The traditional research of building fire probability analysis is from statistics or fire science. This paper combines the two methods and aims to improve the statistical method of building ignition probability determination according to the research conclusion of fire science. The specific factors that affect the ignition probability are divided into three aspects: humans, ignition sources and combustibles and environments. On this basis, the Bayesian network of building ignition probability is constructed, the nodes and conditional probability table in the Bayesian network are introduced in detail, according to which the ignition probability of building can be calculated quantitatively and objectively. Then some typical buildings are chosen as examples for the application of the method, the posterior probability value is calculated by obtaining the relevant building information and substituting them into the Bayesian network. The ignition probability is dynamic, and the comparison with the statistical data of building fire also proves its rationality.

中文翻译：

---

基于贝叶斯网络的建筑物点火概率确定改进方法

建筑火灾概率分析的传统研究来自统计学或火灾科学。本文结合这两种方法，旨在根据火灾科学的研究结论，改进建筑物起火概率确定的统计方法。影响着火概率的具体因素分为三个方面：人、火源和可燃物和环境。在此基础上构建了建筑物着火概率的贝叶斯网络，详细介绍了贝叶斯网络中的节点和条件概率表，据此可以定量、客观地计算建筑物的着火概率。然后选取一些典型的建筑物作为应用该方法的例子，后验概率值是通过获取相关的建筑物信息并将其代入贝叶斯网络来计算的。点火概率是动态的，与建筑火灾统计数据的对比也证明了其合理性。