i-HUMO: An Interactive Human and Machine Cooperation Framework for Entity Resolution with Quality Guarantees.

摘要: Even though many approaches have been proposed for entity resolution (ER), it remains very challenging to find one with quality guarantees. To this end, we propose an interactive HUman and Machine cOoperation framework ER, denoted by i-HUMO. Similar the existing HUMO framework, i-HUMO enforces both precision recall levels dividing ER workload between human machine. It essentially makes machine label easy instances while assigning more human. However, is a major improvement over in that interactive: its process of selection optimized based on real-time risk analysis human-labeled results as well pre-specified metrics. In paper, first introduce then present technique prioritize manual labeling. Finally, empirically evaluate i-HUMO's performance real data. Our extensive experiments show effective enforcing guarantees, compared state-of-the-art alternatives, can achieve better control reduced cost.