Figure 4
From: Piecewise multivariate modelling of sequential metabolic profiling data
Modelling and prediction of the simulated data set. (A) Cumulative magnitude weights (Wdist.cum) for each sub-model. (B) Distance between time points in each sub-model (di) (C) Probabilities for one observation over all time points to belong to each sub-model (in prediction). (D) Predicted time for each observation in the test-set plotted against the true time represented as a boxplot.