Table 1 Comparison of uncertainty strategies
From: Uncertainty in big data analytics: survey, opportunities, and challenges
Uncertainty models | Features |
|---|---|
Probability theory Bayesian theory Shannon’s entropy | Powerful for handling randomness and subjective uncertainty where precision is required Capable of handling complex data [50] |
Fuzziness | Handles vague and imprecise information in systems that are difficult to model Precision not guaranteed Easy to implement and interpret [50] |
Belief function | Handle situations with some degree of ignorance Combines distinct evidence from several sources to compute the probability of specific hypotheses Considers all evidence available for the hypothesis Ideal for incomplete and high complex data Mathematically complex but improves uncertainty reduction [50] |
Rough set theory | Provides an objective form of analysis [47] Deals with vagueness in data Minimal information necessary to determine set membership Only uses the information presented within the given data [51] |
Classification entropy | Handles ambiguity between the classes [39] |