A Granular Computing approach to the design of optimized graph classification systems

Abstract

Research on Graph-based pattern recognition and Soft Computing systems has attracted many scientists and engineers in several different contexts. This fact is motivated by the reason that graphs are general structures able to encode both topological and semantic information in data. While the data modeling properties of graphs are of indisputable power, there are still different concerns about the best way to compute similarity functions in an effective and efficient manner. To this end, suited transformation procedures are usually conceived to address the well-known Inexact Graph Matching problem in an explicit embedding space. In this paper, we propose two graph embedding algorithms based on the Granular Computing paradigm, which are engineered as key procedures of a general-purpose graph classification system. Tests have been conducted on benchmarking datasets relying on both synthetic and real-world data, achieving competitive results in terms of test set classification accuracy

Two density-based k-means initialization algorithms for non-metric data clustering

“…Future directions include the application of the herein presented representative selection algorithm as the core granulation engine of a clustering-based graph classification system [4]. Moreover, we will evaluate DBCRIMES also as a pure representative selection algorithm.…”

“…The need is even more pressing when the problem is conceived on data characterized by non-trivial geometry [4,12,20,23,24,26,27,29,32,36,37]. For this very reason, researchers working in related areas have produced several clustering techniques [5,6,8,13,14,18,19,40].…”

Granular computing: from granularity optimization to multi-granularity joint problem solving

“…Liu et al (2014) proposed a granular computing classification algorithm based on distance measure and verified its superior accuracy by experiments. Bianchi et al (2014) proposed a granular computing model to optimize pattern classification and found the optimal classification by this model. It has become a key point to design common pattern classification systems.…”

Test-cost-sensitive attribute reduction on heterogeneous data for adaptive neighborhood model

“…A major challenge is how to deal with test-cost-sensitive attribute reduction on heterogeneous data. Granular computing (Li et al 2015;Bianchi et al 2014; offers an unified framework related to how information is grouped together and how these groups can be utilized to make decisions. Although granular computing can express groups, more commonly known as information granules, it can use a variety of representations to express such granules, which could be fuzzy sets (Sanchez et al 2014(Sanchez et al , 2015, rough sets (Lin 1998(Lin , 2002Jing 2014;Pawlak 1982Pawlak , 1991Yao and Zhong 2002), neighborhood rough sets (Hu et al 2008b;Zhang et al 2013Zhang et al , 2014Zhu 2007;Zhu and Wang 2003).…”