Knowledge graphs are becoming ubiquitous in many scientific and industrial domains, ranging from biology, industrial engineering to natural language processing. In this work we explore how one of the largest currently available knowledge graphs, the Microsoft Concept Graph, can be used to construct interpretable features that are of potential use for the task of text classification. By exploiting graph-theoretic feature ranking, introduced as part of the existing tax2vec algorithm, we show that

more »

... massive, real-life knowledge graphs can be used for the construction of features, derived from the relational structure of the knowledge graph itself. To our knowledge, this is one of the first approaches that explores how interpretable features can be constructed from the Microsoft Concept graph with more than five million concepts and more than 80 million IsA relations for the task of text classification. The proposed solution was evaluated on eight real-life text classification data sets.