Structural learning of dynamic Bayesian Networks using multiobjective parallel evolutionary algorithm

• Grant number: 18/23139-8
• Support Opportunities: Scholarships in Brazil - Doctorate (Direct)
• Effective date (Start): May 01, 2019
• Effective date (End): February 29, 2024
• Field of knowledge: Engineering - Electrical Engineering
• Principal Investigator: Carlos Dias Maciel
• Grantee: Rafael Rodrigues Mendes Ribeiro
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated research grant: 14/50851-0 - INCT 2014: National Institute of Science and Technology for Cooperative Autonomous Systems Applied in Security and Environment, AP.TEM
• Associated scholarship(s): 21/09396-0 - Analysis of the Impact of data discretization in the bayesian network search space, BE.EP.DD

Abstract

Now a days there is a need for a statistical theory for signal analysis. To meet this need, in recent years Bayesian inference has been used, in which data is used to infer the structure and parameters of the Bayesian Network (which is a graphical description of the conditional probability among random variables). The Signal Processing Laboratory has a long history of work involving Bayesian Networks. The learning of the structure of the Bayesian Network is an NP-hard problem and of great importance for several studies. In order to create a structural learning algorithm for Bayesian Networks, we intend to develop a multi-objective optimization algorithm, since it seeks to find a solution that has a compromise between the various criteria defined. An efficient approach to developing this algorithm is to use evolutionary algorithms. In order to optimize the execution time of the evolutionary algorithm, we intend to use a parallel evolutionary algorithm. It is expected that the final method obtained will present a very short execution time compared to the currently available methods. (AU)