欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
随着时间的变化，用户对汽车产品评论的情感信息会有所波动，然而，通过挖掘这些情感信息可为潜在用户和企业提供决策依据。对于大量的汽车评论，仅靠人工去分析这些评论数据的情感显得无能为力。因此，文章采用迁移学习思想，通过前一时期标注数据获得当前时期数据的初始标注样本，利用主动学习不断优化分类模型，进而预测当前时期汽车评论的情感倾向。实验结果表明，该方法在较少人工标注量的情况下取得了较好的情感预测结果。
英文摘要：
The user sentiment information about the car product reviews is always changing over time, which can provide the decision-making basis for potential users and enterprises. It＇s impossible to analyze these car reviews only by artificial. Therefore, this paper uses transfer learning idea to get initial marked samples of unlabeled data of present period through the labeled data of prior period and uses active learning to optimize classification model constantly, then predicts the sentiment orientation of present period. The experimental results show that this method can achieve a good sentiment prediction performance in the case of less manual annotation.
同期刊论文项目
同项目期刊论文
期刊信息
位置：