Novel Span Measure, Spanning Sets and Applications
This work provides an incremental improvement for researchers in rough set theory and NLP by offering a more convenient uncertainty measure.
The paper tackles the problem of measuring uncertainty in spanning sets for natural language processing by proposing a novel span measure based on upper approximations, which simplifies computation compared to using boundary regions, and discusses its properties and applications.
Rough Set based Spanning Sets were recently proposed to deal with uncertainties arising in the problem in domain of natural language processing problems. This paper presents a novel span measure using upper approximations. The key contribution of this paper is to propose another uncertainty measure of span and spanning sets. Firstly, this paper proposes a new definition of computing span which use upper approximation instead of boundary regions. This is useful in situations where computing upper approximations are much more convenient that computing boundary region. Secondly, properties of novel span and relation with earlier span measure are discussed. Thirdly, the paper presents application areas where the proposed span measure can be utilized.