A new iterative algorithm is proposed to reconstruct an unknown sparse signal from a set of projected measurements. Unlike existing greedy pursuit methods which only consider the atoms having the highest correlation w...A new iterative algorithm is proposed to reconstruct an unknown sparse signal from a set of projected measurements. Unlike existing greedy pursuit methods which only consider the atoms having the highest correlation with the residual signal, the proposed algorithm not only considers the higher correlation atoms but also reserves the lower correlation atoms with the residual signal. In the lower correlation atoms, only a few are correct which usually impact the reconstructive performance and decide the reconstruction dynamic range of greedy pursuit methods. The others are redundant. In order to avoid redundant atoms impacting the reconstructive accuracy, the Bayesian pursuit algorithm is used to eliminate them. Simulation results show that the proposed algorithm can improve the reconstructive dynamic range and the reconstructive accuracy. Furthermore, better noise immunity compared with the existing greedy pursuit methods can be obtained.展开更多
Considering atomic property vector and atomic correlative function, the 3-dimensional structural vector of atomic property correlation （3D-VAPC）, a novel descriptor,is defined to characterize a 3-dimensional molecul...Considering atomic property vector and atomic correlative function, the 3-dimensional structural vector of atomic property correlation （3D-VAPC）, a novel descriptor,is defined to characterize a 3-dimensional molecular structure by introducing self-adaptability regulation mechanism and the idea of orientating to customers. Characterizing the structures of 25 bisphenol A compounds by this vector, the QSAR models of three kinds of estrogen activities （ER affinities, gene induction and cell proliferation） have high multiple correlation coefficient （Rcum^2=0.933, 0.813, 0.959） and cross verification coefficient （Qcum^2=0.847, 0.953, 0.798） by support vector machine （SVM）, which suits for nonlinear circumstances. The above results show that the models successfully express the correlation between structure and three kinds of estrogen activities. Therefore, 3D-VAPC exactly reflects the molecular structural information and SVM method correctly describes the correlation between information and property of the compounds.展开更多
基金supported by the National Natural Science Foundation of China (61172138)the Specialized Research Fund for the Doctoral Program of Higher Education (200807011007)the Natural Science Basic Research Program in Shannxi Province of China (2013JQ8040)
文摘A new iterative algorithm is proposed to reconstruct an unknown sparse signal from a set of projected measurements. Unlike existing greedy pursuit methods which only consider the atoms having the highest correlation with the residual signal, the proposed algorithm not only considers the higher correlation atoms but also reserves the lower correlation atoms with the residual signal. In the lower correlation atoms, only a few are correct which usually impact the reconstructive performance and decide the reconstruction dynamic range of greedy pursuit methods. The others are redundant. In order to avoid redundant atoms impacting the reconstructive accuracy, the Bayesian pursuit algorithm is used to eliminate them. Simulation results show that the proposed algorithm can improve the reconstructive dynamic range and the reconstructive accuracy. Furthermore, better noise immunity compared with the existing greedy pursuit methods can be obtained.
基金This work was supported by the Natural Science Foundation of CQ CSTC (No. 2006BB5177)
文摘Considering atomic property vector and atomic correlative function, the 3-dimensional structural vector of atomic property correlation （3D-VAPC）, a novel descriptor,is defined to characterize a 3-dimensional molecular structure by introducing self-adaptability regulation mechanism and the idea of orientating to customers. Characterizing the structures of 25 bisphenol A compounds by this vector, the QSAR models of three kinds of estrogen activities （ER affinities, gene induction and cell proliferation） have high multiple correlation coefficient （Rcum^2=0.933, 0.813, 0.959） and cross verification coefficient （Qcum^2=0.847, 0.953, 0.798） by support vector machine （SVM）, which suits for nonlinear circumstances. The above results show that the models successfully express the correlation between structure and three kinds of estrogen activities. Therefore, 3D-VAPC exactly reflects the molecular structural information and SVM method correctly describes the correlation between information and property of the compounds.
