We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
ASSESSING THE PERFORMANCE OF COMPRESSION BASED CLUSTERING FOR TEXT MINING.
- Authors
CERNIAN, Alexandra; CARSTOIU, Dorin; OLTEANU, Adriana; SGARCIU, Valentin
- Abstract
The nature of the human brain is to find patterns in whatever surrounds us. Thus, we are all developing models of our personal universe. In an extended form, a constant preoccupation of philosophers has been to model the universe. Clustering is one of the most useful tools in the data mining process for discovering groups and identifying patterns in the underlying data. This paper addresses the compression based clustering approach and focuses on validating this method in the context of text mining. The idea is supported by the evidence that compression algorithms provide a good evaluation of the informational content. In this context, we developed an integrated clustering platform, called EasyClustering, which incorporates 3 compressors, 4 distance metrics and 3 clustering algorithms. The experimental validation presented in this paper focuses on clustering text documents based on informational content.
- Subjects
PERFORMANCE evaluation; TEXT mining; DOCUMENT clustering; DATA compression; DATA mining; CLASSIFICATION algorithms
- Publication
Economic Computation & Economic Cybernetics Studies & Research, 2016, Vol 50, Issue 2, p197
- ISSN
0424-267X
- Publication type
Article