Solution

ID: past-exam-of-the-computer-science-course-of-the-university-of-oxford/2015/computational-learning-theory/1/c/solution

Choose distinct points of . The vectors lie in the -dimensional vector space , so they are linearly dependent. Hence there are real numbers , not all zero, such that
The second equality forces both a positive and a negative coefficient. Put , , and
Rearranging the first equality gives
Each side is a convex combination: the coefficients are positive and sum to one. Thus belongs to both convex hulls. Place the points with indices in in , those with indices in in , and assign every zero-coefficient point and every remaining point of to . Then are nonempty, disjoint, and have union . Adding points cannot shrink a convex hull, so
This proves the intersecting convex hull partition theorem, including when has more than points or is infinite: only the initial finite collection was needed.

New to topics? Read the docs here!