Novel heuristic bidirectional-recurrent neural network framework for multiclass sentiment analysis classification using coot optimization

Lakshmi Revathi Krosuri

doi:10.1007/s11042-023-16133-y

2023-07-07

PRE

Lakshmi Revathi Krosuri *

Aravapalli Rama Satish

DOI：10.1007/s11042-023-16133-y

原文链接

原文求助

摘要

En 中文

In recent years, the popularity of social networking sites has skyrocketed. Customer's feedback is critical for organizations and social media that can serve as a promising tool to enhance and improve business opportunities if reviews have been evaluated on social media in a timely manner. Sentiment analysis (SA) reveals contextual interpretations in user sentiment, permitting businesses and individuals to comprehend how customers see their goods and services. Variations in textual layout, sequence length and complex logic, are some of the hurdles to effectively predict the sentiment score of customer feedback. Furthermore, a key problem in previous approaches was that they exclusively concentrated on binary or tri-classification of reviews, i.e., categorizing the opinion as positive, neutral, or negative. Ignorance of both extremely positive and extremely negative reviews can result in a misinterpretation of a consumer's feedback on a service or product, which leads a business or trend to degrade. As a result, a novel heuristic Bidirectional-Recurrent Neural Network (NHBi-RNN) for multiclass sentiment classification along with coot optimization is proposed in this study. Data acquisition, pre-processing of raw data, feature extraction and sentiment classification steps are all a part of the proposed multiclass sentiment analysis classification. Thus, the proposed framework effectively categorizes the polarity of a sentence from the consumer feedback as very positive, very negative, positive, negative, and neutral. Additionally, the efficacy of our suggested framework is evaluated by utilizing standard performance metrics. When compared to other prominent algorithms, the proposed framework performed well in terms of multiclass sentiment categorization.

Keyword:

Multiclass sentiment analysis

Natural language processing

Hybrid feature extraction

Bi-RNN

Optimization

Classification