Consider a set $U \subset \mathbb{R}^2$ to be star-shaped with respect to a point $\epsilon \in \mathbb{R}^2$ if for every $x \in U$, the segment joining $\epsilon$ and $x$ is contained in $U$.
Consider the $H = \{U \subset \mathbb{R}^2 | U \text{ is starshaped with respect to } \epsilon, \epsilon \in \mathbb{R}^2\}$ And take
$H_{+}$ to be all the hypotheses in $H$ that assign $1$ if $x$ is within the starshaped set and $0$ otherwise. I am interested in finding the VC dimension of $H$. I have a feeling that this can account for an infinite number of points on $\mathbb{R}^2$, but I am having trouble formalizing this proof. Many of my friends said that these problems are best answered using "matrix" methods over all dichotomies of my observations. Is there a better method for handling this?