Efficient top-k high utility itemset mining on massive data

•Current top-k HUIM algorithms only concern small and moderate transaction database.•A novel PTM algorithm is presented to mine top-k HUIs on massive data efficiently.•Prefix-based partitioning is devised to obtain prefix-based partitions.•Set enumeration tree in depth-first growing mode is presented to process partitions.•Subtree pruning rule is devised to reduce exploration space of set enumeration tree.