Gender-based multi-aspect sentiment detection using multilabel learning

J, Ashok Kumar; Trueman, Tina Esther; Cambria, Erik

doi:10.1016/j.ins.2022.05.057

Cited by 13 publications

(5 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Out of the nine studies, six studies [ 58 , 59 , 60 , 61 , 62 , 63 ] worked on sentiment analysis, while two studies [ 64 , 65 ] carried out work on depression detection. Study [ 66 ] approached suicide detection as their objective.…”

Section: Discussionmentioning

confidence: 99%

“…Study [ 66 ] approached suicide detection as their objective. LSTM and BiLSTM, as SDL or as HDL, were the main approaches used in studies [ 58 , 60 , 62 , 63 ], while studies [ 59 , 64 ] used statistical ML algorithms, and the authors of [ 65 ] implemented decision trees with pruning for classification. Four out of the nine studies [ 58 , 60 , 61 , 62 , 63 ] used SDL models (columns C5, C6).…”

Section: Discussionmentioning

confidence: 99%

“…LSTM and BiLSTM, as SDL or as HDL, were the main approaches used in studies [ 58 , 60 , 62 , 63 ], while studies [ 59 , 64 ] used statistical ML algorithms, and the authors of [ 65 ] implemented decision trees with pruning for classification. Four out of the nine studies [ 58 , 60 , 61 , 62 , 63 ] used SDL models (columns C5, C6). In contrast, the authors of [ 64 , 65 , 66 ] used EDL models as the main architecture through the use of stacking, feature extraction, or max voting.…”

Section: Discussionmentioning

confidence: 99%

“…Our study (R10) outperformed existing studies and yielded a mean accuracy of 91.44% across all EDL models and 95.01 % in the best EDL5 model, and an F1-Score of 0.8941 . Seven out of the nine studies [ 58 , 59 , 60 , 62 , 63 , 64 , 65 ] presented some sort of scientific or clinical validation by performing an ablation study, providing p -values of less than 0.001, or by performing cross-validation (columns C12, C13), unlike in our proposed study (R19), where we conducted exhaustive tests, including six individual statistical tests that yielded p -values of less than 0.001, deployed four cross-validation protocols, and achieved an overall standard deviation of less than 2.5% . Further, it is noted that only our proposed work (R10) conducted a true unseen paradigm (column C14) by training and testing the model on datasets of different domains, thus proving its generalizability over cross-domains.…”

Section: Discussionmentioning

confidence: 99%

See 3 more Smart Citations

Attention-Enabled Ensemble Deep Learning Models and Their Validation for Depression Detection: A Domain Adoption Paradigm

Singh

Fouda

et al. 2023

Diagnostics

View full text Add to dashboard Cite

Depression is increasingly prevalent, leading to higher suicide risk. Depression detection and sentimental analysis of text inputs in cross-domain frameworks are challenging. Solo deep learning (SDL) and ensemble deep learning (EDL) models are not robust enough. Recently, attention mechanisms have been introduced in SDL. We hypothesize that attention-enabled EDL (aeEDL) architectures are superior compared to attention-not-enabled SDL (aneSDL) or aeSDL models. We designed EDL-based architectures with attention blocks to build eleven kinds of SDL model and five kinds of EDL model on four domain-specific datasets. We scientifically validated our models by comparing “seen” and “unseen” paradigms (SUP). We benchmarked our results against the SemEval (2016) sentimental dataset and established reliability tests. The mean increase in accuracy for EDL over their corresponding SDL components was 4.49%. Regarding the effect of attention block, the increase in the mean accuracy (AUC) of aeSDL over aneSDL was 2.58% (1.73%), and the increase in the mean accuracy (AUC) of aeEDL over aneEDL was 2.76% (2.80%). When comparing EDL vs. SDL for non-attention and attention, the mean aneEDL was greater than aneSDL by 4.82% (3.71%), and the mean aeEDL was greater than aeSDL by 5.06% (4.81%). For the benchmarking dataset (SemEval), the best-performing aeEDL model (ALBERT+BERT-BiLSTM) was superior to the best aeSDL (BERT-BiLSTM) model by 3.86%. Our scientific validation and robust design showed a difference of only 2.7% in SUP, thereby meeting the regulatory constraints. We validated all our hypotheses and further demonstrated that aeEDL is a very effective and generalized method for detecting symptoms of depression in cross-domain settings.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Attention-Enabled Ensemble Deep Learning Models and Their Validation for Depression Detection: A Domain Adoption Paradigm

Singh

Fouda

et al. 2023

Diagnostics

View full text Add to dashboard Cite

show abstract

“…Untuk memahami ulasan, dibutuhkan metode Analisis sentimen atau opinion mining yang merupakan cabang Natural Language Processing (NLP) untuk mengukur dan memprediksi sentimen pengguna [12], Mengingat perkembangan web 2.0 dan jejaring sosial berkembang pesat [10]. Secara tradisional analisis sentimen mengevaluasi ulasan berdasarkan kalimat atau dokumen [13], sedangkan tingkatan yang lebih baik menggunakan Aspect-Based Sentiment Analysis (ABSA) atau jenis analisis sentimen berbasis aspek [14].…”

Section: Pendahuluanunclassified

Analisis Sentimen Berbasis Aspek Ulasan Pelanggan Restoran Menggunakan LSTM Dengan Adam Optimizer

2023

View full text Add to dashboard Cite

Consumers believe that restaurant reviews are very important when choosing a restaurant. Due to the fact that reviews have become one of the most effective ways to influence customer decisions, research that has been done on restaurant customer reviews is about sentiment analysis. Previous studies have only used sentiment analysis at the sentence or document level, while a better level uses Aspect-Based Sentiment Analysis (ABSA), or a type of aspect-based sentiment analysis. LSTM is a variant of RNN that stores long-term information in memory cells. Use of global max pooling to reduce output resolution features and prevent overfitting. In addition, the optimization method used by Adam Optimizer is an adaptive learning rate optimization algorithm specifically designed to train deep neural networks. This study aims to classify restaurant customer opinions based on aspects (food, place, service, and price) based on restaurant customer reviews on Indonesian-language TripAdvisor with LSTM and global max pooling for sentiment classification (negative, half negative, neutral, half positive, positive). The results of this study indicate that the ABSA in restaurant customer reviews for sentiment classification accuracy is 78.7% and the aspect category accuracy is 78%, both are interconnected and can help understand restaurant customer opinions on TripAdvisor.

show abstract

Depression and Disclosure Behavior via Social Media Using CNN

Tayade¹,

Kulkarni²

2022

ICT Infrastructure and Computing

View full text Add to dashboard Cite

Gender-based multi-aspect sentiment detection using multilabel learning

Cited by 13 publications

References 34 publications

Attention-Enabled Ensemble Deep Learning Models and Their Validation for Depression Detection: A Domain Adoption Paradigm

Attention-Enabled Ensemble Deep Learning Models and Their Validation for Depression Detection: A Domain Adoption Paradigm

Analisis Sentimen Berbasis Aspek Ulasan Pelanggan Restoran Menggunakan LSTM Dengan Adam Optimizer

Depression and Disclosure Behavior via Social Media Using CNN

Contact Info

Product

Resources

About