A Cross-Domain Aspect-Based Sentiment Classification by Masking the Domain-Specific Words

The Aspect-Based Sentiment Classification (ABSC) models often suffer from a lack of training data in some domains. To exploit the abundant data from another domain, this work extends the original state-of-the-art LCR-Rot-hop++ model that uses a neural network with a rotatory attention mechanism for a cross-domain setting. More specifically, we propose a Domain-Independent Word Selector (DIWS) model that is used in combination with the LCR-Rot-hop++ model (DIWS-LCR-Rot-hop++). It uses attention weights from the domain classification task to determine whether a word is domain-specific or domain-independent, and discards domain-specific words when training and testing the LCR-Rot-hop++ model for cross-domain ABSC. Overall, our results confirm that DIWS-LCR-Rot-hop++ outperforms the original LCR-Rot-hop++ model under a cross-domain setting in case we impose a domain-dependent threshold value for deciding whether a word is domain-specific or not. For a target domain that is highly similar to the source domain, we find that a moderate attention threshold yields the best performance, while a target domain that is dissimilar requires a high attention threshold. Also, we observe information loss when we impose a too strict restriction and classify a small proportion of words as domain-independent.