Decision Tree-Based Expert System Planning To Support Temporary Housing Design Decision Making After Earthquake Disasters In Indonesia

Abstract

Indonesia is one of the countries with a high level of vulnerability to earthquake disasters, which has an impact on the urgent need for the provision of temporary housing that is fast, appropriate, and adaptive to local conditions. However, the determination of the design of Temporary Modular Shelter (TMS) so far is still reactive and lacks data-based, so it is often not in accordance with the capacity of the affected areas. This research aims to develop an expert system based on a decision tree algorithm to support the policy decision-making process in determining the design of post-earthquake TMS in Indonesia. The method used is a quantitative approach with a classification system design based on three main variables: the availability of human resources (HR), the availability of local materials, and the affordability of the location. The research instrument was tested through two trial stages with a total of 15 respondents from across earthquake-prone areas. The results of the validity test showed that all items had a Pearson correlation of > 0.900 (p < 0.001), and very high reliability with Cronbach's Alpha values of 0.975 (HR), 0.981 (local materials), and 0.984 (location). The expert system successfully mapped 27 combination conditions into 27 TMS design alternatives that were tailored to local capacity and visualized in the form of an easy-to-interpret decision tree. The conclusion of this study shows that the decision tree is an effective and practical approach to support a more contextual and standardized TMS design policy. It is recommended that this system be further developed into a GIS-based digital platform and tested for implementation in national disaster management policies.