【摘要】 为探索准确预测邵阳县烟草产量的方法,首先对邵阳县70个植烟区土壤样本中的碱解氮、有效磷、速效钾等19个养分指标进行主成分分析,得出邵阳县烟草产量主要受有机质、有效锌、有效硼、有效锰、有效硫、交换性钙、全钾、有效铁和速效钾等9个养分含量的影响。在此基础上,基于支持向量机回归算法SVR对70个植烟区的烟草产量进行回归预测。结果发现,以主成分分析后的9个养分指标作为特征变量预测得到的烟草产量的均方误差明显小于以19个养分指标作为特征变量预测得到的烟草产量的均方误差。同时,对比SVR算法和随机森林回归算法发现,SVR算法的预测精度明显优于随机森林回归算法。基于主成分与支持向量机的回归算法是预测邵阳县烟草产量的有效方法。更多还原

【Abstract】 To accurately predict the yield of tobacco in Shaoyang, soil samples from 70 tobacco growing areas were taken as the research objects. Firstly, the principal component analysis(PCA) was used to analyze the impact of 19 soil nutrient indexes on tobacco yield. The results showed that the yield of tobacco in Shaoyang was mainly affected by 9 soil nutrient indexes, including organic matter, available zinc, available boron,available manganese, available sulfur, exchangeable calcium, total potassium, available iron and available potassium. After that, support vector regression(SVR) was used to predict the tobacco yield of the 70 tobacco growing areas in Shaoyang. The results showed that the mean square error(MSE) of prediction results with 9 soil nutrient indexes was significantly less than the MSE of prediction results with 19 soil nutrient indexes.Finally, compared with that of random forest regression algorithm, the prediction accuracy of SVR was obviously better. The method based on principal component analysis and support vector regression is effective to predict tobacco yield in Shaoyang.更多还原