Dimensional Sentiment Analysis Using a Regional CNN-LSTM Model

By Jin Wang, Liang-Chih Yu, K. Robert Lai, Xuejie Zhang. Annual Meeting of the Association for Computational Linguistics, 2016

517 citations recorded by the source corpus.

Dimensional sentiment analysis aims to recognize continuous numerical values in multiple dimensions such as the valencearousal (VA) space. Compared to the categorical approach that focuses on sentiment classification such as binary classification (i.e., positive and negative), the dimensional approach can provide more fine-grained sentiment analysis. This study proposes a regional CNN-LSTM model consisting of two parts: regional CNN and LSTM to predict the VA ratings of texts. Unlike a conventional CNN which considers a whole text as input, the proposed regional CNN uses an individual sentence as a region, dividing an input text into several regions such that the useful affective information in each region can be extracted and weighted according to their contribution to the VA prediction. Such regional information is sequentially integrated across regions using LSTM for VA prediction. By combining the regional CNN and LSTM, both local (regional) information within sentences and long-distance dependency across sentences can be considered in the prediction process. Experimental results show that the proposed method outperforms lexicon-based, regression-based, and NN-based methods proposed in previous studies.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/6f72ca2f58e4e7b0a4c21f99ee2fd5f6662ec5b2

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bing Liu
2. B. Liu
3. Shuai Wang
4. Lei Zhang
5. Hu Xu
6. Sven Buechel
7. Udo Hahn
8. Lei Shu

Citations in the corpus

• Predicting residential energy consumption using CNN-LSTM neural networks (2019, methodology): “analyzed the emotions using the regional CNN-LSTMmodel with text input [18].”
• Deep learning for sentiment analysis: A survey (2018, methodology): “Wang et al. 57 proposed a regional CNN-LSTM model, which consists of two parts: regional CNN and LSTM, to predict the valence arousal ratings of text.”
• Word Emotion Induction for Multiple Languages as a Deep Multi-Task Learning Problem (2018, methodology): “For word emotion induction, a very similar approach to S ENT P ROP has been proposed by Wang et al. (2016a).”
• Modeling Empathy and Distress in Reaction to News Stories (2018, background): “Only very recently, researchers started exploring more sophisticated models of human emotion on a larger scale (Wang et al., 2016; Abdul-Mageed and Ungar, 2017; Mohammad and Bravo-Marquez, 2017a; Buechel and Hahn, 2017, 2018a,b).”
• Bidirectional LSTM with attention mechanism and convolutional layer for text classification (2019, methodology): “Wang et al. [33] proposed a regional CNN-LSTM model consisting of two parts: regional CNN and LSTM to predict the valence-arousal (VA) ratings of texts.”
• EmoBank: Studying the Impact of Annotation Perspective and Representation Format on Dimensional Emotion Analysis (2017, background): “…of Ekman’s six Basic Emotions as determined by Russell and Mehrabian (1977). models have only recently received increased attention in tasks such as word and document emotion prediction (see, e.g., Yu et al. (2015), K¨oper and Schulte im Walde (2016), Wang et al. (2016), Buechel and Hahn (2016)).”
• Iterative Network Pruning with Uncertainty Regularization for Lifelong Sentiment Classification (2021, background): “In recent years, deep learning has achieved great success and been almost dominant in the field of sentiment classification [35, 36, 38].”
• Sentiment Classification by Leveraging the Shared Knowledge from a Sequence of Domains (2019, methodology): “Many aspect-based sentiment analysis methods are also reported [5,14,17,26,33,36,37,39].”
• TranS^3: A Transformer-based Framework for Unifying Code Summarization and Code Search (2020, methodology): “Unlike the traditional CNN-based approaches that su‚er from long-distance dependency problem [27] and RNN-based approaches that su‚er from excessive load imposed by sequential computation [28], transformer advances in applying the self-aŠention mechanism which can parallelize the computation and pre”