Loading [a11y]/accessibility-menu.js
AspectEmo: Multi-Domain Corpus of Consumer Reviews for Aspect-Based Sentiment Analysis | IEEE Conference Publication | IEEE Xplore

AspectEmo: Multi-Domain Corpus of Consumer Reviews for Aspect-Based Sentiment Analysis


Abstract:

Aspect-based sentiment analysis (ABSA) is a text analysis method that categorizes data by aspects and identifies the sentiment assigned to each aspect. Aspect-based senti...Show More

Abstract:

Aspect-based sentiment analysis (ABSA) is a text analysis method that categorizes data by aspects and identifies the sentiment assigned to each aspect. Aspect-based sentiment analysis can be used to analyze customer opinions by associating specific sentiments with different aspects of a product or service. Most of the work in this topic is thoroughly performed for English, but many low-resource languages still lack adequate annotated data to create automatic methods for the ABSA task. In this work, we present annotation guidelines for the ABSA task for Polish and preliminary annotation results in the form of the AspectEmo corpus, containing over 1.5k consumer reviews annotated with over 63k annotations. We present an agreement analysis on the resulting annotated corpus and preliminary results using transformer-based models trained on AspectEmo.
Date of Conference: 07-10 December 2021
Date Added to IEEE Xplore: 20 January 2022
ISBN Information:

ISSN Information:

Conference Location: Auckland, New Zealand

Funding Agency:


I. Introduction

Sentiment analysis has been one of the most active fields of research in Natural Language Processing (NLP) [1]–[5]. It has been researched as a response to the growing availability of opinion-rich resources, such as personal blogs and online review sites [6]. Rapid development can also be attributed to high business demand for user feedback. One of such use cases is processing user reviews, for instance, in ecommerce systems. In most cases, sentiment analysis focuses on inferring the sentiment of the entire review or sentence. However, it is not always the optimal choice since one review or sentence might contain information on both positive and negative aspects of a product or service. A user might like the quality of a product or service but loathe the price. The Aspect Based Sentiment Analysis (ABSA) is proposed as a method to extract user attitudes towards specific aspects of a product or service. This approach allows inferring more indepth sentiment, which could be invaluable for business as it could, e.g., allow a company to focus on refining negative aspects of their products or services while maintaining aspects valued by their users. ABSA focuses on the relationship between three clearly defined elements – a subject, an aspect of the subject, and a sentiment expressed towards subject and aspect.

Contact IEEE to Subscribe

References

References is not available for this document.