上海交通大学学报

上海交通大学学报（自然版）

• 化学工程 • 上一篇    

批量甜玉米低场核磁共振数据的统计分析

邵小龙a,b，朱将伟b，李云飞b
  

  1. (上海交通大学a. 制冷与低温工程研究所； b. 农业与生物学院， 上海 200240)
  • 收稿日期:2010-01-07 修回日期:1900-01-01 出版日期:2011-01-27 发布日期:2011-01-27

Statistical Analysis for LowField Nuclear Magnetic Resonance Batch Data of Sweet Corn

SHAO Xiaolonga,b,ZHU Jiangweib,LI Yunfeib
  

  1. (a. Institute of Refrigeration and Cryogenic Engineering; b. School of Agriculture and Biology, Shanghai Jiaotong University, Shanghai 200240, China )
  • Received:2010-01-07 Revised:1900-01-01 Online:2011-01-27 Published:2011-01-27

摘要： 以烫漂甜玉米的低场核磁共振数据为例，采用统计分析系统（SAS）得到烫漂温度对甜玉米中水分分布的影响规律，利用SAS批量读入数据并对数据进行多指数模型拟合、主成分分析和偏最小二乘法模型预测，提供了相应的SAS代码.结果表明，当弛豫时间为450~750和50~70 ms时，相应水组分的弛豫强度分数随处理温度的变化而呈现出一定的变化规律；烫漂温度可初步划分为3个温度段，即20~40、50~70和80~100 °C；束缚水含量模型具有较高的预测准确性（决定系数R2=0.974，标准差RMSECV=0.32%）；SAS方法在批量数据处理过程中具有较高分析和处理数据的能力.

关键词: 统计分析系统, 批量数据, 指数拟合, 主成分分析, 偏最小二乘回归法

Abstract: To obtain the effect and quantitative information of water components in sweet corn by different blanching temperature, Statistical analysis system (SAS) was applied to deal with lowfield nuclear magnetic resonance (LFNMR) data of blanched sweet corn. Statistic analysis on batch raw data of LFNMR was performed by SAS system, including exponential fitting, principal component analysis (PCA) and partial least squares regression (PLSR). The corresponding SAS codes were provided. The fitting result of multiexponential model indicates that the percentages of two components with relaxation times (405-750 ms) and (50-70 ms) change distinctly. Three blanching temperature ranges: 20-40, 50-70 and 80-100 °C are roughly discriminated by PCA. PLSR does well in prediction of bound water in blanched sweet corn (determined coefficient is 0.974, root mean square error of crossvalidation is 0.32%). From the whole data processing, SAS programming performs efficiently on data management and analysis and gives valuable reference for LNNMR application.

中图分类号: