sits_patterns: Create time series patterns for classification

In luizassis/sits: Satellite Image Time Series Analysis

Description

This function allows the user to select different alternatives to define a set of patterns, given his samples. The alternatives are: "gam" - uses a generalised additive model to approximate a smooth spline for each pattern "dendogram" - uses a herarchical clustering method to group the patterns "centroids" - uses a positional clustering method to group the patterns

Usage

sits_patterns(data.tb, bands = NULL, pt_method = sits_gam(data.tb = NULL,
  bands = bands, from = NULL, to = NULL, freq = 8, formula = y ~ s(x)))

Arguments

data.tb	a SITS tibble time series with an alignment column
bands	the bands to be used for determining patterns
pt_method	a pattern fitting method

Value

result a model fitted into input data given by train_method parameter

Author(s)

Rolf Simoes, [e-mail]

Gilberto Camara, [e-mail]

luizassis/sits documentation built on May 30, 2019, 7:15 p.m.