CBSSD: community-based semantic subgroup discovery

Abstract

Modern data mining algorithms frequently need to address the task of learning from heterogeneous data, including various sources of background knowledge. A data mining task where ontologies are used as background knowledge in data analysis is referred to as semantic data mining. A specific semantic data mining task is semantic subgroup discovery: a rule learning approach enabling ontology terms to be used in subgroup descriptions learned from class labeled data. This paper presents Community-Based Semantic Subgroup Discovery (CBSSD), a novel approach that advances ontology-based subgroup identification by exploiting the structural properties of induced complex networks related to the studied phenomenon. Following the idea of multi-view learning, using different sources of information to obtain better models, the CBSSD approach can leverage different types of nodes of the induced complex network, simultaneously using information from multiple levels of a biological system. The approach was tested on ten data sets consisting of genes related to complex diseases, as well as core metabolic processes. The experimental results demonstrate that the CBSSD approach is scalable, applicable to large complex networks, and that it can be used to identify significant combinations of terms, which can not be uncovered by contemporary term enrichment analysis approaches.

Multi-resolution visualization and analysis of biomolecular networks through hierarchical community detection and web-based graphical tools

“…The idea of applying enrichment methods to identify relevant subnetworks from a biological standpoint is a well-known approach performed in literature using: a) ontology-based enrichment analysis [ 62 ], b) de novo enrichment analysis [ 63 ], c) community-based algorithms followed by semantic rule induction [ 64 ] to link biological explanations to each discovered subgroups. Similarly, in our work we used community detection followed by enrichment analysis to dissect our biomolecular network in subgroups described by enriched pathways coherent and correlated inside each community.…”

New Cross-Talks between Pathways Involved in Grapevine Infection with ‘Candidatus Phytoplasma solani’ Revealed by Temporal Network Modelling

“…Conceptually, community-based semantic subgroup discovery (CBSSD) consists of two main steps, i.e., community detection followed by a one-versus-all enrichment procedure, which here, was additionally corrected for multiple hypothesis testing. We refer the reader to Škrlj et al (2019) [ 24 ] for a more detailed overview of the enrichment process and for additional theoretical insights. Here, we used the method ‘as-is’, with the default hyperparameter settings including Bonferroni’s multiple test correction and the significance threshold of 0.05 (Fisher’s exact test).…”

“…Applied network-based methodologies have shown promising results in many branches of plant biology, including studies of immunity [ 21 ] and regulatory pathways [ 22 , 23 ]. When considering, for example, community enrichment [ 24 ], drug design [ 25 ] or the structural analysis of protein binding sites [ 26 , 27 ], network-based approaches have also been applied to organisms other than plants.…”

Semantic Reasoning from Model-Agnostic Explanations

“…The use of background knowledge for improving the understandability of machine learning systems has also been referred to as semantic data mining [11]. For example, in relational domains, the CBSSD methodology [12] focuses on data mining tasks that use ontologies as background knowledge when explaining emergent structures in complex networks. Similarly, the NetSDM approach [13] explored how redundant ontologies are for the purposes of inductive rule learning.…”