manual for reproducing results
Disclaimer: This code is provided for research use only. The authors do not have any responsibility for the results generated by the code or the accuracy thereof. For any other use, please contact the Authors.
paper1_refactor.py
idealsearch=2 extreme point search
idealsearch=5 corner 1 search + all evaluation
idealsearch=4 corner 1 search + selective evaluation
idealsearch=6 corner 1 search + selective evaluation + Silhouette analysis
idealsearch=3 corner 2 search + selective evaluation + Silhouette analysis
How to generate results:
(1) raw process: trainy_summary2csv(resultfolder, resultconver)
input: resultfolder(where all results are stored) resultconver(where results processing outcome is stored)
(2) next step
In the folder named by resultconver (e.g. paper1_convert) move matlab script: randsum_4hv.m to this folder and run it the file 'hvcompare_sig.csv' file will be generated it is latex compatible file for the paper table 2-4
(3) figure plots: (3-1) convergence plot: hvconverge_averageplot similar as (1) provide result folder and the second input is not used output is in process_plot folder under the result folder provide problem json file p/resconvert_plot3.json, which problem to process
(3-2) final results plot in plot_for_paper as above provide which problem to solve in resconvert_plot3.json results are found in process_plot folder under the result folder
as for the init plot, uncomment the init script run as above
(4) demo figures degenerate front: plot_run in paper1_refactor.py corner demo: demo_plot2 in paper1_resconvert.py