Distribution of errors based on five repeated barcode-position measurements, n = 163,810 repeated barcodes (left).
Laplace proved that the distribution of errors in large data samples from astronomical observations can be approximated by a Gaussian or normal distribution.
In this paper we address the assumptions about the distribution of errors made by voxel-based morphometry.
The non-linear relationship between response (body score) and explanatory variables (day and food treatment) warranted the application of a generalized additive model (GAM) considering Cox Proportional hazards distribution of errors.
Distribution of errors from measurement of a single barcode's position when the original and repeated measurements were separated by the entire experiment lasting 96 h n = 40,032 repeated barcodes (right).
Distribution of errors from measurement of a single barcode's position where the original and repeated measurements are separated by less than one hour (h), n = 31,506 repeated barcodes (centre).
The distribution of errors conditioned on the sample composition and city showed that estimates were accurate (average mean bias = 0.006%, maximum mean bias = 0.3%).
(A) Distribution of errors and predictions for the training pool.
(C) Distribution of errors and predictions for the validation set.
Although we have no knowledge about the distribution of errors, we assumed a normal distribution of errors to employ a probit regression.
Normal distribution of errors and variance homogeneity were checked via visual inspection and Levene's test.
