中文摘要：

为了提高商品评论情感分类准确率,解决传统SVM分类时参数难以选择问题,在基本人工蜂群算法基础上,提出一种改进人工蜂群算法AABC（Advanced Artificial Bee Colony）来优化支持向量机（SVM）参数。以最小化商品评论分类错误率为优化目标,在人工蜂群算法的引领蜂阶段引入监督-响应机制增强蜂群算法开发能力,在跟随蜂阶段改进概率选择作用保证蜜源个体的差异性,提高算法收敛速度,避免算法陷入局部最优。不同商品评论情感分类结果表明,相比于GA-SVM模型、PSO-SVM模型和ABC-SVM模型,所提出的AABC-SVM模型能够寻优到更好的SVM参数组合,其分类准确率平均多提高了1%~3%,验证了所提模型的有效性。

英文摘要：

In order to improve the accuracy of sentiment classification for online product reviews and solve the problem that the traditional SVM parameters are difficult to choose,based on the standard artificial bee colony algorithm,an advanced artificial bee colony（ AABC） algorithm is proposed,which can further optimize the SVM parameter. This model puts the sentiment classification accuracy of the texts as the optimization objective. The supervision and response mechanism is adopted to enhance the capacity of population exploitation,and the probabilistic selection is enhanced to maintain the population diversity,thus it can effectively avoid the algorithm falling into local optimal. Compared to the GA-SVM model,PSO-SVM model and ABC-SVM model,experiments on different data sets,the AABC-SVM model can achieve better SVM parameters and the average classification accuracy increased by 1% ~ 3%,which verifies the effectiveness of the proposed model.