These instances, whenmapped to an N-dimensional space, represent a core set that can be used to construct an approximation to theminimumenclosing ball. Solving the SVMlearning problem on these core sets can produce a

支持向量机（svm）的学习资料和响应的代码学习。(Support vector machine (svm) learning data and some code learning.)