论文部分内容阅读
针对稻瘟病气象预警中样本含有模糊信息,支持向量机对含有模糊信息样本无法处理的问题,建立适合稻瘟病气象预警特点的分类预警算法(强模糊支持向量机)。以模糊事件的可信性测度为基础,将模糊分类问题转化为求解模糊机会约束规划问题;将模糊机会约束规划化转化为与其等价的二次规划,据此给出强模糊支持向量机。并且研究了强模糊支持向量机在稻瘟病气象预警中的应用方法。对浙江省宁波市某水稻种植区2004—2007年稻瘟病气象预警试验,数据结果与实际情况吻合。由此可说明强模糊支持向量机能较好地解决样本中含有模糊信息的分类问题,基于强模糊支持向量机的稻瘟病气象预警方法对于稻瘟病气象预警有较大的优越性。
Aiming at the problem that the samples of the early warning of rice blast contain fuzzy information and the support vector machine can not process the samples containing the fuzzy information, a classification and early warning algorithm (strong fuzzy support vector machine) is proposed to be suitable for the early warning of rice blast. Based on the credibility measure of fuzzy event, the fuzzy classification problem is transformed into solving the fuzzy chance constrained programming problem. The fuzzy chance constrained programming is transformed into the equivalent quadratic programming, and the strong fuzzy support vector machine is given. And the application of strong fuzzy support vector machine in early warning of rice blast is studied. The meteorological early warning test of rice blast in 2004-2007 in a rice planting area of Ningbo City, Zhejiang Province was carried out. The data are in good agreement with the actual situation. It can be shown that the strong fuzzy support vector function can better solve the classification problem of fuzzy information contained in the sample, and the meteorological early warning method based on the strong fuzzy support vector machine has a greater superiority for meteorological warning of rice blast.