A comparative study on bio-inspired algorithms for sentiment analysis
- PDF / 1,033,809 Bytes
- 21 Pages / 595.276 x 790.866 pts Page_size
- 110 Downloads / 301 Views
(0123456789().,-volV)(0123456789().,-volV)
A comparative study on bio-inspired algorithms for sentiment analysis Ashima Yadav1 • Dinesh Kumar Vishwakarma1 Received: 9 October 2019 / Revised: 9 December 2019 / Accepted: 27 January 2020 Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Data mining is one of the most explored and ongoing areas of research. Sentiment analysis is a popular application of data mining, where the information regarding the customer’s emotions or attitude is extracted by applying various methods or techniques. The earlier work in sentiment analysis deals with supervised, unsupervised machine learning-based approaches and lexicon-based approaches. Nature-inspired algorithms are recently becoming an emerging topic of research for developing new algorithms and for optimizing the results as nature serves as an excellent source of inspiration. These techniques are divided into bio-inspired algorithms, physics–chemistry based algorithms, and others. This survey mainly deals with bio-inspired algorithms, which consist of swarm intelligence based and non-swarm intelligence-based algorithms. We present a comprehensive review of the significant bio-inspired algorithms that are popularly applied in sentiment analysis. We discuss state-of-the-art on these significant algorithms along with a comparative study on these algorithms by reviewing eighty articles from various journals, conferences, book chapters, etc. Finally, this review draws some essential conclusions and identifies some research gaps to motivate researchers in this area. Keywords Data mining Evolutionary computing Optimization Review Sentiment analysis
1 Introduction Social media has become an authoritative source of communication among people to share their opinion and views about any entity. Many business organization needs to study these opinions of the users. Moreover, people rely on the feedback provided by various users on the web. This can significantly affect the buying behavior of the product. Hence, analyzing the opinions or sentiments of the user emerges as an essential field to study. Sentiment analysis (SA) is the study which deals with identifying and categorizing people’s opinion, sentiment, emotion, attitude towards any topic, people, product or event. The opinion is further categorized as positive, negative, or neutral [1].
& Dinesh Kumar Vishwakarma [email protected] Ashima Yadav [email protected] 1
Biometric Research Laboratory, Department of Information Technology, Delhi Technological University, New Delhi 110042, India
As the number of electronic devices like smartphones, tablets, cameras are increasing, the amount of data generated from them is growing tremendously [2]. Hence, the field of SA is getting a lot of attention from the researchers. People can express their opinions in the form of text, speech, videos, emoticons, etc. [3, 4]. The widespread applications of SA justify the importance of this field. It is popularly applied for business review analysis where
Data Loading...
