Skip to main content

Table 2 The prediction accuracy of PCA-LDA models

From: Detection and imaging of lipids of Scenedesmus obliquus based on confocal Raman microspectroscopy

Time Modeling schemesa
F P L PL
Modeling (%) Prediction (%) Modeling (%) Prediction (%) Modeling (%) Prediction (%) Modeling (%) Prediction (%)
d1 100 100 100 100 80 40 100 100
d3 100 100 100 100 100 40 100 100
d5 100 100 100 80 90 100 100 100
d7 100 100 100 100 80 60 100 80
d9 100 80 100 80 60 60 100 60
Total 100 96 100 92 82 60 100 88
  1. aThe modeling scheme is defined based on the Raman characteristic bands used to build the model: F—full band modeling; P—characteristic band of pigment (β-carotene, chlorophyll a) modeling (965, 1007, 1156, 1189, 1526 cm−1); L—characteristic band of lipid modeling (1266, 1302, 1445, 1656, 1742 cm−1); PL—partial characteristic band of pigment (1007, 1156, 1526 cm−1) and partial characteristic band of lipid (1266, 1445 cm−1) modeling