基于随机森林原理的郑州汛期降水量预报

Precipitation Forecast in Flood Period of Zhengzhou Based on Random Forest Theory

计算机网络时代,机器学习方法不断更新并被广泛应用于金融、医学、生物学等多个领域。以进一步提高降水量预报准确率为目的,将机器学习方法应用于降水量预报,介绍了一种以随机森林为基础的汛期降水量预报模型,以郑州为例,使用随机森林预报汛期降水量,将随机森林的预报结果与BP神经网络的预报结果进行比较,结果显示,随机森林预报精度越高,准确率越高,同时避免了过度拟合的问题,稳定性强。

Since the era of computer network,machine learning methods have been continuously updated and applied in many fields,such as finance,medicine,biology and so on.To improve the accuracy of precipitation prediction,in this paper,we apply machine learning method to precipitation forecast,and introduces a precipitation forecasting model in flood season based on random forest.Taking Zhengzhou as an example,compared the precipitation in flood period which was forecasted by random forest with those of BP neural network.The random forest’s prediction accuracy was proving to be better than BP’s,and avoids the problem of overfitting and has good stability.