Hindarto, Djarot (2026) Klasifikasi Data Crypto-Spatial Menggunakan Kecerdasan Artifisial Untuk Pemetaan Kerentanan Tanah Longsor. Doctoral thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 7022231004-Doctoral.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
7022231004-Doctoral.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Tanah longsor merupakan bencana alam yang berpotensi menimbulkan kerugian besar terhadap keselamatan manusia, infrastruktur, lingkungan, dan aktivitas sosial-ekonomi. Penelitian ini bertujuan mengembangkan klasifikasi data Crypto-Spatial menggunakan kecerdasan artifisial untuk mendukung pemetaan kerentanan tanah longsor secara akurat, adaptif, dan dapat dipercaya. Data yang digunakan meliputi stabilitas tanah, curah hujan, vegetasi, jarak terhadap sungai, kemiringan lereng, longitude, latitude, dan wilayah pengamatan. Pendekatan Multi-Criteria Decision Making, yaitu AHP-TOPSIS, TOPSIS, dan PROMETHEE, digunakan untuk membentuk pemeringkatan awal kerentanan wilayah berdasarkan data geospasial statis. Selanjutnya, model machine learning yang terdiri atas XGBoost, Random Forest, dan Support Vector Machine diterapkan untuk mengklasifikasikan tingkat kerentanan tanah longsor secara adaptif berdasarkan pola data. Hasil evaluasi menunjukkan bahwa XGBoost memberikan performa terbaik dengan akurasi 95,0%, presisi 93,0%, recall 92,0%, dan F1-score 92,5%. Pada aspek keamanan dan kepercayaan data, blockchain digunakan sebagai trust layer untuk menjamin integritas, transparansi, keterlacakan, dan auditabilitas data Crypto-Spatial. Mekanisme previous hash, current hash, transaction hash, gas, dan status ledger memungkinkan setiap data serta hasil klasifikasi dapat diverifikasi dan dideteksi apabila terjadi manipulasi. Hasil penelitian menunjukkan bahwa integrasi kecerdasan artifisial, MCDM, machine learning, dan blockchain mampu menghasilkan pemetaan kerentanan tanah longsor yang lebih informatif, aman, transparan, dan relevan sebagai dasar pengambilan keputusan mitigasi bencana bagi pemerintah, peneliti, dan pemangku kepentingan. Kerangka ini juga memperkuat proses validasi hasil analisis karena setiap perubahan data, proses evaluasi, dan keluaran klasifikasi dapat direkam secara konsisten dalam ledger sehingga mendukung system rekomendasi mitigasi yang lebih bertanggung jawab di wilayah rawan bencana.
================================================================================================================================
Landslides are natural disasters that have the potential to cause significant losses to human safety, infrastructure, the environment, and socio-economic activities. This study aims to develop Crypto-Spatial data classification using artificial intelligence to support accurate, adaptive, and trustworthy landslide susceptibility mapping. The data used include soil stability, rainfall, vegetation, distance to rivers, slope gradient, longitude, latitude, and observation area. Multi-Criteria Decision Making approaches, namely AHP-TOPSIS, TOPSIS, and PROMETHEE, are used to establish an initial ranking of regional susceptibility based on static geospatial data. Furthermore, machine learning models consisting of XGBoost, Random Forest, and Support Vector Machine are applied to adaptively classify landslide susceptibility levels based on data patterns. The evaluation results show that XGBoost achieves the best performance, with an accuracy of 95.0%, precision of 93.0%, recall of 92.0%, and F1-score of 92.5%. In terms of data security and trust, blockchain is used as a trust layer to ensure the integrity, transparency, traceability, and auditability of Crypto-Spatial data. The mechanisms of previous hash, current hash, transaction hash, gas, and ledger status enable each dataset and classification result to be verified and allow data manipulation to be detected. The results indicate that the integration of artificial intelligence, MCDM, machine learning, and blockchain can produce landslide susceptibility mapping that is more informative, secure, transparent, and relevant as a basis for disaster mitigation decision-making by governments, researchers, and stakeholders. This framework also strengthens the validation process of analytical results because every data change, evaluation process, and classification output can
be consistently recorded in the ledger, thereby supporting a more accountable mitigation recommendation system in landslide-prone areas.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | Crypto-Spatial, kecerdasan artifisial, tanah longsor, machine learning, blockchain Crypto-Spatial, artificial intelligence, landslide susceptibility, machine learning, blockchain |
| Subjects: | T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > TA Engineering (General). Civil engineering (General) > TA158.7 Computer network resources T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20001-(S3) PhD Thesis |
| Depositing User: | Djarot Hindarto |
| Date Deposited: | 27 Jul 2026 22:03 |
| Last Modified: | 27 Jul 2026 22:03 |
| URI: | http://repository.its.ac.id/id/eprint/139510 |
Actions (login required)
 |
View Item |