Automatic Generation of an Evaluation Model of Regional Disaster Prevention Activities Based on Self-Evaluation Questionnaire

Abstract

Regional disaster prevention activities must be evaluated in terms of their effectiveness and suitability, and then improved on the basis of this evaluation. Those who can evaluate such activities are required to have abundant on-site experience in and extensive knowledge on disaster prevention. However, there is a shortage of such talent, and the training and nurturing thereof requires considerable resources. To address these issues, machine learning was introduced in our previous study to automate the evaluation of such activities. In the present study, we propose the automatic generation of the evaluation model of such activities using the responses of a self-evaluation questionnaire as the input variables. The output variables are the results of a review committee consisting of experts on disaster prevention. This paper describes the application of the model to the fourth Disaster Prevention Map Contest, examines the predicted results, and discusses the application conditions and issues to be resolved.