FCA-Based Data Analysis for Discovering Association Rules in Social Network Service

Abstract

Recently, Formal Concept Analysis (FCA) have been widely used for various purposes in many different domains such as data mining, machine learning, knowledge management and so on. In this paper, we introduce FCA as the basis for a practical and well founded methodological approach for data analysis which identifies conceptual structures among data sets. As well as, we propose a FCA-based data analysis for discovering association rules by using polarity from social contents. Additionally, we show the experiments that demonstrate how our data analysis approaches can be applied for knowledge discovery by using association rules.

An Efficient Algorithm for Decreasing the Granularity Levels of Attributes in Formal Concept Analysis

“…First, the procedure Generate creates a new concept and adds the new concept to concept lattice (lines 1-2). According to Proposition 9, we test every candidate in c.Children to find real children of newConcept (lines [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. Note that the concept c.children.indicator points to has already been obtained after executing the Preprocessprocedure.…”

“…Algorithm 1 reveals that c can be marked directly if c is a merged concept. When c is a deleted or modified concept, it requires comparisons between c andc.Parentsfor finding a parent with c.Intent = parent.Intent (lines [5][6][7][8][9][10][11][12][13]. This operation requires only one comparison in the best case or |G| comparisons at worst case between sets (intents) which takes at most O(|G||M | 2 ) time.…”

“…Each concept of a formal lattice consists of an extent and an intent, which are closely connected by the conceptual relationship. FCA focuses on obtaining different forms of outputs from original formal context for interesting information, and it has a wide range of applications in various fields including information retrieval [4]- [6], data mining [7], social network analysis [8]- [10], gene expression [11]- [13] machine learning [14], [15], ontologies [16], [17], graph mining [18]- [20]. A comprehensive survey about research trends and applications on formal concept analysis is given by Singh et al [21].…”