CLTR: Collectively linking temporal records across heterogeneous sources

Abstract

A huge volume of data on the Web are continually made available, which provides users rich amount of information to learn more about entities. In addition to attribute values of entities, there is often additional relational information, such as friendship on social networks, coauthorship of papers. However, to understand how these facts across heterogeneous data sources are related is challenging for users due to entity evolution over time. In this paper, we propose a novel system to help users find how records are temporally related and understand how entity profiles evolve over time. Our system is able to Collectively Link Temporal Records (CLTR) by taking advantage of evidence from both attribute and relational information on multiple sources. We demonstrate how CLTR allows users to explore time-varying history of targeted entities and visualizes multi-type relations among entities.