C4.5 Algorithm To Predict The Impact Of The Earthquake

Transcription

1 C4.5 Algorith To Predict The Ipact Of The Earthquake Efori Buulolo 1 Departeent of Coputer Engineering STMIK Budi Dara Jl. Sisingaangaraja XII. 338, Siti Rejo I, Medan Kota, Kota Medan, Suatera Utara, buuloloefori21@gail.co Natalia Silalahi 2 Departeent of Inforatics Manageent AMIK STIEKOM SUMUT Jl. Abdul Haris Nasution.19, Kwala berkala, Kota Medan, Suatera Utara, natali.cieechan.silalahi@gail.co Abstract: One of the ipacts of the quake was heavily daaged, the even tsunai killed at no less. One cause any deaths is because any can not predict the ipact of earthquakes. Data earthquakes that occurred earlier can be used to predict the incidence of the quake will probably happen soeday. One algorith that can be used to predict is the algorith C4.5. The results of the algorith C4.5 decision tree for, decision trees characteristic or condition of the earthquake and the decision, where the decision is a fruit of the earthquake that occurred odeling Keywords earthquake; ipact; the algorith C4.5 I. INTRODUCTION Earthquakes often cause assive daage, and huan casualties are not sall, one reason is that any people can not predict the incidence of the quake which occurred ainly in the earthquake-ravaged region. The earthquake can not predict when it would happen, but the expected ipact of the quake based on seisic data that never happened before. One of the ethods used to dig or search for inforation on old data is data ining algorith C4.5. The output of the algorith C4.5 in predicting the ipact of the quake is divided into three parts. Naely, there is no ipact / inor daage, severe daage, and the daage and tsunai. With predictions of the iplications of the earthquake is expected to be iniized as a result of the quake victis [1] [2]. Fadlina 3 Departeent of Inforatics Manageent AMIK STIEKOM SUMUT Jl. Abdul Haris Nasution.19, Kwala berkala, Kota Medan, Suatera Utara, fadlinako11@gail.co Robbi Rahi 4 Departeent of Coputer Engineering Medan Institute of Technology Jl. Gedung Arca.52 Kota Medan, Suatera Utara, usurobbi85@zoho.co A. Earthquake II. THEORY An earthquake is a vibration or shock caused by the release of energy fro the earth suddenly and creates seisic waves. Usually, earthquakes caused by the oveent of the earth's crust or plates [3]. Several theories have been aking the quake is the collapse of caverns below the surface of the Earth, eteor ipact on Earth's surface, due to the volcanic eruption aga activity that occurred before the volcanoes and tectonic activity. Daage caused by earthquakes is death and disability living beings, and the environental daage and the collapse of the construction of buildings and tsunai waves [4]. B. Algoriths C4.5 The c4.5 algorith is one of the data ining algoriths that included in the classification groups. C4.5 algoriths are used to for a decision tree. The resulting decision tree is the result of the algorith C4.5 and can represent and odel the results of the exploration of significant data, so the knowledge or inforation fro these data ore easily identified [5][6]. Yes Class B A2 Yes A1 Class A Class A

2 Fig 1. Decision tree exaple C.45 Entropy(S)= n ( pi log 2 pi) i=1 C4.5 algorith forula in the for of a decision tree as follows: Gain(S,A)=Entropy(S)- S i Entropy (Si) S n i 1 With: S: Set Case A: Attributes n: nuber of partitions attribute A Si : Nuber of cases in the partition toi S : Nuber of cases in S To find the value of Entropy is With: S: Set Case A: Features n: nuber of partitions S pi: a proportion of Si to S The steps of the algorith C4.5 is 1. Calculate the value Entropy (S) and Gain (S, A) to seek early roots. Old sources taken fro one of the attributes table and the value of Gain (S, A) is the est. 2. Create a branch for each value 3. For cases in branch 4. Repeat the process for each branch, until all the cases to the branches have the sae class[7] III. ANALYSIS AND DISCUSSION To predict the ipact of earthquakes with C4.5 algorith then takes the old data of the earthquake never happened before. Below are the seisic data that never happened[8] TABLE I. EARTHQUAKE DATA Region earthquake The epicenter Distance fro the beach Depth Scale (second) 1 Deli Serdang Medan I Land ,9 6 2 Deli Serdang Medan II Land , Aceh Pidie Land ,5 59 Broken 4 Nias Sea ,2 60 Broken and Tsunai 5 Aceh Sea ,1 600 Broken and Tsunai 6 Padang Sea ,6 60 Broken 7 Mentawai Sea ,8 65 Broken and Tsunai 8 Yogyakarta Land 0 17,1 5,9 57 Broken 9 Sendai, Jepang Sea , Broken and Tsunai 10 Illapel, Chile Sea ,3 180 Broken and Tsunai 11 Nepal Land ,8 25 Broken 12 Afghanistan Land ,5 30 Broken 13 West Southeast Maluku Effect Sea Morotai Sea Karo Land ,8 4 Attributes distance fro shore, depth, and duration olded into the for of the categories of data, based on the value of each attribute. TABLE II. CATEGORY DISTANCE FROM THE BEACH Distance fro the beach Categories 0 <= 100 Far > 100 Very far TABLE III. CATEGORY DEPTH Depth Categories <= 10 Deep > 10 Deeper

