Abstract
SIGenBench is a generalization and benchmark framework for the notion of semantic importance. Semantic importance is proposed to model and formalize data importance from various data orderings. It is primarily used in a stream reasoning context, where hidden information can be extracted out of the data streams. There are two main features for SIGenBench: Generalization: it refers to make semantic importance reusable by others. SIGenBench generalizes semantic importance by connecting semantic importance with the state of the art stream reasoning techniques, such as window operational semantics, continuous query languages, etc. Benchmark: it refers to quantify any benefits brought by semantic importance. SIGenBench provides a benchmark system that records the key performance indicators including precision, response time, memory consumption, and throughput. It also provides a data generator that can simulate 9 different patterns for streaming data, which is flexible and sufficient to test semantic importance in various streaming scenarios.
Exploratory License
Eligible for exploratory license
Market Sector
Data Sciences