Ternary Sentiment Classification of Airline Passengers’ Twitter Text Based on BERT

The fast growing digital culture has attracted people’s attention and motivated them to interact with fascinating multimedia materials, including images, videos, voice notes and messages. Video is the most often used digital material format globally; it is used for marketing for companies, education, proof of concern, memorial ceremonies, and amusement among other purposes. The primary goal of video/image captioning is to represent the dynamics of a video clip using simple and understandable conversational language. The motivation for captioning comes from its ability to enhance video accessibility for those who are deaf or hard-of-hearing, improve concentration and memory of information and allow viewing in sound-sensitive environments. In this proposed framework, we demonstrate how to create captions for videos using Gated Recurrent Unit (GRU) along with a Sentence Vector Branch (SVB) and skip connections employing two word embedding methods: Bidirectional Encoder Representations from Transformers (BERT) and Embeddings from Language Model (ELMO). The encoder section of the model uses skip connections to address the issue of vanishing gradients and the sentence vector branch is used to improve the efficiency of the encoder-decoder architecture. The decoder generates a suitable description using the sentence vector branch together with past generated words. All of our experiments make use of the Standard MSVD databases. BLEU@4, METEOR, ROUGE and CIDEr measures have been applied for the performance assessment. Comparative analysis of BERT and ELMO word embeddings reveals their respective strengths. Features are extracted from pre-trained CNNs NASNet-Large, VGG16, Inceptionv3, DenseNet-169 and ResNet50. In the MSVD dataset, BERT beats ELMO in the case of the GRU-based model, whereas ELMO beats BERT in the case of the LSTM-based model; among LSTM and GRU, LSTM outperforms GRU. DenseNet-169 beats every feature extraction method used in the MSVD dataset.

This study copes with extracting adverse effects (AEs) from biomedical texts. An adverse effect is a noxious, unintended, and undesired effect caused by the administration of an external entity such as medication, dietary supplement, radiotherapy, and others. A binary classifier is proposed to filter out irrelevant texts from AE assertive texts and a sequence labeling model for extracting the AE mentions. Both models are built by consolidating the cutting-edge deep learning technologies: Bidirectional Encoder Representations from Transformers (BERT), Embeddings from Language Models (ELMo), and Bidirectional Gated Recurrent Units. The performances of our models are evaluated on an Adverse Drug Effects dataset constructed by sampling from Medline case studies. Both models perform significantly better than previously published models with an F1 score of 0.906 for binary classification and an approximate match F1 score of 0.925 for text labeling. The proposed models can be adapted to any tasks with similar interests.

Recently, people used social media to convey their emotions in the text. Text classification is required for group a text containing emotion or not and identify the type of emotion. Most of the research in emotion classification use machine learning methods. So this research offers the use of deep learning for the classification of emotions in text. One of the algorithms that is widely used for deep learning is BERT (Bidirectional Encoder Representations from Transformers). For the experiment, classification is divided into two phases, namely the first phase groups emotion or not emotion tweets, and the second phase groups emotion tweets into five types of emotions, namely happy, angry, sad, scared and surprised. The experiment uses three data scenarios, namely the commuter line, transjakarta and combination data. The BERT method was compared with the BiLSTM and LSTM methods for emotion classification. The result of this study is BERT generates the highest accuracy compared to other methods for the three data scenarios with values of 87.24%, 86.27%, and 86.99% in the first phase. Then, the next accuracy value followed by obtained BiLSTM, and the last is LSTM. In the second phase, BERT also obtained the highest accuracy values for the three data scenarios with values of 94.26%, 93.71 %, and 93.98%. Then, the evaluation using Precision, Recall and F1-Measure results that BERT provides the best overall method. In this study, BERT has good performance for emotion classification.

