Implementation Fuzzy Weighted Product Preparation Post Disaster

Reconstruction and Rehabilitation Action based Dynamics Decision

Support System

Agung Teguh Wibowo Almais

, Fatchurrohman

, Khadijah Fahmi Hayati Holle

, Kurnia Siwi

Kinasih

, Dyah Ayu Wiranti

and Septian Yustina Yasin

Informatics Engineering State Islamic University Maulana Malik Ibrahim Malang, Indonesia

Keywords:

Decision Support System Dynamic; Fuzzy-Weighted Product; pattern data; test data; Surveyor

Abstract:

The preparation of post-disaster reconstruction rehabilitation action is conducted to determine the level of dam-

age and loss of post-natural disasters that must be handled by the government. In order to prevent the level

of post-disaster damage and loss in accordance with the conditions in the ﬁeld then conducted research im-

plementing Decision Support System Dynamic (DSSD) using the method of Fuzzy-Weighted Product (FWP).

The Decision Support System Dynamic (DSSD) is the development of the latest Decision Support System

(DSS) model, while Fuzzy is an algorithm for determining the level of importance of each criterion used in the

Weighted Product method. Whereas Weighted Product is used as a system pattern data formation. The result

is a test of the confusion matrix value of the pattern data calculated using the FuzzyWeighted Product (F-WP)

method compared to the different test data. The results of the test resulted in three different types of data i.e.

the same test data as the pattern data, the test data not the same as the pattern data, and the test data that could

not be ﬂown for testing. Each of these types of test data has a percentage of 73% of the same test data like the

pattern data, 22% of test data is not equal to the pattern data, and 5% is data that can not be used as test data.

From the results of the test can be concluded that the method Fuzzy-Weighted Product (F-WP) can be applied

to the Decision Support System Dynamic (DSSD) to assist surveyors in the preparation of the rehabilitation

of post-disaster reconstruction action.

1 INTRODUCTION

Determination of the level of damage and loss of the

post-disaster sector is the action for the preparation of

rehabilitation and post-disaster reconstruction. This

activity is one of the things that must be done by one

of the teams in the Regional Disaster Management

Agency (BPBD), which is the planning and Control

Team (P3B) to determine the level of loss and damage

to the sector Affected by natural disasters, knowing

the level of loss and damage will be easy in drafting

the rehabilitation action and postdisaster reconstruc-

tion to determine the amount of assistance that should

be channeled to victims affected by the disaster nature

(Almais et al., 2016).

In his research (Oetari, 2014) explained that to de-

termine the level of damage and damage to the sector

due to natural disasters is to use a reference called

the Economic Commission for Latin America and the

Caribbean (ECLAC). The journal has also been ex-

plained that ECLAC has been used by the Indonesian

government to calculate losses and damage to earth-

quakes in Yogyakarta’s special region in 2006 and

tsunami in special regions of ACEH. At ECLAC there

are references to how to determine the level of loss

and breakdown of postdisaster sectors. The current

Problem is how the ECLAC is applied easily at the

time of the surveyors at P3B conducted an assessment

of the level of loss and damage to the post-disaster

sector. Problems can be solved by implementing In-

formation Technology (IT). One of the areas in IT

can help the problem is the Decision Support System

(DSS). According to (Suryadi and Ramdhani, 2000)

in his book mentioned that to do research on the De-

cision Support System Many requirements that must

be met, including the problems discussed must have a

high scientiﬁc level and semi Structured.

With the development of the DSS era developed

into a dynamic system with the term Decision Sup-

port System Dynamic (DSSD). Decision Support Sys-

tem (DSSD) has a difference with the Decision Sup-

port System (DSS), which is the Decision Support

272

Almais, A., Fatchurrohman, ., Holle, K., Kinasih, K., Wiranti, D. and Yasin, S.

Implementation Fuzzy Weighted Product Preparation Post Disaster Reconstruction and Rehabilitation Action based Dynamics Decision Support System.

