中文摘要：

针对已有的文本情感分析方法并没有关注到医疗社交媒体中用户评论数据呈现非均衡分布的问题，将非均衡数据分类方法应用于医疗社交媒体用户评论情感分析研究中，该方法主要包括基于取样的方法和基于集成学习的方法，分别从数据层面和算法层面来解决医疗社交媒体中数据非均衡分布问题。与其他的方法相比，Random Subspace方法取得了最好的分类效果。实验结果验证了非均衡数据分类方法在医疗社交媒体用户评论情感分析中应用的有效性。

英文摘要：

Little attention has been paid to the imbalanced distribution of reviews datasets in heahhcare social media. In this paper, the imbalanced data classification methods are applied to analyze users＇ sentiment in heahhcare social media. Imbalanced data classification methods include sampling methods and ensemble learning methods. These methods solve the above problem from the data level and algorithm level. Compared with other methods, Subspace Random obtained the best classification results. The experimental results reveal the validity of the imbalanced data classification methods in the application of user＇ reviews sentiment analysis research in heahheare social media.