A Data-Sensitive Approach for Fuzzy Concept Extraction

Abstract

Nowadays, extracting fuzzy concepts using fuzzy formal concept analysis (FFCA) is an increasingly important process in many fields including data-mining, information retrieval, and Ontology construction. However, recent studies have heightened the need for more efficient approaches for extracting a reduced count of distinct fuzzy concepts in a reasonable time. This paper aims mainly to address such challenge. Generally, the proposed approach is composed of two main stages to handle quantitative data effectively. Firstly, a data-sensitive fuzzification stage maps the many-valued context to a more consistent fuzzy one. Secondly, an enhanced algorithm generates a reduced count of valuable fuzzy concepts and merges the more similar ones. Accordingly, the proposed approach can efficiently handle very large datasets where the process of extracting all fuzzy concepts is an intractable task. Surprisingly, the proposed approach reduces the overall processing time and complexity when compared with some other previous approaches.

A comprehensive review on updating concept lattices and its application in updating association rules

“…FFCA (Pollandt, 1997) is a generalized form of the classical FCA that can handle broader data types over the classical ones by incorporating the fuzzy set theory (Shemis et al, 2018) proposed by (Zadeh, 1965). FFCA handles not only binary formal contexts, but also quantitative and fuzzy formal contexts and hence has a wide range of real‐world applications.…”

“…To show how the one‐sided fuzzy concept looks like, we processed the fuzzy context shown in Table 6 and extracted the whole set of fuzzy concepts in Table 7. In literature, an extensive number of algorithms exist for extracting the one‐sided fuzzy concepts such as (Majidian et al, 2011; Shemis et al, 2018; Shemis & Gadallah, 2016; Zheng et al, 2009; Zou et al, 2017). Definition Consider two fuzzy concepts ( ϕ (( A 1 ), B 1 ) and ( ϕ (( A 2 ), B 2 ) of the fuzzy context .…”

“…Nowadays, real‐world data are always changed continually due to the insertion of new data, deletion, or modification of existing ones. The process of extracting formal concepts and drawing concept lattice is a #P‐complete problem (Kuznetsov & Obiedkov, 2002; Shemis, Gadallah, & Hefny, 2018). Therefore, one of the most promising research areas is how to update the already extracted formal concept lattice upon data change rather than constructing it from scratch (Zhi & Li, 2018).…”