Weighted Hesitant Fuzzy Finite Rough Automaton: A Computing Model for Natural Language Processing

Abstract

The well-known computing models, classical/fuzzy/rough automata, in spite of their multifarious applications, fail to model those complex real-world systems that contain both vagueness and incomplete information in datasets of real-world complex systems. The computing model, [Formula: see text]-fuzzy rough automata, incorporates both vagueness and incomplete information in datasets of such systems but fails to incorporate the weights of fuzzy attributes. To overcome this issue, we first introduce the concept weighted hesitant fuzzy finite automaton (WHFFA), as a generalized notion of the hesitant [Formula: see text]-fuzzy automaton (HLFA), where weights show that the decision maker has distinct confidence in providing the possible valuation of the membership degree. We used the concept of weighted hesitant fuzzy rough set (WHFRS) as a hybrid concept of hesitant fuzzy rough set (HFRS) and weighted hesitant fuzzy set (WHFS) to introduce a novel computing model weighted hesitant fuzzy finite rough automaton (WHFFRA). The introduced WHFFRA is efficient for dealing with vagueness and incomplete information inherent in our natural languages and in datasets of the real-world complex systems. Finally, we discuss determinization of WHFFRA and demonstrated the application of introduced WHFFRA in decision-making scenarios of medical diagnosis problems.