矩阵奇异值分解及其在高维数据处理中的应用

矩阵奇异值分解能够实现对高维数据的局部特征提取及维数约减,在智能信息处理和模式识别研究领域具有十分重要的应用价值.首先分析了高维数据处理所面临的困境,并对常用的降维算法进行简单的归纳总结;然后阐述了矩阵奇异值分解的基本原理及其在维数约减和数据压缩中的物理意义;接着通过分析两种建立在奇异值分解基础上的PCA与LSA降维算法的数学导出过程,进一步给出了两者的等价性证明;最后总结了矩阵奇异值分解的优缺点,并且预测了高维数据处理技术未来的发展趋势.     

Activation/Inhibition Effects during the Coelectrodeposition of PtAg Nanoparticles: Application for ORR in Alkaline Media

PtAg bimetallic nanoparticles for oxygen reduction reaction (ORR) in alkaline media were prepared by pulse electrodeposition (PED). During PED the reduction of Ag⁺ ions predominates, thus an increased Ag content in the co‐deposit is accomplished. The mechanism for this anomalous co‐deposition was elucidated by potential pulse experiments, which revealed that nuclei formation mainly occurs via the reduction of Pt²⁺ ions. The growth of the particles is diffusion controlled leading to the formation of a Ag shell covering a PtAg alloyed region. However, the shell is not growing homogeneously on the PtAg alloy. Hence, regions of the PtAg alloy are exposed, which exhibit an enhanced ORR activity compared to a pure Ag surface.