A Smart Decision Support System for Post-Disaster Refugee Repatriation Based on Decision Tree and Facial Emotion Analysis

Abstract

Evaluating the eligibility for post-disaster refugee repatriation plays a crucial role in the reconstruction efforts following a disaster. However, agencies such as the Batu City Regional Disaster Management Agency (BPBD) face major challenges in making objective decisions. The absence of systematic evaluation tools often leads to delays and subjectivity, especially when considering the psychological condition of refugees. This study proposes the development of an intelligent decision support system that combines a multi-criteria decision tree model with facial expression recognition to evaluate eligibility for refugee repatriation. The system assesses five main factors: (1) disaster status, (2) physical condition of refugees, (3) house condition, (4) conditions around the house, and (5) psychological conditions analyzed through facial expression detection. System testing was conducted using a decision tree with training and testing stages on data to evaluate the system's performance in decision-making. Meanwhile, the facial expression recognition module test achieved an accuracy level of 90% according to the detected expression. The proposed system provides a technology-based solution to improve the speed, objectivity, and accuracy of assessing the eligibility of refugee repatriation. This system encourages collaborative decision-making between disaster response teams, medical personnel, and psychologists, thereby supporting more effective post-disaster recovery strategies.