Puteri, Qonitah Alia (2026) Analisis Pola Keterkaitan Penyebab, Faktor Lingkungan, dan Jenis Bencana Menggunakan Association Rule Mining pada Data Berita Bencana di Indonesia. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6026241007-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6026241007-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy |
Abstract
Berita bencana daring memuat informasi penyebab kejadian yang beragam, tetapi sebagian besar masih berbentuk teks tidak terstruktur. Penelitian ini menganalisis pola keterkaitan antara penyebab, konteks, dan jenis bencana di Indonesia menggunakan Association Rule Mining. Data akhir terdiri atas 849 artikel, dengan 842 artikel berkonteks lengkap yang dibentuk menjadi transaksi untuk banjir, tanah longsor, kebakaran hutan dan lahan, kekeringan, serta pencemaran air. Penyebab diekstraksi melalui pencocokan kata kunci dan regular expression, contextual cue, serta semantic rescue menggunakan multilingual Sentence-BERT, kemudian dinormalisasi menjadi 27 label. Setiap transaksi diperkaya dengan kategori kepadatan penduduk, curah hujan tiga harian, dan temperatur maksimum. Genetic Algorithm menghasilkan parameter terbaik berupa minimum support 0,0145, minimum confidence 0,3770, dan maximum itemset length 4. Penerapan Apriori optimasi GAmenghasilkan 699 frequent itemsets, 1.388 aturan asosiasi, 197 target rules, dan 70 representative rules. Pola akhir menunjukkan bahwa banjir dan tanah longsor banyak dikaitkan dengan hujan deras, kebakaran hutan dan kekeringan berbagi pola cuaca panas serta musim kering, sedangkan pencemaran air didominasi sumber limbah. Hasil triangulasi menunjukkan bahwa penyebab dominan tersebut umumnya sejalan dengan literatur ilmiah, tetapi tetap dipahami sebagai pola asosiasi dalam korpus berita.
==========================================================================================================================================================================================
Online disaster news contains diverse information about the causes of disaster events, but most of this information remains in unstructured text. This study analyzes association patterns among disaster causes, contextual factors, and disaster types in Indonesia using Association Rule Mining. The final dataset consisted of 849 articles, of which 842 contained complete contextual information and were transformed into transactions covering floods, landslides, forest and land fires, droughts, and water pollution. Disaster causes were extracted using keyword matching and regular expressions, contextual cues, and semantic rescue with multilingual Sentence-BERT, and were subsequently normalized into 27 labels. Each transaction was enriched with population density categories, three-day rainfall, and maximum temperature. The Genetic Algorithm identified the best parameter combination, consisting of a minimum support of 0.0145, a minimum confidence of 0.3770, and a maximum itemset length of 4. Apriori with GA-based parameter optimization produced 699 frequent itemsets, 1,388 association rules, 197 target rules, and 70 representative rules. The final patterns indicated that floods and landslides were frequently associated with heavy rainfall; forest and land fires and droughts shared patterns related to hot weather and prolonged dry conditions; while water pollution was predominantly associated with waste-related sources. Literature triangulation showed that these dominant causes were generally consistent with findings reported in scientific studies. However, the results should be interpreted.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Association Rule Mining, Genetic Algorithm, berita bencana daring, ekstraksi penyebab, konteks, Association Rule Mining, Genetic Algorithm, online disaster news, cause extraction, contextual factors. |
| Subjects: | Q Science > Q Science (General) Q Science > Q Science (General) > Q350 Information theory |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis |
| Depositing User: | Qonitah Alia Puteri |
| Date Deposited: | 28 Jul 2026 03:11 |
| Last Modified: | 28 Jul 2026 03:11 |
| URI: | http://repository.its.ac.id/id/eprint/138250 |
Actions (login required)
 |
View Item |