Double-quantified linguistic variable

Abstract

Regarding the natural language representation, there are two main methods, namely, the linguistic variable and multi-granularity linguistic term set. Although these two techniques are big achievements in the natural language representation researches, there are still some challenges: 1) the linguistic variable cannot handle the probability information, and 2) the multi-granularity linguistic term set cannot handle the negative words and incomplete probability information. To bridge these research gaps, this study proposes the double-quantified linguistic variable to quantify the natural language from two aspects: membership degree and probability. Firstly, we compare the linguistic variable with the multi-granularity linguistic term set and identify the differences and shortages of them. Based on these analyses, the double-quantified linguistic variable as a quintuple is introduced and every part of the double-quantified linguistic variable is explained in detail. After that, the basic operations and comparison method of the double-quantified linguistic variables are given. This study ends with some concluding remarks and future research directions.