Identifying the Most Influential Data Objects with Reverse Top-k Queries