Hybrid Pareto-front meta-heuristic Algorithm for time series automatic spectral clustering using community detection in complex networks

Mojtaba Manochehri , Syeed Mohammad Bagher Davoodi , Mohammad Hossain Sajadnia

One of the issues in the field of social network is finding similar time series using community detection in complex networks, which every community, include one or several complex time patterns of mass data that these patterns called partition. In recent years spectral clustering has been an important issues in clustering algorithm which is a part of np-hard issues and to solve it we can use multi-objective meta-heuristic algorithms such as particle swarm and biogeography, and teaching learning and the improved hybrid algorithm - which we presented in this article. Multi-objective meta-heuristic algorithms has a set of solutions that each can be the most optimal answer from a different perspective. This set of answers in the field of meta-heuristic algorithms and multi-objective optimization is known as Pareto Front. The result of the implementation of multi-objective algorithms shows that the improved algorithm has been able to provide a relatively better solution to rescue from local optimal traps, and the outcomes indicate the promising performance of the hybrid algorithm over the Biogeoraphy based optimization (BBO) and Differential Evolution (DE) algorithms.