基于机器学习的多灾种灾害链预警模型重用框架设计

Design of Reusing Framework for Multi Disaster Chain Early Warning Model Based on Machine Learning

一种灾害的发生通常会伴随多种灾害同时出现,并对环境产生影响,对多灾种之间的内在关系进行分析,确定其领域知识结构,并基于知识领域对已有的单灾种灾害链预警模型进行分析,最终获得通用性较强且可重复使用的多灾种灾害链预警模型。这种方法不需要大量数据集,可以同时处理多个任务,实验表明,通过对已有单灾种预警模型的重用,预警的准确性得到一定提升。

The occurrence of one disaster will lead to multiple disasters, which will has an impact on the environment. The internal relationship among multiple disasters is analyzed, the domain knowledge structure is determined, based on the knowledge domain,the existing single disaster chain early warning model is analyzed, and finally a multi disaster chain early warning model with strong universality and reusability is obtained. This method does not need a large number of data sets, and can handle multiple tasks at the same time. Experiment results show that the accuracy of early warning can be improved by reusing the existing single disaster early warning models.