Abstract
Taxonomy is generated to effectively organize and access data that is large in volume, as taxonomy is a way of representing concepts that exist in data. It needs to be evolved to reflect changes occur continuously in data. Existing automatic taxonomy generation techniques do not handle the evolution of data, therefore their generated taxonomies do not truly represent the data. The evolution of data can be handled either by regenerating taxonomy from scratch, or incrementally evolving taxonomy whenever changes occur in the data. The former approach is not economical subject to time and resources. Taxonomy
incremental evolution (TIE) algorithm, proposed in this paper, is a novel attempt to handle an evolving data. It serves as a layer over an existing clustering-based taxonomy generation technique and incrementally evolves an existing taxonomy. The algorithm was evaluated on scholarly articles selected from computing domain. It was found that the algorithm evolves taxonomy in a
considerably shorter period of time, having better quality per unit time as compared to the taxonomy regenerated from scratch.
incremental evolution (TIE) algorithm, proposed in this paper, is a novel attempt to handle an evolving data. It serves as a layer over an existing clustering-based taxonomy generation technique and incrementally evolves an existing taxonomy. The algorithm was evaluated on scholarly articles selected from computing domain. It was found that the algorithm evolves taxonomy in a
considerably shorter period of time, having better quality per unit time as compared to the taxonomy regenerated from scratch.
Original language | English |
---|---|
Pages (from-to) | 763–782 |
Number of pages | 20 |
Journal | Frontiers of Information Technology and Electronic Engineering |
Volume | 19 |
Issue number | 6 |
DOIs | |
Publication status | Published - Jun 2018 |
Keywords
- Taxonomy
- Clustering algorithms
- Information science
- Knowledge management
- Machine learning