Nonparametric Preference Completion
This addresses the problem of personalized recommendation systems for users, but it is incremental as it extends nonparametric methods to preference completion with a consistency proof.
The paper tackles collaborative preference completion by proposing a nonparametric k-nearest neighbors-like algorithm to recover personalized rankings from partially observed ratings, proving its consistency and demonstrating performance on Netflix and Movielens datasets.
We consider the task of collaborative preference completion: given a pool of items, a pool of users and a partially observed item-user rating matrix, the goal is to recover the \emph{personalized ranking} of each user over all of the items. Our approach is nonparametric: we assume that each item $i$ and each user $u$ have unobserved features $x_i$ and $y_u$, and that the associated rating is given by $g_u(f(x_i,y_u))$ where $f$ is Lipschitz and $g_u$ is a monotonic transformation that depends on the user. We propose a $k$-nearest neighbors-like algorithm and prove that it is consistent. To the best of our knowledge, this is the first consistency result for the collaborative preference completion problem in a nonparametric setting. Finally, we demonstrate the performance of our algorithm with experiments on the Netflix and Movielens datasets.