Fenomena bullying di media sosial, khususnya di Twitter, telah menjadi isu yang semakin memprihatinkan dengan dampak signifikan terhadap kesehatan mental pengguna. Dalam rangka mengatasi masalah ini, deteksi otomatis tweet yang mengandung konten bullying menjadi sangat penting. Penelitian ini bertujuan untuk membandingkan performa dua model pemrosesan bahasa alami terbaru, yaitu BERT (Bidirectional Encoder Representations from Transformers) dan XLNet, dalam klasifikasi tweet yang mengandung bullying. Metodologi penelitian ini melibatkan pengumpulan dataset tweet yang telah dilabeli sebagai bullying atau non-bullying. Proses preprocessing teks dilakukan untuk membersihkan dan menyiapkan data sebelum digunakan dalam pelatihan model. Kedua model, BERT dan XLNet, dilatih dan diuji menggunakan dataset yang sama. Evaluasi performa dilakukan dengan menggunakan metrik akurasi, presisi, recall, dan F1-score. Hasil penelitian menunjukkan bahwa kedua model memiliki kemampuan yang baik dalam mengidentifikasi tweet bullying, akan tetapi XLNet menunjukkan performa yang lebih unggul dibandingkan BERT dengan tingkat akurasi sebesar 95%. Dengan nilai presisi = 100%, recall = 0,87%, dan F1-score = 0,88%. XLNet mampu menangkap konteks dan nuansa bahasa yang lebih kompleks dalam tweet, yang berkontribusi pada akurasi klasifikasi yang lebih tinggi. Penelitian ini memberikan kontribusi penting dalam bidang deteksi bullying di media sosial dengan menunjukkan bahwa penggunaan model XLNet lebih efektif dibandingkan BERT. Temuan ini dapat membantu platform seperti Twitter dalam mengidentifikasi dan mencegah konten bullying, sehingga menciptakan lingkungan online yang lebih aman bagi pengguna, serta dapat digunakan sebagai dasar untuk pengembangan sistem deteksi bullying yang lebih canggih dan efisien di masa depan. Abstract The phenomenon of bullying on social media, particularly on Twitter, has become an increasingly concerning issue with significant impacts on users' mental health. In order to address this issue, automatic detection of tweets containing bullying content is crucial. This study aims to compare the performance of two recent natural language processing models, namely BERT (Bidirectional Encoder Representations from Transformers) and XLNet, in the classification of tweets containing bullying. The research methodology involves collecting a dataset of tweets that have been labelled as bullying or non-bullying. Text preprocessing is done to clean and prepare the data before it is used in model training. Both models, BERT and XLNet, were trained and tested using the same dataset. Performance evaluation was conducted using accuracy, precision, recall, and F1-score metrics. The results show that both models have a good ability to identify bullying tweets, but XLNet shows superior performance compared to BERT with an accuracy rate of 95%. With precision = 100%, recall = 0.87%, and F1-score = 0.88%. XLNet is able to capture more complex context and language nuances in tweets, which contributes to higher classification accuracy. This research makes an important contribution to the field of bullying detection on social media by showing that the use of the XLNet model is more effective than BERT. These findings can help platforms like Twitter identify and prevent bullying content, thereby creating a safer online environment for users, and can be used as a basis for the development of more sophisticated and efficient bullying detection systems in the future.

BackgroundThe increasing prevalence of cognitive impairment and dementia threatens global health, necessitating the development of accessible tools for detection of cognitive impairment. This study explores using a transformer‐based approach to detect cognitive impairment using acoustic markers of spontaneous speech.MethodRecordings of unstructured interviews from baseline visits were obtained from participants of The 90+ Study, a longitudinal study of individuals older than 90 years. For this analysis, participants without dementia at baseline who had stable cognitive diagnoses over the subsequent two assessments were included. Of 120 total participants (Table 1), 78 had normal cognition (NC) and 42 had cognitive impairment with no dementia (CIND). From baseline interviews, 1,273 segmented 20‐second denoised audio samples were generated. Samples from 80% of participants trained our models while the remaining 20%, unseen data, were test data. After filtering from Mel Spectrogram, Mel‐frequency cepstral coefficients (MFCC), spectral centroid (SC), and Chroma representations, 154 acoustic features were extracted. Features were transformed from 2D images into 1D string representations for interpretation by transformer‐based language models, namely Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre‐Trained Transformer‐2 (GPT‐2). Transformer‐generated feature embeddings were inputted into Random Forest (RF) and Logistic Regression (LR) models. Soft voting ensemble approach combined classifier predictions to maximize prediction power for final classification (Figure 1). To reduce the risk of overfitting, 5‐fold cross‐validation determined evaluation metric averages for baseline models, including random forest and logistic regression.ResultOur test data comprised 142, 20‐second samples from 16 NC and 113, 20‐second samples from 9 CIND. Our model achieved 100% AUC with either BERT or GPT‐2 feature embeddings. We also investigated F1 scores: BERT reached 99.56% and GPT‐2 achieved 100% (Table 2).ConclusionOur approach converts acoustic features from 2D images into 1D patterned value representations interpretable by transformer‐based language models. This renders acoustic markers suitable for detection of cognitive impairment based on language model interpretation. Both GPT‐2 and BERT models outperformed baseline models. The model will be validated with external data and increasing sample size in the next step.

