calculate_best_config: Function that returns for each candidate C the best number of...

In RCTS: Clustering Time Series While Resisting Outliers

Function that returns for each candidate C the best number of groups and factors, based on the PIC.

Description

Function that returns for each candidate C the best number of groups and factors, based on the PIC.

Usage

calculate_best_config(df_results, df_pic, C_candidates, limit_est_groups = 20)

Arguments

df_results	dataframe with results for each estimated configuration
df_pic	dataframe with the PIC for each configuration and for each candidate C
C_candidates	candidates for C (parameter in PIC)
limit_est_groups	maximum allowed number of groups that can be estimated

Value

Returns a matrix with a row for each candidate value for C. The first column contains the optimized number of groups (for each candidate C). The second columns does the same for the number of common factors. Column 3 until 22 do the same for the number of group specific factors. This is set to NA if the configuration has less than 20 groups estimated.

Examples

df_results <- add_configuration(initialise_df_results(TRUE),
  3, 0, c(3, 3, 3, rep(NA, 17))) #data.frame with one configuration
calculate_best_config(df_results, data.frame(t(1:5)), 1:5)

RCTS documentation built on May 31, 2023, 9:20 p.m.