Pars-ABSA: an Aspect-based Sentiment Analysis dataset for Persian

ataei, Taha Shangipour; Darvishi, Kamyar; Javdan, Soroush; Minaei-Bidgoli, Behrouz; Eetemadi, Sauleh

doi:10.48550/arxiv.1908.01815

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This improvement can be increased to about 2% more by cleaning the data. The first dataset called Pars-ABSA, which is completely based on aspects in Persian language, was presented by Ataei et al [23]. To test the dataset, 6 models that have recently been used for the sentiment analysis based on aspects, in different fields in English, and their focus is on deep learning methods, were used.…”

Section: Machine Learning Based Methodsmentioning

confidence: 99%

A Hybrid Approach to Sentiment Analysis of Iranian Stock Market User’s Opinions

Ahangari

Sebti

2023

IJE

View full text Add to dashboard Cite

With the significant growth of social media, individuals and organizations are increasingly using public opinion in these media to make their own decisions. The purpose of sentiment analysis is to automatically extract people's sentiments from those social networks. Social networks related to financial markets, including stock markets, have recently attracted the attention of many individuals and organizations. people in these networks share their opinions and ideas about each share in the form of a post or tweet. In fact, sentiment analysis in this field is the assessment of people's attitude towards each share. There are different approaches in sentiment analysis, in this article, a hybrid approach is proposed for sentiment analysis. In this way the feature vector used in machine learning is obtained from a lexicon that is automatically extracted from user's tweets. This lexicon is made by using stock price information related to user's opinion. Also, by using the next day's price information of each share, amendments were suggested to this lexicon. Therefore, the lexicon generated for the feature vector was constructed in three ways, and all three methods reported about an 8% improvement over the baseline method in terms of Fscore. The baseline method that is considered for this work, is the Persian version of SentiStrength lexicon which is designed for general purpose.

show abstract

Section: Machine Learning Based Methodsmentioning

confidence: 99%