针对物联网中的评论等短文本进行情感分析时, 出现上下文依赖性差和严重的特征稀疏, 以及评论类文本的情感分析具有时效性等问题, 提出了基于词嵌入和时间加权的高斯LDA算法(TG-LDA).实验结果证明, 与同类的主题模型相比, 该模型的关键词的区分度强, 主题的一致性高.%In the sentiment analysis of short texts such as comments in the Internet of Things (IOT), there are some problems, such as poor context dependence and serious sparse features, as well as the timeliness of the sentiment analysis of commentary text.A Gaussian LDA algorithm based on word embedding and time weighting (TG-LDA) is proposed.Experimental results show that compared with the similar topic model, the model has strong keyword discrimination and high topic consistency.
展开▼
