Drink2Vec: Improving the classification of alcohol-related tweets using distributional semantics and external contextual enrichment
作者:
Highlights:
• Methods to generate features for drunk tweets classification using distributional semantics and external enrichment.
• Task-oriented word embeddings generation method using deep learning that outperforms alternative strategies.
• Use of ensembles to improve drunk tweets classification using word embeddings and semantic features.
• Assessment of each method using a broad experimental setup.
• State-of-the-art classifiers for drunk tweets.
摘要
•Methods to generate features for drunk tweets classification using distributional semantics and external enrichment.•Task-oriented word embeddings generation method using deep learning that outperforms alternative strategies.•Use of ensembles to improve drunk tweets classification using word embeddings and semantic features.•Assessment of each method using a broad experimental setup.•State-of-the-art classifiers for drunk tweets.
论文关键词:Alcohol-related tweets,Contextual enrichment,Domain embeddings learning,Convolutional neural network,Twitter
论文评审过程:Received 13 December 2019, Revised 7 July 2020, Accepted 31 July 2020, Available online 2 September 2020, Version of Record 20 October 2020.
论文官网地址:https://doi.org/10.1016/j.ipm.2020.102369