|
|
|
|
|
by jvanderbot
883 days ago
|
|
The title explains literally the whole concept: Representing the "opinions" (or more appropriately dispositions) of each person as a vector, the belief system of a "tribe" (connected set of individuals) is probably the mean of these vectors. And relgions seek to optimize the "morality" they preach (vector they propose as "god") to maximize membership / agreement / centralness of their god. Which of course (author does not point this out) is the mean of the population. They could have expanded a little: * By what metric do we define "closeness"? A higher norm would penalize dissent / divergence from "god" more.
* In the limit, how does dissent penalties affect the "coverage" of available religions? This is basically k-means with different defintions of "mean"
* You could model the problem as unsupervised clustering where k is unknown to predict how many religion clusters form, for a given population - if opinions could be measured. on and on. Title was the most interesting part, the rest was just coffee kicking in. |
|
https://kitsunesoftware.wordpress.com/2019/05/25/dot-product...