We study the problem of incorporating prior knowledge into a deep Transformer-based model,i.e.,Bidirectional Encoder Representations from Transformers (BERT), to enhance its performance on semantic textual matching tasks. By probing and analyzing what BERT has already known when solving this task, we obtain better understanding of what task-specific knowledge BERT needs the most and where it is most needed. The analysis further motivates us to take a different approach than most existing works. Instead of using prior knowledge to create a new training task for fine-tuning BERT, we directly inject knowledge into BERT's multi-head attention mechanism. This leads us to a simple yet effective approach that enjoys fast training stage as it saves the model from training on additional data or tasks other than the main task. Extensive experiments demonstrate that the proposed knowledge-enhanced BERT is able to consistently improve semantic textual matching performance over the original BERT model, and the performance benefit is most salient when training data is scarce.

Public opinion is important for decision-making on numerous occasions for national growth in democratic countries like Bangladesh, the USA, and India. Sentiment analysis is a technique used to determine the polarity of opinions expressed in a text. The more complex stage of sentiment analysis is known as Aspect-Based Sentiment Analysis (ABSA), where it is possible to ascertain both the actual topics being discussed by the speakers as well as the polarity of each opinion. Nowadays, people leave comments on a variety of websites, including social networking sites, online news sources, and even YouTube video comment sections, on a wide range of topics. ABSA can play a significant role in utilizing these comments for a variety of objectives, including academic, commercial, and socioeconomic development. In English and many other popular European languages, there are many datasets for ABSA, but the Bengali language has very few of them. As a result, ABSA research on Bengali is relatively rare. In this paper, we present a Bengali dataset that has been manually annotated with five aspects and their corresponding sentiment. A baseline evaluation was also carried out using the Bidirectional Encoder Representations from Transformers (BERT) model, with 97% aspect detection accuracy and 77% sentiment classification accuracy. For aspect detection, the F1-score was 0.97 and for sentiment classification, it was 0.77.

Multitask learning (MTL) and data augmentation are becoming increasingly popular in natural language processing (NLP). These techniques are particularly useful when data are scarce. In MTL, knowledge learned from one task is applied to another. To address data scarcity, data augmentation facilitates by providing additional synthetic data during model training. In NLP, the bidirectional encoder representations from transformers (BERT) model is the default candidate for various tasks. MTL and data augmentation using BERT have yielded promising results. However, a detailed study regarding the effect of using MTL in different layers of BERT and the benefit of data augmentation in these configurations has not been conducted. In this study, we investigate the use of MTL and data augmentation from generative models, specifically for category classification, sentiment classification, and aspect-opinion sequence-labeling using BERT. The layers of BERT are categorized into top, middle, and bottom layers, which are frozen, shared, or unshared. Experiments are conducted to identify the optimal layer configuration for improved performance compared with that of single-task learning. Generative models are used to generate augmented data, and experiments are performed to reveal their effectiveness. The results indicate the effectiveness of the MTL configuration compared with single-task learning as well as the effectiveness of data augmentation using generative models for classification tasks.

In this paper we focus on an Natural Language Inference task. Being given two sentences, we classify their relation as NEUTRAL, ENTAILMENT or CONTRADICTION. Considering the achievements of BERT (Bidirectional Encoder Representations from Transformers) in many Natural Language Processing tasks, we use BERT features to create our base model for this task. However, several questions arise: can other features improve the performance obtained with BERT? If we are able to predict the situations in which BERT will fail, can we improve the performance by providing alternative models for these situations? We test several strategies and models, as alternatives to the standalone BERT model in the possible failure situations, and we take advantage of semantic features extracted from Discourse Representation Structures.

Text classification algorithms investigate the intricate relationships between words or phrases and attempt to deduce the document's interpretation. In the last few years, these algorithms have progressed tremendously. Transformer architecture and sentence encoders have proven to give superior results on natural language processing tasks. But a major limitation of these architectures is their applicability for text no longer than a few hundred words. In this paper, we explore hierarchical transfer learning approaches for long document classification. We employ pre-trained Universal Sentence Encoder (USE) and Bidirectional Encoder Representations from Transformers (BERT) in a hierarchical setup to capture better representations efficiently. Our proposed models are conceptually simple where we divide the input data into chunks and then pass this through base models of BERT and USE. Then output representation for each chunk is then propagated through a shallow neural network comprising of LSTMs or CNNs for classifying the text data. These extensions are evaluated on 6 benchmark datasets. We show that USE + CNN/LSTM performs better than its stand-alone baseline. Whereas the BERT + CNN/LSTM performs on par with its stand-alone counterpart. However, the hierarchical BERT models are still desirable as it avoids the quadratic complexity of the attention mechanism in BERT. Along with the hierarchical approaches, this work also provides a comparison of different deep learning algorithms like USE, BERT, HAN, Longformer, and BigBird for long document classification. The Longformer approach consistently performs well on most of the datasets.

With the development of Internet technology and the transformation of news media to informatization, news images, texts, and other information have exploded. The news image and text classification can effectively solve the disorder problem of news information. The early news image text classification is to establish artificial classifiers according to specific classification rules, but the classification error rate is high and the classification speed is slow. Later, machine learning technology replaced manual classification of news image texts. Although the classification efficiency has been greatly improved, the classification time is still a bit long. The bidirectional encoder representations from transformers (BERT) model uses transformer and encoder to pretrain news image text to improve classification efficiency. By comparing the differences between machine learning and BERT models in news image text classification, the experiments showed that the average precision, recall, and F1 values of the news image text classification algorithm using the BERT model were 96.6%, 95.7%, and 96.1%, respectively. All three evaluation criteria were about 5% more than the classification algorithm of the machine learning model. The classification speed of the news image text classification algorithm using the BERT model was 1.8 times that of the news image text classification algorithm based on support vector machine. Therefore, the news image text classification algorithm using the BERT model can improve the classification accuracy and efficiency of news image text.

Automatic assessment of concept map quality is an important challenge in the field of education, particularly in evaluating students' conceptual understanding objectively and efficiently. This study aims to compare the performance of two machine learning algorithms, namely Random Forest and Ordinal Regression, in classifying the quality of concept maps. The evaluation was conducted on three approaches to text feature representation: Term Frequency-Inverse Document Frequency (TF-IDF), Bidirectional Encoder Representations from Transformers (BERT), and a combination of both (TF-IDF + BERT). Additionally, this study compares the performance of the models under two dataset conditions: original data and data balanced using the Synthetic Minority Over-sampling Technique (SMOTE), to address the class imbalance that often occurs in educational data. The data used consists of a collection of propositions from students' concept maps that have been labeled with ordinal scores based on quality. Text representation is extracted using the TF-IDF and BERT approaches, and then used as input to build the classification model. Performance evaluation was conducted using the metrics of Accuracy, Precision, Recall, F1-score, Cohen’s Kappa, and MAE. The results show that the Ordinal Regression model with TF-IDF representation combined with SMOTE achieved the best performance, with an accuracy of 0.8777, an F1-score of 0.8773, and a Cohen’s Kappa of 0.7701. These results indicate that classical feature representations like TF-IDF remain effective in limited data scenarios, and that the SMOTE technique successfully improved the model's performance by reducing bias towards the majority class. This research contributes to the development of an automatic concept map assessment system and suggests optimal classification strategies for educational datasets with ordinal and imbalanced characteristics