DOI: 10.5220/0009909002720277

In Proceedings of the International Conferences on Information System and Technology (CONRIST 2019), pages 272-277

ISBN: 978-989-758-453-4

System (DSS) that does not change the system that

is already running when adding criteria and alterna-

tives. With the concept of Decision Support System

dynamic (DSSD), the Decision Support System Dy-

namic (DSSD) is suitable if applied to help the Gov-

ernment (P3B) in conducting the assessment of loss

rate and damage to the postdisaster sector due to stan-

dard The criteria used to do so will someday increase

or less depending on government policy. At this time

the standard criteria to conduct an assessment of the

level of loss and damage post-natural disasters using

standard criteria from the Public Works ofﬁce on cri-

teria to determine the home or building earthquake re-

sistant natural disasters. These criteria are 1). State

of Building 2). State of the building structure 3). The

physical state of the building is damaged by 4). Build-

ing function 5). Other supporting conditions (Almais

et al., 2016).

These criteria can be used to build a Decision Sup-

port System Dynamic (DSSD) because one of the re-

quirements for building a Decision Support System

(DSS) is to have an alternative, criteria, and level of

importance. These criteria have a level of interest

each depending on the type of cautiousness. In previ-

ous research (Suhada et al., 2018) explained that the

Fuzzy-Weighted Product method is used as the Deci-

sion Support System (DSS) Determination of the cus-

tomer in obtaining credit in a BPR. The Fuzzy in the

journal is used as a level of importance (weight) of

each criterion converted to a crisp number. Because

the criteria used have different levels of importance

depending on the type of criticism, the level of im-

portance can be converted to a crisp number using

the Fuzzy method (Kusumadewi et al., 2006). Based

on the research, it applied the fuzzy-weighted product

method to determine the level of loss and damage to

the old post-disaster sector that will be used as a reha-

bilitation and reconstruction action of post-disaster.

The fuzzy result is used as a scale for the Weighted

Product (WP) method in order that each criterion has

its own scaling scale depending on the criteria. Then

the result of the WP method will be saved to make the

system data pattern to become a reference surveyor in

determining the damage and loss of the post-disaster

sector. To test the accuracy level of the Decision Sup-

port System Dynamic (DSSD) system, you can use

the method of Confusion matrix in which there is re-

call, precision, f-measure, and accuracy.

2 STATE OF THE ART

In the journal (Suhada et al., 2018) The fuzzy use

of the weighted product method is found at the level

of importance (weight) of each criterion using crisp

numbers resulting from the conversion of fuzzy num-

bers. The result of the fuzzy number conversion is

crisp numbers using reasoning theory where numbers

close to number 1, the higher the dependency rate.

According to (Kusumadewi et al., 2006) method,

the weighted Product is the classic formula of the

Multi-Criteria Decision Making method. To develop

these methods need to be developed with the addition

of a fuzzy method so that the Multi-Criteria Decision

Making (Weighted Product) method can distinguish

the use of the assessment scale of each criterion that

corresponds to each criterion.

Weighted Product is one of the solution models

on the problem FMADM (Fuzzy Multi-Attribute De-

cision Making), this method evaluates several alterna-

tives to a set of criteria whereby each criterion inter-

dependent one with Others (Suhada et al., 2018).

Weighted Product method requires normalization

process because this method multiplies the judgment

result of each criterion, the multiplication result is not

meaningful if not compared (divided) with the default

value. The importance of the criteria serves as a posi-

tive rank in the multiplication process, while the cost

weight serves as a negative rank (Suhada et al., 2018).

1. Advantages and disadvantages Weighted Product

like an analytical method, Weighted product also

has an advantage in the analysis system that can

provide value of cost and beneﬁt to the value of

each. But having weaknesses is only used in the

process of values that have a range value.

