Registration

Use functional data analysis to register different data sets with eachother.

The algorithm identifies areas of high density that are shared across all most of the data sets, then applies a warp function to align those areas of high density. This is commonly used to correct sample-to-sample variation across large data sets. This is not a multidimensional algorithm – if you apply it to multiple channels, each channel is warped independently.

Channels

The channels to apply the decomposition to.

Scale

Re-scale the data in the specified channels before fitting.

Smoothing parameters

Kernel

The kernel to use for the smoothing.

Bandwidth

The bandwidth for the kernel, controls how lumpy or smooth the kernel estimate is. Choices are:

• scott (the default)

• silverman

• A floating point number. Note that this is in scaled units, not data units.

Grid Size

The number of times to evaluate the smoothed histogram.

By

A list of metadata attributes to aggregate the data before estimating the model. For example, if the experiment has two pieces of metadata, Time and Dox, setting By to ["Time", "Dox"] will fit the model separately to each subset of the data with a unique combination of Time and Dox.