|
Evaluation of Logistic Regression Model in Predicting Probable Landslide Areas: A Case Study, Qazvin County
|
Rouholla Jafarisirizi *  , Mohammad Hassan Khamesi-Meybodi , Javad Sadidi |
|
|
|
Abstract: (12 Views) |
Landslides are known as a potential risk. Identifying potential landslide areas can improve its management and prediction. For this purpose, 10 criteria including soil type, lithology, precipitation, distance from fault, vegetation, distance from watercourse, distance from city, distance from village, slope and elevation were selected and prepared using GIS spatial database processing. Then, a landslide risk prediction map was prepared using the logistic regression model and training data. Precision-recall and area under the ROC curve indices were used to assess accuracy. The results showed that at low thresholds, initial positive samples were well detected, but with increasing recall, the precision rate decreases to about 0.6-0.7. On the other hand, the AUC rate is 0.557, which is close to random. The results indicate that the logistic regression model and the criteria used are not highly effective in identifying landslide risk.
|
|
| Keywords: Landslide, Logistic Regression, Risk Prediction, GIS, Qazvin County |
|
|
Full-Text [DOCX 2056 kb]
(3 Downloads)
|
Type of Study: Applicable |
Subject:
Special
Received: 2025/11/28 | Accepted: 2026/08/9 | Published: 2026/08/9
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
