLR) model and nonlinear radial basis function neural network (RBFNN). Both of the models were verified by internal validation and external validation. The RBFNN model gives better results: N=40,R²=0.9524, LOOq²=0.8815, F=578.758, RMS=17.4988 for training set, and R²=0.9231, q²_ext =0.9202, F=84.010, RMS=25.3406 for the external test set. While the MLR model identify the most appropriate descriptors and gives a simple and intuitionistic model whose statistic parameter are: R²=0.8887, LOOq²=0.8627, F=231.552, and RMS=28.6358 for training set; R²=0.8622, q²_ext=0.8308, F=423.809, and RMS=33.8613 for the external test set. The appropriate results proved the models to be meaningful and useful to predict the maximum absorption of 1, 4-naphthoquinone derivatives.