Spatial Prediction of Flood Probability in Juwana Watershed: A Comparative Analysis of Logistic Regression and Decision Tree Algorithm

Khoirurrizqi, Yusri and Ramadhan, Luthfi Rahendra and Pratama, Andhika Wisnu and Durida, Yona Putri Nengah Septiani and Ainurrafiqi, Hafidz and Setyaningsih, Dwiana Putri and Dharma, Fransiskus Assisi Wikan Surya and Pramudya, Widyaswara Angger and Widyatmanti, Wirastuti (2023) Spatial Prediction of Flood Probability in Juwana Watershed: A Comparative Analysis of Logistic Regression and Decision Tree Algorithm. In: 14th International Conference on Information and Communication Technology and System, ICTS 2023, 4 Oktober 2023, Surabaya.

[thumbnail of Spatial_Prediction_of_Flood_Probability_in_Juwana_Watershed_A_Comparative_Analysis_of_Logistic_Regression_and_Decision_Tree_Algorithm.pdf] Text
Spatial_Prediction_of_Flood_Probability_in_Juwana_Watershed_A_Comparative_Analysis_of_Logistic_Regression_and_Decision_Tree_Algorithm.pdf - Published Version
Restricted to Registered users only

Download (686kB) | Request a copy

Abstract

Flooding is a devastating natural disaster in Indonesia regarding the number of people affected, socioeconomic damage, severity, and scale of the impact. The phenomenon of floods is generally caused by multiple contributing factors and occurs within river basins. The occurrence of flood phenomena can vary in terms of intensity, frequency, and spatial distribution between different river basins, depending on the characteristics and health status of the basins. Based on these conditions, it is necessary to conduct research that can provide information regarding the prediction of which areas are potentially at high risk of flooding and identify the variables that influence floods. The objective of this paper is to examine the application of spatial data integration with statistical models (logistic regression and decision tree) to develop a flood prediction model in the Juwana Watershed. The modeling was conducted using ten variables, including landform, elevation, slope, drainage density, distance to river, (TWI), Land Use, NDWI, NDVI, and precipitation. From the modeling results, it can be observed that both models perform well in creating flood prediction models. This is evidenced by the obtained AUC values of 0.9416222 for the logistic regression model and 0.9255993 for the decision tree model. Although the logistic regression model has a higher accuracy value compared to the decision tree model, the difference in accuracy values between the two models is not substantial, and both values are close to one. This indicates that both models have good performance in making predictions.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Library Dosen
Uncontrolled Keywords: Data integration; Decision trees; Disasters; Forecasting; Land use; Rivers; Watersheds; Decision-tree model; Flood prediction; Floodings; Juwanum; Logistic Regression modeling; Logistics decisions; Logistics regressions; Prediction modelling; River basins; Spatial prediction; Floods
Subjects: G Geography. Anthropology. Recreation > G Geography (General)
Divisions: Faculty of Geography > Departemen Sains Informasi Geografi
Depositing User: Purwani Istiana ISTIANA
Date Deposited: 11 Dec 2024 03:18
Last Modified: 11 Dec 2024 03:18
URI: https://ir.lib.ugm.ac.id/id/eprint/2289

Actions (login required)

View Item
View Item