Cell cycle phase classification in 3D in vivo microscopy of Drosophila embryogenesis

BMC Bioinformatics

Table 1 Comparison of cell cycle phase classification accuracy obtained with different classification models (columns) and feature reduction techniques (rows).

Accuracy	SVM	PNN	KNN	BPNN
Original features	93.52±0.62%	91.67±0.69%	90.18±0.56%	89.97±0.65%
PCA	92.45±0.73%	90.12±0.67%	90.02±0.54%	89.82±0.64%
LDA	93.12±0.45%	89.94±0.70%	89.12±0.56%	88.54±0.54%
MDS	93.23±0.44%	91.12±0.56%	91.34±0.65%	90.22±0.65%

The training set consisted of 3119 samples in 5 cell cycle phases (see breakdown in Table 2) of the post-cellular blastoderm (gastrulation). Training was performed using 10 fold cross validation. SVM=support vector machine, PNN= probabilistic neural network, KNN=k nearest neighbour, BPNN=neural network with back propagation, PCA=principal component analysis, LDA=linear discriminant analysis, MDS=multidimensional scaling.

ISSN: 1471-2105