2. Stages Weighted Product The stages in the cal-

culation of the weighted product method include

a) multiplying the entire attribute for all alterna-

tives with weights as a positive rank for the cost

attribute. b) The multiplication result is sum to

generate value on each alternative. c) Divide the

value of V for each alternate with value on each

alternate. d) found the best alternative sequence

that will be the decision of the calculation of vec-

tor V then carried out the alignment sorted from

vector value V of the largest value to the smallest

and the largest vector value V (Vi) is the alter-

native Ai Elected to the best. Preference for Ai

alternatives using equations (1):

∏

j=1

(1)

While the Σ WJ = 1 and WJ is a positive value

rank for the attribute of proﬁt and negative value

to the cost attribute. The relative presentation of

each alternative uses the following equation (2)

(Nofriansyah, 2014):

Implementation Fuzzy Weighted Product Preparation Post Disaster Reconstruction and Rehabilitation Action based Dynamics Decision

Support System

273

∏

j=1

∏

j=1

)

(2)

3 RESEARCH METHODS

In this research, there are two types of data, namely

pattern data and test data. Pattern data is used to

form a data pattern using the Fuzzy-Weighted Product

(F-WP) method, while test data is used to test built-

in pattern data. The pattern data is the result of the

breakdown of Fuzzy-Weighted Product (F-WP), us-

ing the damage and loss data of the East Java provin-

cial disaster in 2010-2013. The result of the pattern

data is a result of the damage and loss of each sector

(damaged, mild, moderate and severely damaged).

The result of the pattern data will be tested using

damage and loss data after the East Java provincial

disaster in 2018. For test data only use the criteria

and assessment of each criterion to get the result of

the type of damage and loss of each sector (damaged

light, moderate damage, and severely damaged). The

result of the test data will be searched in the pattern

data already created. For more details on how to ﬁnd

or match test data with existing pattern data in the sys-

tem, see the following ﬁgure 1:

Figure 1: System ﬂows

4 RESULTS AND DISCUSSION

4.1 RESULT

4.1.1 Modeling Systems

The Model weighted product uses multiplication to

meet the rating of the criteria, where the rating of

each criterion should be attached to the corresponding

weights. This process is similar to the normalization

process, in this case, the weighted product method

is used to determine the level of loss and damage of

sectors affected by natural disasters. The result of a

weighted product method will be stored in a data store

for use as pattern data.

The ﬁrst phase of the admin to input cases using

existing criteria is the building state, the state of the

building structure, the physical state of the building

is damaged, building functions, and other supporting

conditions. The second stage gives a level of impor-

tance to each of these criteria.

The third stage gives each of these criteria based

on the alternatives that are lightly damaged, medium

damaged, and heavily damaged. The fourth stage of

the Decision Support System Dynamic (DSSD) will

automatically process the case and will result in a

number of alternatives. The ﬁfth stage tests the resul-

tant from step four above with data that has the same

characteristics. The sixth stage gets the result that is

wanted by the user (surveyor).

4.1.2 The System Needs Analysis

The need for information on the Decision Support

System Dynamic (DSSD) in preparation for rehabil-

itation and reconstruction action is: the state of the

building (C1), the state of the building structure (C2),

the physical state of the building is damaged (C3), the

function of the building (C3), and Other supporting

conditions (C4).

From each of these criteria, determined the

weight. This weight is then used for the calculation of

the Weighted Product (WP) model. The weight used

is a Fuzzy number that can be converted to a crisp

number. The determination of the number is crisp us-

ing reasoning theory where numbers close to 1, the

higher the dependency rate. Conversely, if the number

approaches 0, the dependency rate is getting lower.

There are ﬁve criteria used to determine the level

of damage and damage to the post-natural disaster

sector according to the journal (Almais et al., 2016)

namely:

CONRIST 2019 - International Conferences on Information System and Technology

274

space

Figure 2: Data criteria.

Each criterion in table 1 above has its own scale.

