Title: Assessing opinion polarity using machine learning by correlated attributes impact fitness

Authors: P.V.R.N.S.S.V. Saileela; N. Naga Malleswara Rao

Addresses: Department of CSE, ANU College of Engineering & Technology, Guntur, AP, India ' Department of Information Technology, RVR & JC College of Engineering & Technology, Guntur, AP, India

Abstract: The social network platform such as Twitter enables the opportunity to express an opinion about an event of entertainment, administration, politics, or product, which is equally available to every individual. The opinions expressed are enabling prevention of other target audiences from the negative impacts or encourage entailing the positive impacts. On another variant, the decisive team of the corresponding target event or product can reframe their decisive factors according to the target audience's views. Hence, sentiment analysis to identify the opinion polarity is becoming a mandatory fact of the decisive factors. Machine learning is a product strategy in this regard. The other option of opinion polarity assessment is the method of analysing polarity by sentiment lexicons, which is considerably scaled down since there is an opportunity of having high volumes of input data with considerable variance in a projection of opinion. Regarding this argument, this manuscript portrayed a novel method that learns from the correlated attributes impact fitness (CAIF) to identify the opinion polarity. In this regard, the attribute impacts have scaled towards positive or negative polarity. The experimental study evinces the performance advantage of the proposal compared to other contemporary models.

Keywords: datasets; support vector machine; naïve Bayes; customer relationship management; sentiment-classification.

DOI: 10.1504/IJSCC.2022.119724

International Journal of Systems, Control and Communications, 2022 Vol.13 No.1, pp.5 - 20

Received: 27 May 2020
Accepted: 19 Aug 2020

Published online: 16 Dec 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article