MPQA 3.0: An Entity/Event-Level Sentiment Corpus

By Lingjia Deng, Janyce Wiebe. 2015

167 citations recorded by the source corpus.

This paper presents an annotation scheme for adding entity and event target annotations to the MPQA corpus, a rich span-annotated opinion corpus. The new corpus promises to be a valuable new resource for developing systems for entity/event-level sentiment analysis. Such systems, in turn, would be valuable in NLP applications such as Automatic Question Answering. We introduce the idea of entity and event targets (eTargets), describe the annotation scheme, and present the results of an agreement study.

Subfields: Artificial Intelligence

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

Citations in the corpus

Modeling Naive Psychology of Characters in Simple Commonsense Stories (2018, background): “In prior literature, annotations of this complexity have typically been performed by experts (Deng and Wiebe, 2015; Ouyang and McKeown, 2015).”
When is multitask learning effective? Semantic sequence prediction under varying data conditions (2016, methodology): “MPQA : The Multi-Perspective Question Answering (MPQA) corpus (Deng and Wiebe, 2015), which contains sentiment information among others.”
Understanding Neural Networks through Representation Erasure (2016, methodology): “• Sentiment: We use the MPQA subjectivity lexicon list (Deng and Wiebe, 2015; Wilson et al., 2005), which consists roughly 8,000 lexicons.”
Beneath the Tip of the Iceberg: Current Challenges and New Directions in Sentiment Analysis Research (2020, methodology): “This trait is desirable as subjectivity and word-senses have strong semantic dependence, as highlighted in Wiebe & Mihalcea [221].”
Multitask learning for semantic sequence prediction under varying data conditions (2016, methodology): “…of ontological types for senses (Vossen et al., 1998) to convert the SUPERSENSES into coarser semantic traits like Animate or UnboundedEvent.1 MPQA: We use the MPQA (Deng and Wiebe, 2015) annotation corresponding to the coarse level of annotation, with labels like attitude and directspeech-event.”
Deep Learning--based Text Classification (2020, methodology): “The Multi-Perspective Question Answering (MPQA) dataset [183] is an opinion corpus with two class labels.”
Author’s Sentiment Prediction (2020, methodology): “We used three resources to select news articles: • MPQA (Deng and Wiebe, 2015; Wiebe et al., 2005): This dataset contains news articles manually annotated for opinions, beliefs, emotions, sentiments, speculations, etc.”
SenticNet 6: Ensemble Application of Symbolic and Subsymbolic AI for Sentiment Analysis (2020, methodology): “We used these six datasets to compare SenticNet 6 with 15 popular sentiment lexica, namely: ANEW [5], a list of 1,030 words created in 1999; WordNet-Affect [50], an extension of WordNet made of 4,787 words developed in 2004; Opinion Lexicon [22], a lexicon of 6,789 words built in the same year by me”