问一个题目# JobHunting - 待字闺中w*y2012-03-02 08:031 楼从数据流里sample k 个documents, 假定现在是有N个document, 怎么做到uniformlyrandom sampling
r*n2012-03-02 08:032 楼这个貌似是我的onsite题...【在 w***y 的大作中提到】: 从数据流里sample k 个documents, 假定现在是有N个document, 怎么做到uniformly: random sampling
l*a2012-03-02 08:033 楼search reservoir sampling【在 w***y 的大作中提到】: 从数据流里sample k 个documents, 假定现在是有N个document, 怎么做到uniformly: random sampling
r*n2012-03-02 08:034 楼其实就是个learning rate为1/n的stochastic gradient descent【在 l*****a 的大作中提到】: search reservoir sampling
l*a2012-03-02 08:036 楼你用的这几个词太专业了基本都不认识【在 r*******n 的大作中提到】: 其实就是个learning rate为1/n的stochastic gradient descent
h*e2012-03-02 08:0310 楼存下前k个。。try to use the incoming ones to replace with a correctprobability.. that's storing and sampling..