我刚刚读了一篇关于 different ranking algorithms 的文章.我有点困惑的一件事是 stumbleupon 的算法:
(Initial stumbler audience / # domain) + ((% stumbler audience / # domain) + organic bonus – nonfriend) – (% stumbler audience + organic bonus) + N
N is a "safety variable" so that the assumed algorithm is flexible. It represents a random number.
有人可以向我解释一下 N 的用途吗?我不明白它如何使算法更加灵活。
最佳答案
它有助于以不可预测的速度降低结果。如果您曾经有过几个小时的跌跌撞撞,您可能会注意到,当您开始摸清桶底时,每次跌跌撞撞的质量和相关性都会严重下降。
如果没有随机变量,结果将以非常可预测的速度下降。您最有可能喜欢的链接将排在第一位,随后的每个链接对您的吸引力可能会降低一些。该变量试图在某种程度上否定这一点。
stumbleupon 是众多巧妙方法之一 designed to be addictive :
The results were unanimous in the fact that when one performed an action over and over again, and was given a reward randomly, dopamine levels rose. If the reward was given consistently, i.e. every four time the action was performed, the dopamine levels remained constant
关于algorithm - 什么是 "safety variable"?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/1759756/