Page 237 - Contributed Paper Session (CPS) - Volume 2

CPS1845 Devni P.S. et al.
where pa( ) is the parent of in BN, and () reflects the properties of BN

(Fenton & Neil, 2013).

3. Result
We will conduct a simple and hypothetical survey that aims to determine
the level of damage to houses due to earthquakes. The results of this survey
can later be used as one of the disaster mitigation efforts and basic guidelines
for insurance companies in determining the mount of premium rates. The data
used is the data on building damage, especially the residence in the city of
Padang due to the earthquake in September 2009 which was 7.9 on the scale
of SR and centered on the coast of West Sumatra. The earthquake caused
severe damage in several areas in West. Data was obtained from the Regional
Disaster Management Agency of Padang (BPBD Padang City) and the
Indonesian Meteorology, Climatology and Geophysics Agency. The data to be
used as a case study consists of 61344 building data. The research data
consists of three independent variables (close to faults, slope, and epicenter
distance) and four dependent variables (construction, landslide risk, PGA, and
damage). In the Figure 2, we present summary data on the level of damage
for each District that we obtained from BPBD Padang City. In this example
application, we will pay attention to the data for each building, in the Table 3
the details of the seven discrete variables used are detailed.
There are two types of relationships between variables, namely direct
relationships and indirect relationships (relationships mediated by other
variables). Both types of relationships can be represented effectively using
directed graphs. Each node in the graph represents each variable in the study.
So, the graph generated from this example consists of seven nodes,
symbolized by C, P, E, L, S, F, and D. The relationship between variables can be
seen in Figure 1, where the connection in this picture is formed based on
expert opinion with reference in a previous research paper (Li, Wang, & Leung,
2010).
Examples of direct dependence are E → P means that P depends on E. It
can be said that node E is the parent (parent) of node P and node P is the child
/ descendant (descendant) of node E. Indirect dependency relationship can be
seen from the graph, i.e. the sequence of arcs that lead from one variable to
another through one or more mediating variables. For example, S → L → D
means that D depends on S through L. However, we must note that the arrow
line only goes in one direction and does not turn back towards its original
node. The arc implies, that for each arc one variable is interpreted as a cause
and the other variable as an effect (eg E → P means that E causes P).

226 | I S I W S C 2 0 1 9