基于SVM的灾害天气预测方法的研究

Support Vector Machines Based Algorithm for the Disastrous Weather Forecast

采用针对小样本的基于统计学习理论的支持向量机进行灾害天气预测预报的研究。将灾害天气作为模式识别分类中的一类,构造气象预测的样本,建立灾害天气分类的模型,然后将其应用于夏天异常高温和凉夏预测,取得了较好的效果,可以进一步用于灾害天气预测的实践。

A new method,which forecasts the disastrous weather using the support vector machines（SVM） based on the statistical learning theory,was presented.The method took the disastrous weather as the classification in pattern recognition,and constructed the model and the training samples,then applied the method to forecast the abnormal high temperature and cool summer in summer.The result shows that this method is simplified,and the prediction is correct,and can be applied to forecast the practical disastrous weather.