1: \begin{abstract}

2:   We analyze convergence of the weighted nonlocal Laplacian (WNLL) on high dimensional randomly distributed data. %random distributed data.

3: The analysis reveals the importance of the scaling weight $\mu \sim P|/|S|$ with $|P|$ and $|S|$ be the number of entire and labeled data, respectively. The result gives theoretical foundation of WNLL for high dimensional data interpolation.

4: \end{abstract}

5: