摘要
非负矩阵因子分解(non-negative matrix factorization,NMF)是对非负数据处理的一种多元统计分析方法。NMF分解结果没有“负值”,易于理解和解释,具有比较明确的物理化学意义。由于其多解的特征,文献介绍的NMF算法并不能直接用于化学混合信号解析。作者根据化学波谱的基本特征(化学波谱的平滑性、色谱的单峰性以及质谱的稀疏性)对NMF算法进行了改进,缩小了其多解范围。应用改进的NMF进行模拟HPLC-DAD型两维数据(其中色谱严重重叠和完全重叠)和苍术GC/MS实验数据解析,得到了比较理想的结果。实验表明,改进后的NMF是一个可用于复杂样品化学信号分析的化学计量学新方法。
Non-negative matrix factorization （NMF） has been proposed for multivariate data analysis, with non-negativity constraints. The resolution results of NMF are non-negative and can be easily interpreted and understood directly. Due to multiple solutions, the original algorithm of NMF is not suitable for resolving mixed chemical signals. And NMF has never been applied to resolve mixed chemical signal. It must be modified according to the characteristics of chemical signals, such as smoothness of spectra, unimodality of chromatograrns, sparseness of mass spectra, etc.. The modified NMF algorithm was used to narrow the feasible solution region for chemical signals resolution, it was found that NMF could get reasonable and acceptable results under certain experimental error. Simulated two-dimensional （high performance liquid chromatography/diode array detection） data and real gas chromatography/mass spectrometry data of cangzhu were resolved by NMF technique salisfactorilly.
出处
《分析化学》
SCIE
EI
CAS
CSCD
北大核心
2006年第U09期45-48,共4页
Chinese Journal of Analytical Chemistry
基金
山东省优秀中青年科学家奖励基金资助项目(No.2005BS10004)
