akmedoids

Advances a novel adaptation of longitudinal k-means clustering technique (Genolini et al. (2015) doi:10.18637/jss.v065.i04) for grouping trajectories based on the similarities of their long-term trends and determines the optimal solution based on either the average silhouette width (Rousseeuw P. J. 1987) or the Calinski-Harabatz criterion (Calinski and Harabatz (1974) doi:10.1080/03610927408827101). Includes functions to extract descriptive statistics and generate a visualisation of the resulting groups, drawing methods from the 'ggplot2' library (Wickham H. (2016) doi:10.1007/978-3-319-24277-4). The package also includes a number of other useful functions for exploring and manipulating longitudinal data prior to the clustering process.