ABSTRACT
This chapter emphasizes the nature of the representation in terms of processes of granulation and degranulation. It looks at how such a description can be realized and which of its parameters may impact the quality of the resulting representation. The chapter shows that the representation of this form is inherently associated with the emergence of higher-type information granules. With regard to encoding and decoding abilities, it is shown that any representation of information granules and numeric entities is associated with degranulation error, which could be absorbed by forming an information granule of higher type. The chapter also shows that a surprisingly simple yet powerful result whose essence could be summarized as follows: a codebook formed of fuzzy triangular fuzzy sets with an overlap of one half between two neighboring elements leads to the lossless reconstruction mechanism. For instance, while the original information granules are fuzzy sets, the results of degranulation become fuzzy sets of type 2.
