Title: Tweets sentiment analysis using multi-lexicon features and SMO
Authors: V.P. Lijo; Hari Seetha
Addresses: School of Computer Science and Engineering, Vellore Institute of Technology, Vellore-632014, India ' School of Computer Science and Engineering, VIT-AP, Amaravati, AP, India
Abstract: Nowadays, people use social networks extensively to share their opinions about events and entities in personal life, politics and business. Twitter sentiment analysis aims at automatic identification and extraction of such opinionated data, gaining the attention of many researchers. Twitter gives a good opportunity to explore mass data, and it opens challenges of processing big data, and automatic identification and extraction of subjective information from short texts. Many methods and resources have been proposed in literature for automatic identification of subjective information from natural language texts. In this paper, we have proposed a fast mode of polarity detection using multi-lexicon features. The binary-valued multi-lexicon features save time and space for handling them for large data. Furthermore, we have used the modified-SMO with Apache Spark to process large datasets effectively. The experiments show that the method supports high scalability, and it improves the efficiency of the polarity detection.
Keywords: sentiment analysis; polarity detection; big data; Twitter analysis.
International Journal of Embedded Systems, 2021 Vol.14 No.5, pp.476 - 485
Received: 30 Mar 2020
Accepted: 17 Dec 2020
Published online: 13 Jan 2022 *