3 TABLE V. CATEGORY DURATION TABLE IV. CATEGORY SCALE Scale Categories <= 5 Low 5,1 7 Mediu >7,1 High Categories <=20 second Short > 20 second long Region earthquake TABLE VI. EARTHQUAKE DATA THAT HAS CATEGORIZE The epicenter Distance fro the beach Depth Scale (second) 1 Deli Serdang Medan I Land Deep Low Short 2 Deli Serdang Medan II Land Deep Mediu Short 3 Aceh Pidie Land Deepen Mediu Long Broken 4 Nias Sea Far Deeper High Long Broken and Tsunai 5 Aceh Sea Very far Deeper High Long Broken and Tsunai 6 Padang Sea Far Deeper Long Broken 7 Mentawai Sea Very far Deep High Long Broken and Tsunai 8 Yogyakarta Land Deeper Mediu Long Broken 9 Sendai, Jepang Sea Very far Deeper High Long Broken and Tsunai 10 Illapel, Chile Sea Length Deeper High Long Broken and Tsunai 11 Nepal Land Deeper High Long Broken 12 Afghanistan Land Deeper High Long Broken 13 Maluku Tenggara Barat Effect Sea Very far Deeper Low Short 14 Morotai Sea Very far Deep Low Short 15 Karo Land Deep Low Short The next step is to calculate the nuber of cases(s), the nuber of declared cases of non-effect(s 1), the nuber of cases for decision broken(s 2) and the nuber of cases reported broken and tsunai(s 3). After that calculating the gain for each attribute. The results show in the following table.

4 TABLE VII. CALCULATION NODES 1 de S S 1 S 2 S 3 Entropy Gain 1 Total , The epicenter 0, Land , Sea , Distance fro the beach 0, , Far , Very far , Depth 0, Deep , Deepen , Scale 0, Low Mediu , High , , Short Long , Fro the Table VII, the calculation could see that the est attribute is a that is equal to Thus the can be the root node. There is three attributes value, low, ediu and. Due to Low Entropy value of 0 eans, the case has classified into (S1) indicates the decision to no effect. While Mediu and High does not have decision-aking needs to be calculated again. 1.1 ediu low 1.2 Fig 2. Decision tree calculation results The next step is to calculate the 1.1 branch nodes of ediu and branch nodes of 2.1 TABLE VIII. CALCULATION NODES 1.1 de S S1 S2 S3 Entropy Gain 1.1 Scale-ediu , The epicenter 0 Land , Sea Distance fro the beach ,

5 Depth Far Very far Deep Deeper Short Long , , Fro Table VIII is the est gain value with the value duration, the duration becoes a branch node of the ediu. The duration has two branches, naely short and long, the two branches already have a decision for entropy value of 0, as shown below branch already has branched decision eans the process stops. The next step is to for a branch node of 2.1 out of. ediu below 1.2 short Long Broken Fig 3. Decision tree node calculation in 1.1 TABLE IX. CALCULATION NODES 1.2 de S S1 S2 S3 Entropy Gain 1.2 Scale , The epicenter 0, Land Sea , Distance fro the beach 0, Far , Very far Depth 0, Deep Deepen , Fro Table IX, the est value gain distance fro the beach is , the distance fro the coast to the branch node. Distance fro the beach is owned by the three branches naely. and very uch with entropy values 0, during the length because the decision did not have entropy value is not 0, then continued the following process.

6 l o ediu low Distance fro the beach short long Very far lengthy broken broken Broken and tsunai Fig 4. Decision tree node results in 1.2 To search for a branch node of the fro calculation table X, like the following. TABLE X. CALCULATION NODES de S S1 S2 S3 Entropy Gain Scale--distance fro the beach , The epicenter 0 Land Sea , Depth 0 Deep Deepen , Fro the calculation table X, The epicenter and depth have the sae gain value. The epicenter and depth ean a siilar position to be a reote branch node. In this case is ore likely to influence the ipact of the earthquake is the epicenter ediu low Distance fro the beach short Long broken broken Far Very far Broken and tsunai The epicenter sea Broken and tsunai

7 The decision tree above is the product of the algorith C4.5. A decision tree can be used to predict the ipact of the earthquake based on the characteristics and condition of the quake. The explanation of the decision tree above are as follows: a. If the is low, does not cause any effect b. If the of ediu and short duration then no effect c. If the of ediu and long duration then cause broken d. If the height and distance fro the coast 0 / happened on land, it causes broken e. If the height and distance fro the coast very far then cause broken and tsunai f. If the of height and distance fro the coast far and The epicenter sea it causes broken and tsunai IV. CONCLUSION Based on the description above can be suarized as follows: a. The data of earthquakes that has ever happened can provide useful inforation or knowledge b. Data ining algoriths can be used to predict C4.5 c. Algoriths C4.5 can predict the ipact of the quake based on seisic data that has ever happened which odeled in the for of a decision tree d. e. Ipact of earthquake affected by soe characteristics or conditions of an earthquake that is the, duration, distance fro the beach and The epicenter V. REFERENCES [1] E.Buulolo,"Ipleentasi algorita apriori pada siste persediaan obat(studi Kasus: Apotik ruah sakit estoihi edan)," Pelita Inforatika, vol.iv, no. 1, pp.71-83, 2013 [2] E.Buulolo, Algorita apriori pada data penjualan disuperarket," dala Sniti, tuktuk siadong, 2015 [3] Nandi,"gepa bui," Bandung,2016 [4] M.Ihsan.(2017,01,24),Teori Gepa bui[online].available: file=digital/ r analisa%20ketahanan- Literatur.pdf [5] M. Purnaasari and sulistiyono,"dicision Support Syste For Classification of Child Intelligence Using C4.5 Algorith", Internasional journal Of advanced research in Coputer Science(IJARCS), vol. 5, no.7,pp.16-21,2014 [6] S.Budi,"Data ining: teknik peanfaatan data untuk keperluan bisnis," Yogyakarta: Graha Ilu,2007 [7] Kusrini and E.T.Luthfi,"algorita data ining," Yogyakarta: Andi Yogyakarta [8] BMKG.(2017,01,23),Data gepa bui[online].available: i/gepabui-dirasakan.bkg Fig 5. Decision tree node calculation 1.2.1