Abstract:The goal of dynamic community discovery is to quickly and accurately mine the network structure for individuals with similar attributes for classification. Correct classification can effectively help us screen out more desired results, and it also reveals the laws of dynamic network changes. We propose a dynamic community discovery algorithm, NOME, based on node occupancy assignment and multi-objective evolutionary clustering. NOME adopts the multi-objective evolutionary algorithm MOEA/D framework based on decomposition, which can simultaneously decompose the two objective functions of modularization and normalized mutual information into multiple single-objective problems. In this algorithm, we use a Physarum-based network model to initialize populations, and each population represents a group of community-divided solutions. The evolution of the population uses the crossover and mutation operations of the genome matrix. To make the population in the evolution process closer to a better community division result, we develop a new strategy for node occupancy assignment and cooperate with mutation operators, aiming at the boundary nodes in the connection between the community and the connection between communities, by calculating the comparison node. The occupancy rate of the community with the neighbor node, the node is assigned to the community with the highest occupancy rate, and the authenticity of the community division is improved. In addition, to select high-quality final solutions from candidate solutions, we use a rationalized selection strategy from the external population size to obtain better time costs through smaller snapshot quality loss. Finally, comparative experiments with other representative dynamic community detection algorithms on synthetic and real datasets show that our proposed method has a better balance between snapshot quality and time cost.

Esnmf: Evolutionary Symmetric Nonnegative Matrix Factorization For Dissecting Dynamic Microbial Networks

Differential Markov Random Field Analysis with an Application to Detecting Differential Microbial Community Networks.

MicroNet-MIMRF: a Microbial Network Inference Approach Based on Mutual Information and Markov Random Fields.

Evolutionary Markov Dynamics for Network Community Detection

Integrative Analysis for Identifying Co-Modules of Microbe-Disease Data by Matrix Tri-Factorization With Phylogenetic Information

Dynamic Network Embeddings for Network Evolution Analysis.

iMS2Net: A multiscale networking methodology to decipher metabolic synergy of organism

Novel Collaborative Weighted Non-negative Matrix Factorization Improves Prediction of Disease-Associated Human Microbes

A novel approach to differential expression analysis of co-occurrence networks for small-sampled microbiome data

Microbial community pattern detection in human body habitats via ensemble clustering framework

An Overlapping Community Detection Approach in Ego-Splitting Networks Using Symmetric Nonnegative Matrix Factorization

Kldm: Inferring Multiple Metagenomic Association Networks Based on the Variation of Environmental Factors.

Improving temporal smoothness and snapshot quality in dynamic network community discovery using NOME algorithm

Building Predictive Understanding of Microbial Ecology by Bridging Microbial Growth Kinetics and Microbial Population Dynamics

Bayesian modeling of co-occurrence microbial interaction networks

Towards more accurate microbial source tracking via non-negative matrix factorization (NMF)

Longitudinal Microbiome-based Interpretable Machine Learning for Identification of Time-Varying Biomarkers in Early Prediction of Disease Outcomes

Community evolution detection in dynamic heterogeneous information networks.

Dynamic Community Discovery Via Common Subspace Projection

Online Detecting and Tracking of the Evolution of User Communities

mbDriver: identifying driver microbes in microbial communities based on time-series microbiome data