This paper presents a general methodological framework for constructing context-aware fuzzy partitions that extend conventional crisp categorizations. The approach is based on Nov´ak’s theory of fuzzy contexts and is implemented using the R package lfl. It enables smooth and interpretable transitions between adjacent classes while preserving the original categorical structure. To illustrate the procedure, we apply it to derive fitness-specific fuzzy partitions of Body Mass Index, where the conventional four categories (underweight, normal weight, overweight, obese) are adapted according to individual levels of cardiorespiratory fitness.
Alijani et al. (Thu,) studied this question.