By using the fuzzy then the scale of each criterion

can be determined by converting the existing fuzzy

number to crisp numbers on each criterion in Figure

2 above.

a Building state assessment Scale

For the criteria of the state of the building has

three types of valuation scale is still standing, tilt

and collapse Total. Each scale of the assessment

limit has its own value. To scale the rating still

stand has a value of 0-0.33, for the tilt Rating scale

has a value of 0.33 – 0.66 and for a Total Robot

rating scale has a value of 0.66 – 1. More details

can be seen in Figure 3 below.

Figure 3: Scale of the building State assessment

b Building structure State Assessment scale

For the criteria of the state of the building, struc-

ture has three types of valuation scale IE small

partly damaged light, partly damaged and most

damaged. Each scale of the assessment limit has

its own value. To scale a small portion of the light

damaged assessment has a value of 0-0.33, for the

scale of the scoring the damaged part has a value

of 0.33 – 0.66 and for the scale of the assessment

most damage has a value of 0.66 – 1. More details

can be seen in Figure 4 below.

Figure 4: of the building structure State assessment scale

c The large scale of damaged building conditions

The criteria of the big damaged building condi-

tions have three types of assessment scale is <

30%, 30-50%, and > 50%. Each scale of the as-

sessment limit has its own value. To scale the as-

sessment of the< 30% has a value of 0-0.33, for a

rating scale of 30-50% has a value of 0.33 – 0.66

and for the rating scale > 50% has a value of 0.66

– 1. More details can be seen in Figure 5 below.

Figure 5: Scale assessment of large damaged building con-

ditions

d Building function Valuation Scale

For the criteria, building function has three types

of assessment scale is harmless, relatively danger-

ous and harmful. Each scale of the assessment

limit has its own value. The scale of the harmless

rating has a value of 0-0.33, for the relative hazard

rating scale has a value of 0.33 – 0.66, and for the

scale of hazardous assessments has a value of 0.66

– 1. More details can be seen in Figure 6 below.

Figure 6: of the building function rating Scale

e Other supporting state assessment scales

For the criteria of the large damaged building con-

ditions have three types of grading scales that are

partially damaged, mostly damaged and damaged

in Total. Each scale of the assessment limit has its

own value. For the assessment scale, the damaged

part has a value of 0-0.33; for the scale of the as-

sessment, most damage has a value of 0.33 – 0.66,

and for the scale of Total damage assessment has

a value of 0.66 – 1. More details can be seen in

Figure 7 below.

Figure 7: Other supporting conditions assessment scale

The above scale is used to perform assessments

and weight for each criterion. If implemented in

the form of a system like Figure 8 below:

Implementation Fuzzy Weighted Product Preparation Post Disaster Reconstruction and Rehabilitation Action based Dynamics Decision

Support System

275

space

Figure 8: Fuzzy-Weighted Product (F-WP) implementation

4.2 DISCUSSION

After the stage of the design entity relational diagram

and implementation in programming to create data

patterns and test data, then the next step is to test the

system using the calculation of the confusion matrix.

4.2.1 Testing Fuzzy-Weighted Product (F-WP)

The formation of data patterns on the Decision Sup-

port System (DSS) uses the Fuzzy-Weighted Product

(F-WP) method. The pattern data will be tested us-

ing test data to test the role in the pattern data. Test

pattern data will be done three times the experiment

with the test data type used the same but different con-

tent from the data. Figure 9 is a description of the

data composition used for pattern data and test data.

The test data used is the data on the damage after the

East Java provincial disaster in 2010. While the pat-

tern data uses post-disaster data in East Java province

2010-2019.

Figure 9: Data Composition.

Each trial will result in a rule and calculation of

a confusion matrix. From some experiments, there

is the same rule, so that if the test data used is al-

ways the same then the precision, recall, f-measure,

accuracy, and response time values are also the same.

