Customizing Semantics for Individuals with Attitudinal HFLTS Possibility Distributions

Linguistic computational techniques based on hesitant fuzzy linguistic term set (HFLTS) have been swiftly advanced on various fronts during the past five years. However, one of the key issues in the existing theoretical developments is that modeling possibility distribution-based semantics involves a relatively strict constraint that linguistic terms are uniformly distributed across an HFLTS. Releasing the constraint of uniform HFLTS through which individual semantics could be customized is challenging yet intriguing for participants interested in this topic. Comparative linguistic terms (CLEs) generated from context-free grammar facilitate flexible and accurate linguistic elicitation, and in consideration of computational simplicity they are transformed into HFLTSs which are machine manipulatable. It is imperative that the precision of customized individual semantics can be significantly improved in respect of different CLEs. This study proposes a novel possibility computation structure for HFLTS possibility distributions based on the linguistic terms similarity measure. The uniquely established linguistic terms in each and every CLE are initially treated as referential items for comparison. Then, possibilities of linguistic terms in a transformed HFLTS can be calculated as their similarity degrees to the predetermined referential item. Subsequently, the Interweaving method, in which a consistent Inner Interweaving Matrix needs to be constructed, is adopted to adapt Attitudinal Character to attain appealing degrees characterized in the unit interval. The generated attitudinal HFLTS possibility distributions provide a solution to the problem of modeling individually the semantic implications of CLEs. Several illustrative examples and comparative analyses further demonstrate that individual semantics endowed with Attitudinal Character model efficiently individual differences in cognitive styles.