add a function, detect_outliers() to detect multivariate outliers based on Hadi’s (1994) procedure (thanks to Rebecca Steingut for this contribution)
create_profiles() functionadd plot_raw_data and plot_centered_data as arguments (that can be specified as TRUE but default to FALSE) create_profiles() to change plot of profile centroids (thanks again to Rebecca for input that led to making this addition)
add of a new function, cross_validate(), to perform double split-half cross-validation
make minor changes to how centering and scaling of data is carried out
change create_profiles(), to now return a .data slot, so the original data can be used for subsequent analyses
Fixed error in create_profiles() that returned only the cluster assignment for .data, rather than the original data.frame with the addition of the cluster assignment
Added function plot_r_squared() to plot R^2 (r-squared) values for a range of number of profiles and updated vignette to include use of plot_r_squared() function
Added new interface for main (create_profiles()) function so that variables to create profiles must be specified, rather than every variable in the data frame being used to create profiles being used. This change means that the data.frame does not need to be subset before using this package, and also that it is easier to use cluster assignments in subsequent analyses because the original data frame with either a variable for the cluster assignments or dummy-coded variables for each cluster are returned.
Changed n_clusters argument in create_profiles() to n_profiles
Changed axes for plot associated with plot method, so clusters are on the x-axis and variables are on the y-axis
Used R version 3.4
Add URL and BugReports fields to DESCRIPTION
Added a NEWS.md file to track changes to the package.
Fix r-squared, which was calculated using the sum of the within-cluster sum of squares divided by the total sum of squares, rather than the between-cluster sum of squares divided by the total sum of squares
Fix plot, which was reversed