In this study, we explore the potential of machine learning (ML) models after different text representation methods on the balanced IMDB dataset, which is widely regarded as a gold standard in sentiment analysis, one of the Natural Language processing (NLP) tasks. On the open source IMDB movie reviews dataset, we first undertake data cleaning and text representation with data preprocessing steps. Then, we apply sentiment classification using different ML models. In order to evaluate the models, we used precision (P), recall (R), F1-score (F1), and area under curve (AUC), as well as receiver operating characteristic (ROC). It is worth noting that text feature extraction with Bidirectional Encoder Representations from Transformers (BERT) provided the highest performance in all models, with the SVM model offering particularly promising results. In this model, we observed the following results: ACC 0.9033, F1 0.9308, R 0.9015, R 0.9015, P 0.9072, AUC 0.9638, and ROC 0.96. These findings suggest that NLP techniques and, in particular, machine learning models that employ BERT may offer high levels of accuracy and reliability in text classification problems. It would be beneficial for future studies to validate these findings using BERT on different NLP tasks. This would help to evaluate the effectiveness and applicability of the models in practice.

Deep learning approaches, utilizing Bidirectional Encoder Representation from Transformers (BERT) and advanced fine-tuning techniques, have achieved state-of-the-art accuracies in the domain of term extraction from texts. However, BERT presents some limitations in that it primarily captures the semantic context relative to the surrounding text without considering how relevant or central a token is to the overall document content. There has also been research on the application of sequence labeling on contextualized embeddings; however, the existing methods often rely solely on local context for extracting key phrases from texts. To address these limitations, this study proposes a centrality-weighted BERT model for key phrase extraction from text using sequence labelling (CenBERT-SEQ). The proposed CenBERT-SEQ model utilizes BERT to represent terms with various contextual embedding architectures, and introduces a centrality-weighting layer that integrates document-level context into BERT. This layer leverages document embeddings to influence the importance of each term based on its relevance to the entire document. Finally, a linear classifier layer is employed to model the dependencies between the outputs, thereby enhancing the accuracy of the CenBERT-SEQ model. The proposed CenBERT-SEQ model was evaluated against the standard BERT base-uncased model using three Computer Science article datasets, namely, SemEval-2010, WWW, and KDD. The experimental results show that, although the CenBERT-SEQ and BERT-base models achieved higher and close comparable accuracy, the proposed CenBERT-SEQ model achieved higher precision, recall, and F1-score than the BERT-base model. Furthermore, a comparison of the proposed CenBERT-SEQ model to that of related studies revealed that the proposed CenBERT-SEQ model achieved a higher accuracy, precision, recall, and F1-score of 95%, 97%, 91%, and 94%, respectively, than related studies, showing the superior capabilities of the CenBERT-SEQ model in keyphrase extraction from scientific documents.

The omnipresence of smartphones helps people to declare an emergency in realtime. In times of crisis, Twitter has become a big communication platform. As a result, more organizations are interested in tracking Twitter programmatically. Although governments and emergency management agencies work together through their respective national system for response to disasters, the sentiments of the people affected during and after the catastrophe decide the effectiveness of the disaster response and the recovery process. In recent years, sentiment analysis via Twitter-based machine learning has become a common subject. However, the detection of such tweets was often difficult due to the tweets’ language structure’s uncertainty. Thus, it is not always apparent if the words of a person announce a catastrophe. BERT (Bidirectional Encoder Representations by Transformers) is a profound learning model developed by Google. Since Google opened it, several scientists and companies have embraced it and have applied it to many text classification tasks. Therefore, we use BERT in this paper to some dataset disaster tweets. This research will help rescue and emergency responders establish effective knowledge management techniques for a rapidly evolving disaster environment.