microsoft_concept — Microsoft Concept Graph
- Description
Network of "IsA" relations between concepts and instances from the Microsoft Concept Graph, released by Microsoft Research in 2016. The data derive from Probase, a knowledge base that was built by automatically extracting IsA relations from a corpus of billions of web pages with syntactic patterns, and their frequencies. This is the core IsA data set of the public release, containing 33,377,320 relations between 5,376,526 concepts and 12,501,527 instances (named entities, objects, actions, attributes, etc.).
An edge connects a concept to an instance of it, with a weight equal to the number of times the relation was extracted. Nodes are identified by their (lowercase) text, so a term that appears both as a concept and as an instance is represented by a single node; therefore the network is not strictly bipartite. Edges are undirected, and the direction of the relation is not stored.
The network has the following properties:
name: the text of the concept or instance.count: the number of times the IsA relation was extracted from the web corpus.
- Tags
- Informational Relatedness Weighted
- Citation
-
- Z. Wang et al., "An Inference Approach to Basic Level of Categorization." Proc. ACM Internat. Conf. on Info. and Know. Management (CIKM 2015) 10.1145/2806416.2806533 Sci-Hub
- Upstream URL OK
- https://web.archive.org/web/20220122082444/https://concept.research.microsoft.com/Home/Download
- Networks
-
Tip: hover over a column header for a description.
Column legend
- $\left<k\right>$
- Mean degree
- $\sigma_k$
- Standard deviation of the degrees
- $\lambda_h$
- Largest eigenvalue of the Hashimoto (non-backtracking) matrix
- $\tau$
- Random walk mixing time
- $r$
- Degree assortativity
- $c$
- Global clustering coefficient
- $\oslash$
- (Pseudo-) diameter
- $S$
- Relative size of the largest component
- NPs
- Node properties
- EPs
- Edge properties