An Efficient Optimized DenseNet Model for Aspect-Based Multi-Label Classification - Article

Aug 4, 2026

Upload by Maeve

Sentiment analysis is central to natural language processing, yet relying only on overall sentiment fails to capture fine-grained emotional cues. This work targets multi-label aspect extraction, where different emotional aspects must be predicted from review text. The proposed Ensemble of DenseNet based on Aquila Optimizer (EDAO) prioritizes both diversity and accuracy in multi-label learners. Experiments on seven datasets—emotions, hotels, movies, proteins, automobiles, medical, and news—combine preprocessing, Vader with Bag of Words for feature extraction, word2vec-based associations, and DenseNet fine-tuned with Aquila Optimizer. Results on aspect-based multi-labeling show up to 95–97% accuracy and demonstrate consistent gains over standard benchmarks.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages30

Reading Time1 h 16 min

AccessFree to download

Introduction

Sentiment analysis and opinion mining

Multi-label vs multi-class sentiment tasks

Proposed approach (EDAO)

algorithms Article An Efﬁcient Optimized DenseNet Model for Aspect-Based Multi-Label Classiﬁcation Nasir Ayub 1, *, Tayyaba 2, Saddam Hussain 3, Syed Sajid Ullah 4, * and Jawaid Iqbal 5 Citation: Ayub, N.; Tayyaba; Hussain, S.; Ullah, S.S.; Iqbal, J. An Efﬁcient Optimized DenseNet Model for Aspect-Based Multi-Label Classiﬁcation. Algorithms 2023, 16, 548. [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)a16120548 Academic Editor: Jesper Jansson Received: 4 October 2023 Revised: 21 November 2023 Accepted: 23 November 2023 Published: 28 November 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Creative Technologies, Air University Islamabad, Islamabad 44000, Pakistan 2 Department of Computing, Riphah International University, Faisalabad 38000, Pakistan 3 School of Digital Science, Universiti Brunei Darussalam, Jalan Tungku Link, Gadong BE1410, Brunei; [saddamicup1993@gmail.com](saddamicup1993@gmail.com) 4 Department of Information and Communication Technology, University of Agder (UiA), N-4898 Grimstad, Norway 5 Faculty of Computing, Riphah International University, Islamabad 44000, Pakistan * [Correspondence: nasir.ayub@mail.au.edu.pk](Correspondence: nasir.ayub@mail.au.edu.pk) (N.A.); [syed.s.ullah@uia.no](syed.s.ullah@uia.no) (S.S.U.) Abstract: Sentiment analysis holds great importance within the domain of natural language processing as it examines both the expressed and underlying emotions conveyed through review content. Furthermore, researchers have discovered that relying solely on the overall sentiment derived from the textual content is inadequate. Consequently, sentiment analysis was developed to extract nuanced expressions from textual information. One of the challenges in this ﬁeld is effectively extracting emotional elements using multi-label data that covers various aspects. This article presents a novel approach called the Ensemble of DenseNet based on Aquila Optimizer (EDAO). EDAO is speciﬁcally designed to enhance the precision and diversity of multi-label learners. Unlike traditional multi-label methods, EDAO strongly emphasizes improving model diversity and accuracy in multi-label scenarios. To evaluate the effectiveness of our approach, we conducted experiments on seven distinct datasets, including emotions, hotels, movies, proteins, automobiles, medical, news, and birds. Our initial strategy involves establishing a preprocessing mechanism to obtain precise and reﬁned data. Subsequently, we used the Vader tool with Bag of Words (BoW) for feature extraction. In the third stage, we created word associations using the word2vec method. The improved data were also used to train and test the DenseNet model, which was ﬁne-tuned using the Aquila Optimizer (AO) . On the news, emotion, auto, bird, movie, hotel, protein, and medical datasets, utilizing the aspectbased multi-labeling technique, we achieved accuracy rates of 95%, 97%, and 96%, respectively, with DenseNet-AO. Our proposed model demonstrates that EDAO outperforms other standard methods across various multi-label datasets with different dimensions. The implemented strategy has been rigorously validated through experimental results, showcasing its effectiveness compared to existing benchmark approaches. Keywords: classiﬁcation; multi-labeling; natural language processing; deep learning; optimization method; sentiment analysis 1. Introduction Sentiment analysis, also known as opinion mining, has become an increasingly important topic of discussion in various ﬁelds. It is crucial in evaluating customer feedback, understanding public opinions, and tracking real-world

FAQs

Why is aspect-based multi-label sentiment analysis important compared with overall sentiment?

Overall sentiment summarizes text but misses nuanced emotional elements. Aspect-based multi-labeling extracts different emotional aspects covered by the review content.

What is EDAO and how does it improve multi-label learning?

EDAO is an Ensemble of DenseNet based on Aquila Optimizer, designed to enhance both precision and diversity of multi-label learners rather than using only traditional multi-label strategies.

How was the proposed model evaluated and what datasets were used?

Experiments were conducted on seven datasets, including emotions, hotels, movies, proteins, automobiles, medical, and news, using aspect-based multi-label labeling with DenseNet fine-tuned by Aquila Optimizer.

Aug 4, 2026

Upload by Maeve