We include the following resources from our ConcepT system in this repository:
- UCCM dataset.txt The UCCM dataset. It is used to evaluate the performance of our approach for concept mining and it contains 10,000 samples.
- concept_tagging_data.txt The document tagging dataset. It is used to evaluate the document tagging accuracy of ConcepT, and it contains 11,547 documents with concept tags.
- topic_concept_instance_taxonomy_data.sample.txt Topic-concept-instance taxonomy. It contains 1000 topic-concept-instance samples from our constructed taxonomy.
- seed_concept_patterns.txt The seed concept patterns for bootstrapping-based concept mining. It contains the seed string patterns we utilized for bootstrapping-based concept mining from queries.
- topics.txt Pre-defined topic list. It contains our 31 pre-defined topics for taxonomy construction.
The above datasets are published for research purpose.