返回
Smart sampling and probing
DOI:10.1016/j.chemolab.2021.104306.png)
摘要
En 中文
Sampling (collecting samples) and probing (measuring variables) are crucial when accessing key information from any system that must be characterized and understood. In this process, the tendency is to collect as many samples or signals as possible and to obtain information using all available methods. To avoid excessive, usually expensive and time-consuming analytical effort, in this work, we describe how to use simple unsupervised multivariate analysis tools (PCA and POA) to estimate the number of objects and factors in a given multivariate dataset. With this information, one may easily determine if the numbers of analyzed samples and variables are satisfactory in terms of whether key information has been extracted from the investigated system. By simulating several analytical possibilities, we tested the presence of different levels of random noise, sample restrictions and variable restrictions to detect the information contained in the system. After this validation process, the same methodologies were applied to verify different available datasets selected from previously published research papers (a total of 41 datasets) and draw a conclusion about sampling and probing. Finally, some useful ?rules of thumb? were developed to simplify the analysis of complex multivariate systems, especially for the characterization of complex environmental cases.
Keyword:
Sampling
Probing
Multivariate analysis
PCA
POA
Component analysis
Object analysis
AI总结
对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
3.8
论文数:
4.6K
被引数:
1.2W
机构
引用论文
Farm Biosecurity and Influenza A virus detection in Swine Farms: A Comprehensive Study in Colombia农场生物安全与猪场A型流感病毒检测:哥伦比亚综合研究
Surface Enrichment in Equimolar Mixtures of Non‐Functionalized and Functionalized Imidazolium‐Based Ionic Liquids
ChemPhysChem
IF0
A global comparison of the nutritive values of forage plants grown in contrasting environments不同环境中生长的牧草植物营养价值的全球比较


