Review on sentiment analysis for text classification techniques from 2010 to 2021

Arif Ullah, Sundas Naqeeb Khan, Nazri Mohd Nawi

Abstract: Progression in the popularity of social media activities had provided huge amount of data in the form of text that can immeasurably augment its specialty. This textual data offers a platform for the reviewers to share their comments about any product, service or event on social media. These types of discussions among the reviewers boost the demand and supply in business and industry field. Furthermore, for every passing day the textual data is also increasing in amount which makes data mining especially sentiment analysis or opinion mining, a research hungry area. This is mainly because of data is represented in the form of calculations about reviewers’ comments, assessment, attitudes, behavior and emotions to individual issues, events, topics, services and attributes. Previously, researchers focus on systems to recognize and categorize sentiments from the written material where opinions are extremely unstructured, assorted and classified. In this paper, authors try to presents a meticulous survey on sentiment analysis with classification, in which one hundred and forty three articles were reviewed regarding important activities, approaches, applications with multilingual and cross domain jobs. This systematic survey considers published literature during 2010-2021, organized based on machine learning, lexicon and hybrid approaches with multilingual and cross domain knowledge.