Skip to content

cj-wade/Domain-level-Pairwise-Semantic-Interaction-for-Aspect-Based-Sentiment-Classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Domain-level Pairwise Semantic Interaction for Aspect-Based Sentiment Classification (PAKDD 2022)

paper http://arxiv.org/abs/2202.10032

image

structure

  • datasets: This directory stores restaurant14/Restaurant15/Restaurant16/Laptop15/Laptop16 processed datasets
  • models: This directory stores our proposed PSI model
  • data_utils_I_P.py: This python file is the sampling processing of datasets of Interacting Polarity
  • train.py: This python file is the training function of PSI model

To run train.py, you can input following command: python train.py --dataset res14 --lr 2e-5 --n_classes 3 --n_samples 4

'--dataset' parameter can be set to res15/res16/laptop15/laptop16 datasets '--lr' parameter means learning rate, it can be set to 5e-5, 2e-5 '--n_classes' parameter means the number of classes, it can be set to 2, 3 '--n_samples' parameter means the number of samples per class, it can be set to 3, 4, 5

image

About

ABSA - subtask - Aspect-Based Sentiment Classification

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages