Multilingual opinion mining on YouTube – A convolutional N-gram BiLSTM word embedding

作者：

Highlights：

• The model is designed to learn a convolutional N-gram BiLSTM word embedding.

• Words are encoded by semantic and contextual information in short-long periods.

• The approach is robust across domains as well as languages.

• The approach does not rely on linguistic resources.

摘要

•The model is designed to learn a convolutional N-gram BiLSTM word embedding.•Words are encoded by semantic and contextual information in short-long periods.•The approach is robust across domains as well as languages.•The approach does not rely on linguistic resources.

论文关键词：Sentiment analysis,Multilingual opinion mining,Convolutional,N-gram word embedding,BiLSTM

论文评审过程：Received 8 September 2017, Revised 11 January 2018, Accepted 7 February 2018, Available online 16 February 2018, Version of Record 16 February 2018.

论文官网地址：https://doi.org/10.1016/j.ipm.2018.02.001