This function uses package self-organized maps to find clusters in satellite image time series to cluster samples.
sits_kohonen(data.tb, time_series, bands = NULL, grid_xdim = 25,
grid_ydim = 25, rlen = 100, dist.fcts = "euclidean", alpha = 1,
neighbourhood.fct = "bubble", ...)A tibble with samples to be clustered.
Time series extracted from tibble data.
Bands to be clustered.
X dimension of the SOM grid (default = 5).
Y dimension of the SOM grid.
Number of times the complete data set will be presented to the SOM grid
The similiraty measure (distance).
Learning rate, a vector of two numbers indicating the amount of change.
Type of neighbourhood function (bubble or gaussian).
Additional parameters to be passed to kohonen::supersom function.
A tibble with the clusters time series or cluster' members time series according to return_member parameter. If return_members are FALSE, the returning tibble will contain a new collumn called `n_members` informing how many members has each cluster.
`kohonen` package (https://CRAN.R-project.org/package=kohonen)