An Algorithm for the Automatic Precisiation of the Meaning of Adjectives

Computing with words and perceptions is a technique that has been proven to be incredibly powerful to understand and compute with entities with imprecise nature, such as natural language and human perceptions. To be able to correctly handle these elements, however, a high degree of human-generated data is needed for the precisiation of meaning, the first step of computing with words and perceptions systems. This can be achieved via direct user intervention or with a - not existing yet - data set representing the meaning of words for several people. This aspect reduces the practical usability of computing with words and perceptions to cases where only an extremely reduced set of words needs to be understood (precisi-ated). In the current article, an algorithm able to fully automate the precisiation of the meaning of quantitative and qualitative adjectives using already available data - from a thesaurus - is proposed and analyzed in an exploratory study with six experts as participants. Results indicate an accuracy of the algorithm close to human-level accuracy for the task of precisiating another person's perceptions and help to understand some improvements to be taken into account for future developments of the algorithm. The presented algorithm represents a promising step toward human-like information processing.