XSearchKG: A Platform for Explainable Keyword Search over Knowledge Graphs

One of the most user-friendly methods to search over knowledge graphs is the usage of keyword queries. They offer a simple text input that requires no technical or domain knowledge. Most existing approaches for keyword search over graph-shaped data rely on graph traversal algorithms to find connections between keywords. They mostly concentrate on achieving efficiency and effectiveness (accurate ranking), but ignore usability, visualization, and interactive result presentation. All of which offer better support to non-experienced users. Moreover, it is not sufficient to just show a raw list of results, but it is also important to explain why a specific result is proposed. This not only provides an abstract view of the capabilities and limitations of the search system, but also increases confidence and helps discover new interesting facts. We propose XSearchKG, a platform for explainable keyword search over knowledge graphs that extends our previously proposed graph traversal-based approach and complements it with an interactive user interface for results explanation and browsing.