It can be seen in the results of 1st, 2nd, and 3rd ex-

periments. The similarity of precision value, recall,

f-measure, accuracy, and response time of some ex-

periments is caused because the pattern data used pro-

duces the same pattern of rules, and the test data used

is also the same. In summary, test results can be seen

in Figure 10.

space

Figure 10: Test Results Fuzzy-Weighted Product (F-WP)

Method.

Based on Figure 10, it is known that the more pat-

tern data used will be more likely to be, the greater

the value of precision, recall, f-measure, and accuracy

generated. With the greater value of precision, recall,

f-measure, and accuracy are generated, the method

used is better. Therefore, by using the testing in Fig-

ure 10 above, shows the performance of the Fuzzy-

Weighted Product (F-WP) method is good and suit-

able for determining the damage and loss of post-

disaster settlement. To explain the test results of the

Fuzzy-Weighted Product (F-WP) method can be seen

in Figure 11, it illustrates the percentage results be-

tween precision value, recall, F-measure, and accu-

racy. While in Figure 9 is a graph for the test time

response from the Fuzzy-Weighted Product (F-WP)

method.

Figure 11: Test Results of The Fuzzy-Weighted Product (F-

WP) Method

Figure 12: Response Time Test Results of the Fuzzy-

Weighted Product (F-WP) Method

CONRIST 2019 - International Conferences on Information System and Technology

276

5 CONCLUSIONS

Based on the results of the tests that have been done,

it can be concluded that the Fuzzy-Weighted Product

(F-WP) method can be implemented on the Decision

Support System (DSS) to determine the damage and

losses of the post-disaster housing well. Since test-

ing the pattern data done using three different test

data can result in better data based on precision, re-

call, f-measure, and accuracy values. But for relax-

ation time, the method fuzzy-weighted product (F-

WP) the more data, the longer because in the method

of the fuzzy-weighted product (F-WP), there are mea-

sures negation of weights, aggregation of criteria and

aggregation of experts. Therefore, the more pat-

tern data used in the Fuzzy-Weighted Product (F-WP)

method, The longer the response time as each in-

coming data will be searched for the negative weight,

the critical aggregation, and the Expectation aggre-

gation using the method FuzzyWeighted Product (F-

WP). So for further research can use more test data

to get better precision, recall, f-measure, and accu-

racy values. And it can also be developed using the

other Fuzzy-Weighted Product (F-WP) methods or

Machine Learning methods to get even better relax-

ation time results.

REFERENCES

Almais, A. T. W., Sarosa, M., and Muslim, M. A. (2016).

Implementation of multi experts multi criteria deci-

sion making for rehabilitation and reconstruction ac-

tion after a disaster. MATICS, 8(1):27–31.

Kusumadewi, S., Hartati, S., Harjoko, A., and Wardoyo, R.

(2006). Fuzzy multi-attribute decision making (fuzzy

madm). Yogyakarta: Graha Ilmu, pages 78–79.

Nofriansyah, D. (2014). Concept of data mining vs decision

support system. Yogyakarta: Deepublish.

Oetari, F., J. S. Y. . N. A. (2014). Pengembangan sistem

informasi penilaian kerusakan dan kerugian bencana

menggunakan metode eclac (economic commission

for latin american and carribean)(studi kasus erupsi

gunung merapi 2010 di kab. boyolali).

Suhada, S., Hidayatulloh, T., and Fatimah, S. (2018). Pen-

erapan fuzzy madm model weighted product dalam

pengambilan keputusan kelayakan penerimaan kredit

di bpr nusamba sukaraja. JUITA: Jurnal Informatika,

6(1):61–71.

Suryadi, K. and Ramdhani, M. (2000). Decision support

system: A structural discourse of idealization and im-

plementation of the concept of decision making. Teen

Rosdakarya, 7.

Implementation Fuzzy Weighted Product Preparation Post Disaster Reconstruction and Rehabilitation Action based Dynamics Decision

Support System

277