It is well known that , the brownian paths in n - space have no double points with probability 1 if n 4 , butforn 3 , there are double points with probability 1 眾所周知,對(duì)n維brown運(yùn)動(dòng),當(dāng)n 4時(shí),幾乎所有的樣本軌道不存在二重點(diǎn),而當(dāng)n 3時(shí),幾乎所有的樣本軌道存在二重點(diǎn)。
Is that if a set of points in n - space is cut by a hyperplane , then the application of the perceptron training algorithm will eventually result in a weight distribution that defines a tlu whose hyperplane makes the wanted cut )下的結(jié)論是,如果n維空間的點(diǎn)集被超平面切割,那么感知器的培訓(xùn)算法的應(yīng)用將會(huì)最終導(dǎo)致權(quán)系數(shù)的分配,從而定義了一個(gè)tlu ,它的超平面會(huì)進(jìn)行需要的分割。