초록

용어보기논문에서 용어와 풀이말을 자동 추출한 결과로, 시범 서비스 중입니다. AI-Helper

Abstract ▼ AI-Helper

In the country, an average of 46,000 fires occur every year and they cause massive loss of properties and human lives. In this paper, we investigate a model that predicts the probability of fire based on weather information in order to prevent such disastrous fires. We are particularly interested in...

In the country, an average of 46,000 fires occur every year and they cause massive loss of properties and human lives. In this paper, we investigate a model that predicts the probability of fire based on weather information in order to prevent such disastrous fires. We are particularly interested in decision tree as a model of not only good accuracy but also comprehensibility that supports analysis of fire factors and patterns. A decision tree is mainly used for classification but in this paper we use the model to predict the probability of fire. We also investigate Bayesian network and logistic regression models to compare with decision tree for prediction of fire probability. We use Brier score in performance evaluation of those models. From the experiment results we conclude that a decision tree performs better compared with other methods taking both of accuracy and comprehensibility into account.

주제어