论文部分内容阅读
针对烟草化学成分与卷烟制品香级之间确定的数学模型难以建立的问题.提出了一种基于萤火虫群优化算法的烟草香级集成分类方法.方法首先使用混合核SVM独立训练多个个体支持向量机,然后利用改进的离散型萤火虫群优化算法选择部分精度较高、差异度较大的个体分类器参与集成,最后通过多数投票法得到最终的分类预测结果.对比实验结果表明,算法在分类准确度上具有较大的优势,证明了算法的有效性·从而为烟草的香级分类提供了可靠依据.
In order to solve the problem that it is difficult to establish the mathematical model between the chemical components of tobacco and the fragrance level of tobacco products, a fire class integrated classification method based on firefly swarm optimization algorithm is proposed.Methods Firstly, hybrid SVM is used to train multiple individual support vectors Machine, and then use the improved discrete firefly swarm optimization algorithm to select part of the higher accuracy and diversity of individual classifier to participate in the integration, and finally by the majority voting method to get the final classification prediction results.Comparative experimental results show that the algorithm in the classification accuracy It has proved that the algorithm is effective and thus provides a reliable basis for the classification of tobacco.