Concept-Cognitive Learning Model for Incremental Concept Learning

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

“…These basses are generated thanks to the concept lattice (Lakhal & Stumme, 2005) (eg., Duquenne–Guigues bases (Guigues & Duquenne, 1986), Luxenburger (Luxenburger, 1991), and Min–Max basis (Pasquier et al, 2005)). Various applications are related to the formal concept lattice. So, once the concept lattice is generated, it can be used for multiple purposes including the ARM, information retrieval (Kumar, Mouliswaran, Amriteya, & Arun, 2015), clustering, concept learning (Hao et al, 2019; Shi et al, 2018), classification rule mining (Liu & Li, 2017), and much more.…”

“…Consequently, formal concept lattices represent conceptual hierarchies embedded in the data under study (Ganter, Stumme, & Wille, 2002). For its formal representations and analysis of problems, formal concept lattice spans a wide range of applications (Ganter et al, 2005; Ganter & Wille, 2012; Poelmans, Ignatov, Kuznetsov, & Dedene, 2013) including information retrieval (Aloui & Grissa, 2015; Dau, Ducrou, & Eklund, 2008; Formica, 2012; Godin, Pichet, & Gecsei, 1989; Priss, 2000; Zhi & Li, 2018; Zou, Deng, Wan, Wang, & Deng, 2017), knowledge representation and discovery (Li, Mei, Kumar, & Zhang, 2013; Li, Mei, & Lv, 2012; Li, Mei, Xu, & Qian, 2015; Xu & Li, 2014), software engineering (Eisenbarth, Koschke, & Simon, 2001; Fischer, 2000; Godin & Mili, 1993; Krone & Snelting, 1994), and machine learning (Carpineto & Romano, 1996; Kuznetsov, 2004; Mi, Liu, et al, 2020; Mi, Shi, et al, 2020; Shi, Mi, Li, & Liu, 2018). In software engineering, concept lattices are used as a powerful tool to structure random observations.…”

“…Also, FCA can be used for building a lattice‐based learning model in the pattern structures (Kuznetsov, 2004). Furthermore, FCA can be utilized as an incremental mechanism for dynamic learning rather than static learning (Hao, Bouzouane, & Gaboury, 2019; Shi et al, 2018). Also, it can be used in semi‐supervised concept learning by considering the structural information of the concepts (concept spaces) (Mi, Liu, et al, 2020).…”

Fuzzy Rule-Based Classification Method for Incremental Rule Learning

“…Considering that it takes more time to get rules from concepts [12], [14]- [18], Mi et al [19] proposed a fuzzy-based concept learning model by employing the attribute-oriented fuzzy concept similarity degree. To avoid repetitive training, researchers proposed many incremental learning systems [20]- [25] to meet different task requirements. For instance, Shi et al [25] proposed a concurrent incremental learning technique by continuously accommodating newly added data to meet the requirements of classification task.…”

“…To avoid repetitive training, researchers proposed many incremental learning systems [20]- [25] to meet different task requirements. For instance, Shi et al [25] proposed a concurrent incremental learning technique by continuously accommodating newly added data to meet the requirements of classification task. Aiming at the non-stationary environments, Song et al [21] introduced a window-based self-adaptive fuzzy network to continuously modify the network through identifying new knowledge from the previous samples, while Yu et al [20] presented a Gaussian membership-based SOINN (Gm-SOINN) by employing a Gaussian membership degree to indicate which node is a winner for solving the stabilityplasticity dilemma.…”