基于决策树算法的海州湾地区海雾预测

Research on sea fog diagnosis in Haizhou Bay based on decision tree algorithm

基于连云港西连岛站点2014—2018年逐小时气象观测资料,经过对海雾事件及气象要素特征的统计分析探寻海州湾海雾发生发展的基本规律,并基于机器学习中的经典的C4.5算法对海雾天气建立气象要素预测模型。结果表明:基于C4.5算法的决策树预测模型能够较为直观准确的对海州湾海雾进行预测,并且该决策树模型具有较高的泛化能力。利用2014—2017年的样本数据进行学习,模型的学习准确率为92.85%,利用2018年的样本数据对模型的泛化能力进行测试,测试准确率为93.51%。决策树算法在海雾预测中具有方便简洁、科学实用,准确率高等特点。

Based on the hourly meteorological observation data of the Xiliandao station in Lianyungang from 2014 to 2018,the basic law of the occurrence and development of the sea fog in Haizhou Bay was explored through the statistical analysis of the sea fog events and the characteristics of the meteorological elements,and the meteorological element diagnosis model was established based on the classical C4.5 algorithm in machine learning.Results show that the decision tree diagnosis model based on C4.5 algorithm can diagnose the fog in Haizhou Bay intuitively and accurately,and the decision tree model has high generalization ability.The 2014-2017 sample data was used for learning with 92.85%learning accuracy of the model,and the 2018 sample data was used to test the generalization ability of the model with the 93.51%test accuracy.The decision tree algorithm is simple,scientific and practical,and has high accuracy in the sea fog diagnosis.