Utjecaj točkastog inventara klizišta i statističkih metoda na procjenu podložnosti na klizanje (Gornja Dubrava, Grad Zagreb)

master's thesis

• Title: Utjecaj točkastog inventara klizišta i statističkih metoda na procjenu podložnosti na klizanje (Gornja Dubrava, Grad Zagreb) : diplomski rad
• Title (english): The impact of centroid-based landslide inventory and statistical methods on landslide susceptibility assessment (Gornja Dubrava, City of Zagreb) : master's thesis
• Author: Nik Ilić
• Mentor: Sanja Bernat Gazibara (mentor)
• Mentor: Marko Sinčić (pomoć pri izradi, neposredni voditelj)
• Committee member: Sanja Bernat Gazibara (predsjednik povjerenstva)
• Committee member: Martin Krkač (član povjerenstva)
• Committee member: Snježana Mihalić Arbanas (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mining, Geology and Petroleum Engineering; (Department of Geology and Geological Engineering); Zagreb
• Defense date and country: 2024-11-29, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mining, Petroleum and Geology Engineering; Geological Engineering

Abstract

U okviru istraživanja prikazan je utjecaj vrste ulaznog inventara klizišta (poligonski i točkasti) i statističke metode na rezultate procjene podložnosti na klizanje. Za procjenu podložnosti na klizanje korištena je kvantitativna bivarijantna statistička metoda Weight of evidence, te metode strojnog učenja Slučajna šuma i Logistička regresija. Postupak analize podložnosti na klizanje uključuje prikupljanje i obradu ulaznih podataka, izradu geomorfoloških, geoloških, hidroloških i antropogenih karata faktora klizanja, izradu modela i karata podložnosti na klizanje, verifikaciju modela metodom ROC krivulje te klasifikaciju konačnih karata u tri klase podložnosti na klizanje. Usporedba dobivenih rezultata analize podložnosti na klizanje pokazuje da su metode Weight of evidence i Logistička regresija stabilne bez obzira na oblik ulaznog inventara klizišta, dok je metoda Slučajne šume osjetljiva na promjenu geometrijskog oblika inventara, što ju čini manje pogodnom za lokalna mjerila s točkastim inventarom. Iz navedenog proizlazi da odabir metode i oblika ulaznog inventara klizišta treba prilagoditi ciljevima istraživanja i zahtjevima analize, pri čemu točkasti inventar klizišta, koji ne odražava detaljne inženjerskogeološke uvjete na terenu, nije preporučljiv za analize u krupnom mjerilu.

Abstract (english)

The research presented examines the impact of landslide inventory type (polygonal and centorid) and statistical method on the results of landslide susceptibility analysis. The assessment was conducted using the quantitative bivariate statistical method "Weight of Evidence," as well as machine learning methods, specifically Random Forest and Logistic Regression. The process of landslide susceptibility analysis includes collecting and processing input data, creating landslide factor maps (geomorphological, geological, hydrological, and anthropogenic), developing models and susceptibility maps, verifying the models using the ROC curve method, and classifying the final maps into three susceptibility classes. The comparison of the landslide susceptibility analysis results indicates that the Weight of Evidence and Logistic Regression methods are stable regardless of the landslide inventory type, while the Random Forest method is sensitive to changes in inventory type, making it less suitable for local analyses with centroid-based inventories. These results suggests that the choice of method and landslide inventory type should be adapted to specific research goals and analytical requirements and that centroid-based inventories, which do not reflect the 'real' conditions in the field, are not recommended for local-scale analyses.

• Keywords: inventar klizišta; procjena podložnosti na klizanje; strojno učenje; Weight of evidence; Logistička regresija; Slučajna šuma
• Keywords (english): landslide inventory; landslide susceptibility analysis; machine learning; Weight of evidence; Logistic regression; Random forest
• Language: croatian
• URN:NBN: urn:nbn:hr:169:811224
• Study programme: Title: Geological engineering; specializations in: Hydrogeology and engineering geology, Environmental geology; Course: Hydrogeology and engineering geology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka geološkog inženjerstva (magistar/magistra inženjer/inženjerka geološkog inženjerstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2025-04-24 09:13:28