Machine learning for predicting landslide risk of Rohingya refugee camp infrastructure

Abstract

Since the dawn of human civilization, forced migration scenarios have been witnessed in different regions and populations, and is still present in the twenty-first century. The current largest population of stateless refugees in the world, the Rohingya people, reside in the southeastern border region of Bangladesh. Due to rapid expansion of refugee camps and lack of suitable locations, a large proportion of the infrastructure are at risk of landslides. This study aims to use machine learning for predicting landslide risk of camp infrastructure using geospatial features. Four supervised classification algorithms have been employed viz., (i) Logistic Regression (LR), (ii) Multi-Layer Perceptron (MLP), (iii) Gradient Boosted Trees (GBT) and (iv) Random Forest (RF) and applied on preprocessing varied versions of features. Results show that RF achieves accuracy of 76.19% and AUC of 0.76 on un-scaled features which is higher than all other algorithms. The applications of the study reside in refugee management and landslide susceptibility mapping of Rohingya camps, which can both potentially save refugee lives and serve as a case study for global applications.

Improving Spatial Agreement in Machine Learning-Based Landslide Susceptibility Mapping

“…The area receives the annual mean precipitation of 4288 mm [31]. The heavy rainfall triggers both flash floods and landslides in this area [11,12].…”

“…These cells provided the models with the necessary data during the training stage [35,36]. The sample locations were split The area receives the annual mean precipitation of 4288 mm [31]. The heavy rainfall triggers both flash floods and landslides in this area [11,12].…”

“…For instance, a major landslide triggered by heavy rain in June 2017 killed at least 156 people in the south eastern hilly region of Bangladesh where the study area is located [32]. Unplanned urbanization, rapid growth of population, hill cutting, and deforestation are associated with the recent increase in landslide hazards [11,31]. Notably, Rohingya refugee camps, especially the Kutupalong camp of Ukhia sub-district ( Figure 3) is located in areas that are highly susceptible to landslides.…”

“…Notably, Rohingya refugee camps, especially the Kutupalong camp of Ukhia sub-district ( Figure 3) is located in areas that are highly susceptible to landslides. The Kutupalong camp is considered as the most densely populated refugee settlement area in the world, where around 75,000 people live per km 2 [31,33]. Any catastrophic landslide will cause significant damage to human lives and assets.…”

“…In order to represent discrete (categorical) variables semantically, they must be considered as a composite feature (where the number of generated binary features and the number of categories are equal). These discrete variables were encoded using a one-hot encoding scheme [31], implying that multiple binary features were generated to represent a single discrete feature. The number of one-hot encoded features depends on the number of variable classes.…”