Comprehensive Analysis of Aspect Term Extraction Methods using Various Text Embeddings
This work addresses a gap in aspect-based sentiment analysis research, providing insights for practitioners in natural language processing, but it is incremental as it focuses on comparative analysis of existing methods.
The paper tackled the lack of comprehensive studies in aspect-based sentiment analysis by comparing aspect term extraction methods using various text embeddings, finding that BiLSTM outperforms LSTM and that word embedding coverage and source significantly affect performance, with CRF layers consistently improving results.
Recently, a variety of model designs and methods have blossomed in the context of the sentiment analysis domain. However, there is still a lack of wide and comprehensive studies of aspect-based sentiment analysis (ABSA). We want to fill this gap and propose a comparison with ablation analysis of aspect term extraction using various text embedding methods. We particularly focused on architectures based on long short-term memory (LSTM) with optional conditional random field (CRF) enhancement using different pre-trained word embeddings. Moreover, we analyzed the influence on the performance of extending the word vectorization step with character embedding. The experimental results on SemEval datasets revealed that not only does bi-directional long short-term memory (BiLSTM) outperform regular LSTM, but also word embedding coverage and its source highly affect aspect detection performance. An additional CRF layer consistently improves the results as well.