Skip to main content
Figure 4 | Biology Direct

Figure 4

From: Component retention in principal component analysis with application to cDNA microarray data

Figure 4

Predicted dimension for simulated data. Predicted dimension versus noise Level. Shown is the information dimension (line with circle markers), Jolliffe's modification of the Guttman-Kaiser rule (line with star markers) and LEV (line with square markers) plotted against noise level, measured in standard deviations. Note that the predictions given by both the information dimension and Guttman-Kaiser's rule increase as the noise level increases, while LEV drops sharply (see Text).

Back to article page