中文说明:应用背景当评估一个和EOUT由此产生的分类二,使用二进制分类错误阳离子。实际发言:(我)为该作业的目的,没有规模的数据时,你使用libsvm或其他的包,以免改变E 有效核和迪不同结果。(Ⅱ)在某些包中,您需要指定双精度。(三)在10倍交叉验证,如果数据大小是不多的10,大小的10亚群可能是O 1数据点。关键技术考虑多项式核函数n;XM)=(1 + xTnXM)Q,在哪里问的程度多项式。
English Description:
Application backgroundWhen evaluating Ein and Eout of the resulting classier, use binary classication error. Practical remarks: (i) For the purpose of this homework, do not scale the data when you use libsvm or other packages, lest you should change the eective kernel and get dierent results. (ii) In some packages, you need to specify double precision. (iii) In 10-fold cross validation, if the data size is not a multiple of 10, the sizes of the 10 subsets may be o by 1 data point.Key TechnologyConsider the polynomial kernel K(x n ; xm) = (1 + x T n xm) Q , where Q is the degree of the polynomial.