Version 2 / April 18, 2012 / Benjamin Aaron Degenhart
Create all possible subgroups (n^2) for each sequence (length n) by counting up to n^2 in binary numbers (e.g. 101001 = 41) and adding a char at the place of each '1' (e.g. the binary 41 gives the subgroup 'acf' from the sequence 'abcdef'). Count occurrence of each subgroup. Output a list of subgroups ranked by [length * occurrence] across all sequences.
A group has to split into subgroups and has to solve the grouping 'democratically'; ask everyone to give their subgroup-preference and run this algorithm over their sequences. The result will show the social constellation that the system gravitates towards.
A brainstorming-session has generated lots of items and it's now necessary to cluster them in categories. From all possible categorization schemes its often the first applied one that dominates in a social setting, not necessarily the best one. Applying this algorithm you can ask every individual to cluster to the best of their abilities on their own and then look which items gravitate towards each other across all categorization schemes. Same counts for the structure of a website too for instance.
A yoga teacher has 1000 positions to draw from when he assembles the sequence of 10 positions in todays session. We could record his selection over a year and analyze the constellational gravity within the positions. Of course the 'arc' (narrative) in each session is not getting analyzed with this method.
Music playlists, ingredients in food, elements across CVs, narrative elements in stories, any monochronic method/procedures that draws from a shared pools of entities and arise as selection of those ...
input:
('abc' & 'cdae' will not share 'ac') ('abc' & 'cdae' will share 'ac')
('abc' & 'acbd' will not share 'ab') ('abc' & 'acbd' will share 'ab')