K-means and bayesian networks to determine building damage levels

Open

Devni Prima Sari, Dedi Rosadi, Adhitya Ronnie Effendie, Danardono

2019 Telkomnika (Telecommunication Computing Electronics and Control) Vol. 17 Issue 2 Article Cited by 15

Abstract

Many troubles in life require decision-making with convoluted processes because they are caused by uncertainty about the process of relationships that appear in the system. This problem leads to the creation of a model called the Bayesian Network. Bayesian Network is a Bayesian supported development supported by computing advancements. The Bayesian network has also been developed in various fields. At this time, information can implement Bayesian Networks in determining the extent of damage to buildings using individual building data. In practice, there is mixed data which is a combination of continuous and discrete variables. Therefore, to simplify the study it is assumed that all variables are discrete in order to solve practical problems in the implementation of theory. Discretization method used is the K-Means clustering because the percentage of validity obtained by this method is greater than the binning method. © 2019 Universitas Ahmad Dahlan.

Affiliations

Department of Mathematics, Universitas Gadjah Mada, Yogyakarta, Indonesia; Department of Mathematics, Universitas Negeri Padang, Padang, Indonesia