Parameterized Convolutional Neural Networks for Aspect Level Sentiment Classification

About

We introduce a novel parameterized convolutional neural network for aspect level sentiment classification. Using parameterized filters and parameterized gates, we incorporate aspect information into convolutional neural networks (CNN). Experiments demonstrate that our parameterized filters and parameterized gates effectively capture the aspect-specific features, and our CNN-based models achieve excellent results on SemEval 2014 datasets.

Binxuan Huang, Kathleen M. Carley · 2019

Related benchmarks

Task	Dataset	Result
Aspect Sentiment Classification	Rest SemEval 2014 (test)	Accuracy78.43	73
Aspect Sentiment Classification	Laptop (test)	Accuracy86.35	49
Aspect-based Sentiment Classification	15Rest SemEval-2015 (test)	Accuracy0.7501	32
Aspect-based Sentiment Classification	Rest SemEval 2016 (test)	Accuracy81.34	28
Aspect-based Sentiment Analysis	Lap SemEval 2014 (test)	Accuracy69.72	20
Aspect-level 3-way Sentiment Classification	Restaurant (test)	Accuracy0.792	7
Aspect-level Binary Sentiment Classification	Restaurant (test)	Accuracy90.58	7