Research indexJournal article
SIMILANT
Bernhauer, Skopal, Holubová and 2 others
3 Nov 2019
NobleIDark:/48914/w1/20260525/EAAFDC97DOI10.1145/3357384.3357852
SIMILANT
Bernhauer, Skopal, Holubová and 2 others · 2019 · Subscription
NobleIDark:/48914/w1/20260525/EAAFDC97DOI10.1145/3357384.3357852
The full text is at the publisher
You can read the abstract here, or open it at the publisher.
Abstract
We present SIMILANT, a data analytics tool for modeling similarity in content-based retrieval scenarios. In similarity search, data elements are modeled using black-box descriptors, where a pair-wise similarity function is the only way how to relate data elements to each other. Only these relations provide information about the dataset structure. Data analysts need to identify meaningful combinations of descriptors and similarity functions effectively. Therefore, we proposed a tool enabling a data analyst to systematically browse, tune, and analyze similarity models for a specific domain.